Report on socio-economic, market and financial assessment€¦Report on socio-economic, market and financial assessment

Deliverable D6.5.1 Dissemination Level - PU Copyright SAFESPOT

Contract N. IST-4-026963-IP

SF_D6.5.1_Socio-economic assessment_v12.doc Page 1 of 142 BLADE

SAFESPOT INTEGRATED PROJECT - IST-4-026963-IP

DELIVERABLE

SP6 – BLADE – Business models, Legal Aspects, and DEployment

Deliverable No. D6.5.1

Sub-Project No. SP6 Sub-Project Title Business models, Legal Aspects, and DEployment

Work package No. WP6.5 Work package Title Assessment & Evaluation

Task No. T6.5.8 Task Title Reporting the result of WP6.5

Authors (per company, if more than one company provide it together)

BASt, CRF, CSST, MMSE, RWS, TNO, UoC

(see p. 3 for the List of Authors)

Status (F: final; D: draft; RD: revised draft): F

Version No: 12

File Name: SF_D6.5.1_Socio-economic assessment_v12.doc

Planned Date of submission according to TA: November 2009

Issue Date: 19/01/2010

Project start date and duration 01 February 2006, 48 Months

Report on socio-economic, market and financial assessment




Revision Log

Version Date Reason Name and Company

V0.1 14/01/2009 First draft structure of the document; TOC BASt

V0.2 23/01/2009 Amendments and corrections on the structure

UoC

V0.3 23/03/2009 TOC included in the document structure; formatting adjusted

BASt

V1.0 29/05/2009 Partners´ inputs included; first combined draft document

BASt; CRF; TNO; UoC, RWS

V2.0 19/06/2009 Further input of partners included; amendments and corrections on the first draft


V3.0 16/07/2009 Further input of partners included; amendments and corrections on the second draft


V4.0 25/08/2009 Further input of partners included; amendments and corrections on the third draft


V5.0 31/08/2009 Further input of partners included; amendments and corrections on the fourth draft


V6.0 31/08/2009 Final corrections on the fifth draft included, version sent in for SAFESPOT peer-review

BASt, TNO, UoC

V7.0 11/12/2009

Final draft; put in some further explanations taking the SAFESPOT peer-review remarks into account; quality control

BASt, TNO, UoC. SF peer-review by ANAS & RWS.

V8.0 19/01/2009 Final version after revision: comments of partners included; Quality control

BASt, TNO, UoC

V8.1 06/04/2010 First page: minor changes in the document layout; overall peer reading by IP leader

CRF

V9.0 23/04/2010 Minor corrections on typing errors; minor rephrasing according to suggestions by IP leader

BASt, CRF

V10.0 07/05/2010 Minor changes of layout and phrasing

BASt, CRF

V11.0 12/05/2010 Minor changes of phrasing and layout; compiled final version for EC

BASt, CRF, TNO

V11.1 15/10/2010 Final draft version after inserting reactions to the final review report.

BASt

V12.0 21/10/2010 Final version, validated at the IP level BASt, CRF




List of Authors

Company Name

BASt Andreas Luedeke

Roland Schindhelm

CRF Michele Francano

Cristina Levizzani

Sergio Damiani

CSST Simonetta Manfredi

MMSE Matteo Revellino

Piero Mortara

RWS Tom Alkim

TNO Philippus Feenstra

Martijn de Kievit

Ellen Wilschut

Han Zwijnenberg

UoC Torsten Geissler

Ulrich Westerkamp




List of Abbreviations

ABS Antilock Braking System BCR Benefit-cost ratio CARE Community database on Accidents and Roads in Europe CBA Cost-benefit analysis ESC Electronic Stability Control EUCAR European Car Makers association EUR Euro EUROSTAT European Statistics Studies Organisation IRIS Intelligent Cooperative Intersection Safety IRTAD International Road Traffic and Accident Database ITS Intelligent Transport Systems IVSS Intelligent Vehicle Safety System H&IW Hazard and Incident Warning LAT Lateral Collision HMI Human machine interface INS TL Intersection Support – Traffic Light Assistance INS RoW Intersection Support – Right-of-way Support LATC Lateral Collision (aggregation of V2V applications) LONC Longitudinal Collision (aggregation of V2V applications) LOS Level Of Service LRR Long Range Radar Mill. Million OEM Original Equipment Manufacturer RODP Road Departure (aggregation of V2V applications) RSU Road side unit SMA Safety Margin Assistant SP Sub-project SpA Speed Alert SPE Speed Alert (as used in the eIMPACT and CODIA projects) STREP Strategic Targeted Research Project TRACE Traffic Accident Causation in Europe VAT Value-added tax V2I Vehicle-to-infrastructure V2V Vehicle-to-vehicle V2X Vehicle-to-vehicle and vehicle-to-infrastructure WLD Wireless Local Danger WP Work Package WTP Willingness To Pay




Table of Contents Revision Log ................................................................................................................................................... 2

List of Authors ................................................................................................................................................. 3

List of Abbreviations ....................................................................................................................................... 4

Table of Contents ............................................................................................................................................ 5

List of Figures ................................................................................................................................................. 7

List of Tables ................................................................................................................................................... 8

EXECUTIVE SUMMARY ........................................................................................................................... 10

1. Introduction ........................................................................................................................................... 13

1.1. Innovation and Contribution to the SAFESPOT Objectives ........................................................ 13

1.2. Methodology ................................................................................................................................ 15

1.3. Deliverable structure .................................................................................................................... 17

2. Assessment framework and methodology ............................................................................................. 19

2.1. Integrated assessment framework for socio-economic evaluation ............................................... 19

2.1.1. Core of socio-economic evaluation ......................................................................................... 19

2.1.2. Needs for socio-economic assessment ..................................................................................... 20

2.1.3. Needs for an integrated assessment approach .......................................................................... 20

2.1.4. Methodological approaches ..................................................................................................... 21

2.2. Methodological progress and performance restrictions ............................................................... 22

2.2.1. Areas of methodological progress ........................................................................................... 22

2.2.2. Performance restrictions .......................................................................................................... 24

2.3. Cost-benefit analysis (CBA) ........................................................................................................ 25

2.3.1. General methodology of CBA ................................................................................................. 25

2.3.2. CBA process ............................................................................................................................ 26

2.3.3. Scope of CBA .......................................................................................................................... 28

2.3.4. Benefits .................................................................................................................................... 29

2.3.5. Costs ........................................................................................................................................ 31

2.3.6. Benefit-Cost results ................................................................................................................. 33

2.4. Stakeholder financial analysis ...................................................................................................... 34

2.4.1. Break-even analysis ................................................................................................................. 36

2.4.2. Financial assessment for road operators and public authorities ............................................... 38

3. BLADE oriented specification of the selected applications and dedicated accident scenarios.............. 40

3.1. Introduction .................................................................................................................................. 40

3.2. V2V applications ......................................................................................................................... 41

3.2.1. Lateral Collision – LATC: Road intersection safety ............................................................... 41

3.2.2. Road Departure – RODP: Road condition status/ Slippery Road ............................................ 42

3.2.3. Longitudinal Collision – LONC: Speed limitation and safe distance ...................................... 43

3.2.4. V2V applications bundle ......................................................................................................... 44

3.3. V2I applications ........................................................................................................................... 44

3.3.1. Intelligent Cooperative Intersection Safety – IRIS: basic application ..................................... 44

3.3.2. Hazard and Incident Warning – H&IW: Reduced friction or visibility ................................... 45

3.3.3. Speed Alert (SpA): Legal speed limit ...................................................................................... 46

3.3.4. V2I applications bundle ........................................................................................................... 46

4. General accident and traffic data compilation ....................................................................................... 47

4.1. Approach ...................................................................................................................................... 47

4.2. Accident data ............................................................................................................................... 48

4.2.1. Data compilation and synchronisation ..................................................................................... 49

4.2.2. Data correction with regard to underreporting ......................................................................... 54

4.3. Traffic-related data ....................................................................................................................... 55

5. Market assessment ................................................................................................................................. 57

5.1. Business and Service Models ....................................................................................................... 57

5.2. Methodological Approach of the survey ...................................................................................... 59

5.2.1. Survey results .......................................................................................................................... 61




5.2.2. Summary of survey results ...................................................................................................... 65

6. Impact estimation .................................................................................................................................. 66

6.1. Safety impacts .............................................................................................................................. 66

6.1.1. Mechanisms of safety effects ................................................................................................... 67

6.1.2. Effects on driver behaviour for relevant impact mechanisms .................................................. 68

6.1.3. Results of safety effects ........................................................................................................... 70

6.1.4. Summary .................................................................................................................................. 83

6.2. Traffic and environmental impacts .............................................................................................. 84

6.2.1. Assumptions ............................................................................................................................ 84

6.2.2. Results ..................................................................................................................................... 85

7. Results of the Cost-Benefit Analysis ..................................................................................................... 88

7.1. Results of the V2V applications bundle ....................................................................................... 88

7.1.1. Benefits .................................................................................................................................... 88

7.1.2. Costs ........................................................................................................................................ 90

7.1.3. Benefit-Cost-Ratio ................................................................................................................... 94

7.2. Results of V2I applications bundle .............................................................................................. 95

7.2.1. Benefits .................................................................................................................................... 95

7.2.2. Costs ........................................................................................................................................ 97

7.2.3. Benefit-Cost Ratio ................................................................................................................. 103

7.3. Sensitivity and scenario analysis ................................................................................................ 105

7.3.1. Sensitivity analysis ................................................................................................................ 106

7.3.2. Scenario analysis ................................................................................................................... 110

7.3.3. Case study of the Netherlands ............................................................................................... 113

8. Stakeholder financial analysis ............................................................................................................. 117

8.1. Break-even analysis ................................................................................................................... 117

8.2. Financial assessment for road operators / public authorities ...................................................... 120

8.2.1. Road operators ....................................................................................................................... 120

8.2.2. Public authorities ................................................................................................................... 122

9. Conclusions ......................................................................................................................................... 125

10. References ...................................................................................................................................... 133

11. Annex/es ......................................................................................................................................... 137

11.1. Accident data (2005, EU-25) ..................................................................................................... 137

11.2. Market assessment: The sample ................................................................................................. 140




List of Figures Figure 1: Objective of WP5 into BLADE 14

Figure 2: Overview of methodology 16

Figure 3: Task structure of WP6.5 17

Figure 4: Description of applications considered in SAFESPOT-BLADE 24

Figure 5: Evaluation methods 29

Figure 6: Life expectancy of different elements of IVSS (own figure) 32

Figure 7: Impact level of stakeholders 36

Figure 8: Examples of the scenarios for Road Intersection safety [SAFESPOT SP4 WP2 (2006), p. 17] 42

Figure 9: Example of scenario for Slippery Road Condition [SAFESPOT SP4 WP 3 (2008), p. 53] 42

Figure 10: Example of scenario for Speed limitation and safe distance [SAFESPOT SP8 WP8.4 (2008),

p. 33] 43

Figure 11: Methodological Approach for road safety prediction for 2010 and 2020, EU-25 [WILMINK

(2008), p. 37] 53

Figure 12: Stakeholders’ typology 60

Figure 13: Market penetration forecast for Business model base – Public 61

Figure 14: Market penetration forecast for Business model base – Public/Private for V2V solution 62

Figure 15: Market penetration forecast for Service model – Private for complete sample 63

Figure 16: Market penetration forecast for combined SAFESPOT solution for three cluster (pessimistic,

intermediate, optimistic) 63

Figure 17: Increase in market penetration, if additional functions are provided 64

Figure 18: V2V Lateral Collision (LATC) – Road intersection safety effect on fatalities with 100%-fleet

penetration, EU-25. 71

Figure 19: V2V Road departure (RODP) – Road condition status: slippery road safety effect on fatalities

with 100 %-fleet penetration, EU-25 73

Figure 20: V2V Longitudinal (LONC) safety effect on fatalities with 100 %-fleet penetration, EU-25 75

Figure 21: V2I Intersection Safety effect (IRIS) on fatalities with 100 %-fleet penetration, EU-25 76

Figure 22: V2I Hazard and Incident Warning (H&IW) safety effect on fatalities with 100 %-fleet

penetration, EU-25 77

Figure 23: V2I Speed Alert (SpA) safety effect on fatalities with 100 %-fleet penetration, EU-25. 79

Figure 24: Representation of the V2V bundle 81

Figure 25: Representation of the V2I bundle 82

Figure 26: Cost composition for the V2V bundle (without implementation costs) 91

Figure 27: Risk rating of the Spanish road network (EuroRAP) 99

Figure 28: Cost composition for the V2I bundle (for penetration rate 7.9%) 103

Figure 29: Large scale versus black spot deployment of the V2I system 105

Figure 30: Effect on BCR of the scenarios 1 to 4 (difference relating to the baseline V2I case) 113

Figure 31: BCR for different scenarios and penetration rates (case study) 115




List of Tables Table 1: Cost-unit rates for fatalities and injured used in SAFESPOT 30

Table 2: Cost-unit rates for the traffic impacts 31

Table 3: Environmental aspects of the LATC use case 42

Table 4: Environmental aspects of the RODP use case 43

Table 5: Environmental aspects of the LONC use case 43

Table 6: Clusters and Applications developed in SAFESPOT SP8 WP8.4 (2008), p. 26) 44

Table 7: Environmental aspects of the IRIS use case 45

Table 8: Environmental aspects of the H&IW use case 45

Table 9: Environmental aspects of the SpA use case 46

Table 10: Definition of variables for the data enquiry 51

Table 11: One-dimensional distribution of background variables over all relevant accident classes, EU-25

(2005) 52

Table 12: Results of the road safety prediction for fatalities for 2020, EU-25 [EUROSTAT (2009); own

calculations] 54

Table 13: Vehicle stock in EU 25 for the year 2020 [PROGTRANS (2007), own Calculations] 56

Table 14: (Estimated) Vehicle mileage for EU 25 for the year 2020 [PROGTRANS (2007), own

calculations] 56

Table 15: (Estimated) Distribution of vehicle mileage for the year 2020 [INFRAS / IWW (2004), own

Calculations] 56

Table 16: (Estimated) Distribution of the level of services on the different road categories [INFRAS /IWW

(2004), own calculations] 56

Table 17: Distribution of drive configuration [BAUM et al. 2008] 56

Table 18: Overview of the Business and Service Models 58

Table 19: Estimated penetration rates for the Business and Service models 59

Table 20: Market penetration forecasts for Business model base - Public 61

Table 21: Market penetration forecast for Business model base – Public/Private 62

Table 22: Market penetration forecast for Service model – Private 62

Table 23: Market penetration forecasts for combined SAFESPOT solution 63

Table 24: New vehicle market penetration forecasts for the Business and Service Models 65

Table 25: The safety effect of LATC on fatalities and injured for full penetration 71

Table 26: Multiplication factors to determine the safety effect for the RODP application 72

Table 27: The safety effect of the RODP – Road condition status: slippery road for full penetration 73

Table 28: The safety effect of the LONC sub-application 75

Table 29: The safety effect of the IRIS basic application 77

Table 30: The safety effect of the HIW sub-application 78

Table 31: The safety effect of the Speed Alert sub-application 79

Table 32: The estimated safety effect of the V2V bundle 81

Table 33: The estimated safety effect of the V2I bundle 82

Table 34: Estimates for behavioural mechanism effects for selected applications and bundles based on

relevant aggregated accident statistics (EU-25) 83

Table 35: Estimates for behavioural mechanism effects for the bundles based on aggregated total accident

statistics for EU-25 84

Table 36: Effect of more or less congestion on motorway sections on emissions 85

Table 37: Comparison of direct traffic impacts in SAFESPOT, eIMPACT and CODIA 86

Table 38: Safety effects of the V2V bundle 90

Table 39: Safety benefits of the V2V bundle 90

Table 40: Total component costs and system costs per penetration rate 91

Table 41: Base component costs and base system costs 93

Table 42: System costs and total costs for the considered penetration rates 94

Table 43: Benefit-cost-ratios for the considered V2V business and service models 95

Table 44: Parameters for safety benefits 97

Table 45: Safety benefits of the V2I bundle 97

Table 46: Infrastructure costs 100




Table 47: Infrastructure costs for the IRIS basic application for equipping 50% of the intersections 101

Table 48: Infrastructure costs for H&IW sub-application for equipping 50% of the road net 102

Table 49: Infrastructure costs for SpA sub-application for equipping 50% of the road net 102

Table 50: V2I bundle costs for considered penetration rates and infrastructure equipment of 50 % 103

Table 51: Benefit-cost ratios for considered fleet penetration rates and 50% infrastructure equipment 103

Table 52: Results of sensitivity analysis for the V2V bundle 108

Table 53: Results of sensitivity analysis for V2I bundle 109

Table 54: Input parameters for validation scenarios 112

Table 55: Results of the relative change of BCRs for different infrastructure scenarios 112

Table 56: Scenario input for the case study for the Netherlands 114

Table 57: Data for the break-even analysis 117

Table 58: Break-even analysis: calculation parameters and results 119

Table 59: Critical mileage in case of complete private (user) funding of private infrastructure 121

Table 60: Critical mileage with complete private (user) funding of public infrastructure 122

Table 61: VAT earnings (in Mill. EUR) for different scenarios 124

Table 62: Estimates for behavioural mechanism effects of the V2V bundle 126

Table 63: Estimates for behavioural mechanism effects of the V2I bundle 126

Table 64: Benefit-cost-ratios for the considered business/service models of the V2V bundle 127

Table 65: Benefit-cost-ratios for the considered business/service models of the V2I bundle (infrastructure

equipment rate 50 %) 128

Table 66: Critical mileage with complete private (user) funding of private infrastructure 129

Table 67: Critical mileage with complete private (private) funding of public infrastructure 130

Table 68: VAT earnings (in Mill. EUR) for different scenarios 130




EXECUTIVE SUMMARY

The role of WP6.5 within BLADE is to provide a socio-economic assessment which compares two SAFESPOT cooperative system bundles. Both system bundles are based on technically specified safety applications addressing road intersection safety, hazard and incident warning regarding road condition and low visibility, and keeping speed limit and safe distance. The comparison is done for two extremes in the area of possible solutions for cooperative systems: the Vehicle-to-Vehicle concept (V2V) on the one side, and the Vehicle-to-Infrastructure concept (V2I) on the other side. The core of the assessment methodology is a cost-benefit analysis (CBA) estimating possible safety and traffic effects of the SAFESPOT bundle to prove the profitability of the system from a society point of view. To complete the assessment, a stakeholder analysis checking the profitability of the cooperative systems from the point of view of vehicle drivers, road operators and public authorities was undertaken.

The safety impact analysis showed considerable safety effects of the assessed bundles, resulting in about 7.1 % less fatalities for the V2V case, and about 8.9 % for the V2I case, assuming a 100 %-penetration rate of cooperative systems into the vehicle fleet. The figures for the injured were quite similar. Based on an estimation of the trend in fatalities and injured for 2020 in the EU-25 region, an estimation of the maximum number of fatalities and injured that could be avoided was made. From this forecast it can be expected that the V2V based SAFESPOT system has a safety potential of avoiding up to 1,476 fatalities and 63,780 injured. The V2I based system can avoid up to 1,850 fatalities and 74,264 injured. Besides safety impacts, no effects on traffic flow, fuel consumption and resulting emissions were assessed. In accordance with former studies these effects are assumed to be marginal because the applications in the bundles considered were primarily designed for safety purposes.

Benefit-cost ratios (BCR) were derived using the safety effects, the accident trend data, cost estimations, and forecasted market penetration rates for the SAFESPOT systems. For the V2V case, BCRs ranging from 1.0 to 1.1 were found for fleet penetration rates ranging from 6.1 % to 8.7 % in 2020. The fleet penetration rates have been estimated by experts during a market assessment assuming that the deployment in new vehicles starts in the year 2015. From a society point of view, the BCR calculated for the V2V based system are acceptable.

In contrast to the V2V case, the BCR calculated for the V2I based SAFESPOT system are clearly lower than 1. The calculation is based on fleet penetration rates ranging from 5.4 % to 9.5 % in 2020, and on an infrastructure equipment rate of 50 %. This result indicates that the V2I based SAFESPOT system is not efficient under the given assumptions from a society point of view. The low BCR of the V2I system is mainly caused by the high costs resulting from the large-scale equipment of infrastructure. However, the scenario analysis reveals that the profitability of the V2I system from a society point of view could be increased, if the equipment of infrastructure is done on a smaller scale concentrating on




accident black spots. The case study performed for the Netherlands underline this result showing that a partially equipped infrastructure can also create large benefits.

The cost-benefit analysis was completed by a sensitivity analysis showing that the BCR calculations were not very sensitive to variances of the different parameters.

A break-even analysis showed that the SAFESPOT bundle pays off in the V2I case for more than half of the drivers, giving the assumption, that the infrastructure can be used for free. The critical mileages calculated for the V2V case more than doubles compared to the V2I case and ranges between 24 and 30 thousand kilometres, assuming that the V2V based system is completely financed by the end user. Thus, only a small fraction of drivers may account the V2V system as being worthwhile from an economic point of view.

The financial analysis for road operators showed that a large-scale infrastructure equipment of the V2I based SAFESPOT system could not be operated economically by private or public road operators: charging cost prices would result in a level of fees generating a critical mileage for drivers which makes the system unattractive for most of the drivers.

Finally, it was also shown that public authorities would receive a considerable amount of additional VAT earnings based on equipping of vehicles and infrastructure with the SAFESPOT system. The VAT earnings depend on the relevant scenario and penetration rate and range between 181 million EURO for a penetration rate of 5 % and 395 million EURO for a penetration rate of 20 %. This amount of money may be used for fiscal-neutral supporting actions to increase the attractiveness of cooperative systems.

Regarding the question, whether the deployment process should start with the V2V based system or the V2I based system, the socio-economic assessment provides some indication for deployment strategies of combining both technologies. It seems that added value in terms of road safety can be achieved, if in addition to a V2V based solution a V2I based solution is implemented. The main indications are the following:

Firstly, both the V2V based SAFESPOT system and the V2I based SAFESPOT system use the same hardware for the in-vehicle device. Once V2V is implemented, no additional costs for the in-vehicle device arise in order to put V2I into effect for the SAFESPOT applications. Secondly, the V2V deployment demands for exceeding the 5% penetration barrier as fast as possible in order to realize effects from the V2V system. The V2I based solution could be very important in this respect, because it shows no such barrier. Third, the V2I solution requires “smart” equipment of the infrastructure concentrating on a limited number of accident black spots in order to contain costs but keeping the benefits at a high level. Therefore, it seems to be beneficial 1) to start with the V2V system implementation after having some black spots of the infrastructure equipped, 2) then strengthen the market penetration of the V2V system in order to exceed the




critical penetration rate as fast as possible, and 3), if needed, add further infrastructure equipment at other black spots.




1. Introduction

This deliverable reports the results of work package “Assessment & Evaluation” (WP6.5) within the subproject BLADE (Business models, Legal Aspects and DEployment, SP 6) of the SAFESPOT Integrated Project.

1.1. Innovation and Contribution to the SAFESPOT Objectives

The objective of the SAFESPOT Integrated Project is to develop and evaluate a “Safety Margin Assistant” which will extend in space and time the safety information available to drivers. The Safety Margin Assistant takes into account potentially dangerous situations occurring on a specific road segment, dynamic capabilities of the vehicle and the road status, and the driver status as well as the driver´s ability to manage emergency manoeuvres. Thus, the Safety Margin Assistant aims at increasing both the driver perception of near danger and the driver perception of safe driving behaviour.

To achieve this objective intelligent technologies of cooperative systems for road safety are used, which are based on V2V and V2I. Vehicles and road-side infrastructure will serve both as sources and destinations of safety-related information. The enabling technologies (e.g. a new generation of infrastructure-based sensor, accurate relative localisation and vehicle ad-hoc dynamic networking, dynamic local traffic maps) are developed, tested and evaluated in different sub-projects within SAFESPOT.

One of the key questions of SAFESPOT is how the intelligence is to be distributed between vehicles and roadside infrastructure in order to receive maximum benefit at reasonable costs. There are two main concepts under discussion:

• V2V based solution: The intelligence is predominantly in the vehicle.

• V2I based solution: The intelligence is predominantly in the infrastructure.

SAFESPOT is to analyse the implications of the two concepts and aims to identify the “optimal balance” between both concepts. Technical, socio-economic and business related criteria are to be considered.

The SAFESPOT system specified by the technical sub-projects is composed of a number of functions, each addressing a specific safety application, e. g. warning of low visibility on a road segment ahead, and warning of an emergency vehicle approaching an intersection. The bundling of applications leads to the question of how to distribute the applications between V2V and V2I technology in order to achieve the objective of an optimal balanced system. It is assumed that the optimal balanced solution may be a system that uses a combination of both concepts, V2V and V2I. Thus, a large number of hybrid types of the SAFESPOT system can be envisaged, which makes the identification of an optimal balanced system a really challenging task.




BLADE contributes to this objective by performing analytical comparisons and assessments which focus on architecture feasibility from an organisational, socio-economic and business perspective. The results are finally flowing into the definition of a deployment strategy, which is also of high importance for the conditioning of the optimal balanced SAFESPOT system. Due to the particular subject and the complexity involved, sustainability has to be proved involving many aspects (organization, legal, responsibilities, regulations, economical) evaluating risks and defining guidelines and suggestions for the different actors, as well as the government. The role of WP6.5 within BLADE is to provide a socio-economic assessment which compares two SAFESPOT system bundles (see Figure 1).

Figure 1: Objective of WP5 into BLADE

The comparison is done for two extremes in the area of possible solutions for cooperative systems: the V2V concept on the one side, and the V2I concept on the other side. Each of the SAFESPOT systems under consideration includes a bundle of safety applications, i. e. functions which are designed to assist the driver with regard to road safety. Besides safety, the systems may also affect traffic flow and emissions. The selected systems are subject to the specifications by the technical sub-projects of SAFESPOT, namely SP4 and SP5. The technical specifications provided by these sub-projects and the concentration to the extreme scenarios of a pure V2V and V2I based solution for intelligent




cooperative systems provide the framework for the assessment undertaken in WP6.5.

Considering the large number of possible types of SAFESPOT system, this approach allows evidence to be found for the question on which concept (V2V, V2I) the SAFESPOT system should be based in order to start a feasible deployment. Additionally, the approach meets restrictions of the IP regarding the time schedule and budget. The results received from the socio-economic assessment flow into the ranking of candidate business models (WP6.6) and the development of a deployment program (WP6.7). WP6.5 thus contributes to the support of the stakeholders when deciding on actions to be taken with regard to the implementation of a SAFESPOT system.

Such actions are of special importance for reaching a sufficient level of market penetration of cooperative systems for road safety. Without such a sufficient level of market penetration of the SAFESPOT system (on vehicles and/or along the infrastructure), benefits for the final user will be low. In reality, market penetration of IVSS is low. In this sub-project a market assessment will also be undertaken to envisage market penetration and provide prospective cost data of IVSS for the next decade.

The selected socio-economic assessment method also covers the distributional effects of the flow of benefits and costs attributed to IVSS between economic stakeholders in this respect, for example, road operators, users and public authorities. This innovative approach will help to indentify deployment barriers on the level of relevant economic agents and to better target a deployment plan.

1.2. Methodology

This chapter shows a short overview on the elements of the assessment methodology used and how they have been integrated in the organisational structure of the work package. A more detailed description of the methodological framework and the specific adaptations of the methodology is given in chapter 2.

The core of the approach used is a cost-benefit analysis (CBA) of the SAFESPOT system (see Figure 2). In general, a CBA of road safety measures values different physical impacts of the measures, quantifies the impacts in monetary units, and compares them to the costs of the measures by calculating the BCR.

As an input to the CBA, data are needed from system specification, impact assessment, accident and traffic data compilation, and market assessment:

First of all, a detailed specification of the applications provides a functional description of the system, taking into account the corresponding use cases (e.g. collision types, dangerous situations at road junctions). Further, the impact assessment has to be based on traffic flow and accident data which must also be synchronized to the typical use cases. The determination of the impacts provides quantitative results of the benefits in terms of reduced fatalities and injured, prevented emissions and saved travel time.




The special approach of the SAFESPOT system is to consider “co-operative” systems which communicate to each other. The user benefits of the system increase with the number of users. The market assessment therefore analyses possible business/service models and estimates the associated market penetration.

On the output side, the cost-benefit analysis can give advice about the profitability of the cooperative systems from a society´s point of view. In WP6.5, however, a wider economic assessment is performed, showing the profitability for the SAFESPOT applications from the perspective of the users, road operators and public authorities. The goal of the Financial Analysis is twofold: first it shows how the attractiveness of the systems depends on the intensity of using them (including a break-even analysis). Second, the way in which the distribution of the flow of benefits and costs influences the profitability of investments in the SAFESPOT system, from the perspective of road operators and public authorities, is demonstrated.

To complete the cost-benefit analysis and to validate and prove the stability of the results a sensitivity analysis is also undertaken for important parameters of cost-benefit calculation, such as penetration rates, discount factors, safety impacts and accident trends.

System specification

&impact assessment

Applications, impacts on safety, traffic and environment, ...

Accident and traffic

data compilation

Number of accidents, accident severity, vehicle stock, ...

Market assessment

Penetration rates, system costs (incl. economies of scale), ... following business and service models

Cost-benefit

analysis of

system bundles

Benefit-cost ratio

Sensitivity analysis

Optimistic / realistic / pessimistic scenario

Break-even analysis

for users

Benefits and Costs related to different stakeholders of the stakeholder group “users”

Financial assessment

Road operators and public authorities with focus on costs

Figure 2: Overview of methodology

The initial description of work of the SAFESPOT integrated project assumed that the approach of the WP6.5 assessment has to be adapted to specific requirements and decisions resulting from the work of the other subprojects and work packages. A structure of the tasks to be executed in WP6.5 was not provided in the initial description of work. Therefore, in a first step, the structure of




the work package and the tasks enclosed have been planned in detail. Figure 3 shows the flow chart between the eight tasks planned for WP6.5.

Figure 3: Task structure of WP6.5

1.3. Deliverable structure

The main part of the report starts with the description of the framework of the socio-economic assessment in chapter 2. It shows both the core methods (cost-benefit analysis, stakeholder analysis) and the adaptation of the approach to the requirements of the SAFESPOT project.

Chapter 3 provides a BLADE-oriented description of applications selected for the SAFESPOT systems under consideration in BLADE.

Chapter 4 shows the compilation of the relevant accident and traffic data, and provides the related figures.

Chapter 5 explains the empirical foundation for the development of the SAFESPOT business models and the determination of market penetration, and provides quantitative figures of market penetration based on expert interviews.

Chapter 6 describes the analysis and the estimation of impacts of the applications, based on literature and expert opinion.

Chapter 7 presents the results of the socio-economic assessment by showing BCRs of the SAFESPOT system bundles consisting of the applications. The results are shown for the V2V based system and the V2I based system separately. The sensitivity analysis is also presented in this chapter.

WP6.7




Chapter 8 provides the results of the economic assessment of the SAFESPOT system for important stakeholders: car drivers, road operators and public authorities.

Chapter 9 summarizes the deliverable and provides the conclusions.

Chapter 10 shows the references and chapter 11 contains the appendix in which accident, traffic and market assessment data are provided.




2. Assessment framework and methodology

This chapter introduces the methodological frames of the socio-economic assessment. The integrated approach used here consists of a cost-benefit analysis and an additional stakeholder analysis. This wider view of socio-economic impact assessment represents a recent development to evaluate not only overall economic efficiency but also the private profitability of Intelligent Vehicle Safety Systems (IVSS) for important stakeholders, e.g. end users. Such an extended approach was also used in other EU projects which were focused on a socio-economic assessment of road safety measures and IVSS, such as SEiSS, ROSEBUD and eIMPACT. Assessment objects in these projects, however, were mostly stand-alone systems. Cooperative systems, as considered in SAFESPOT, represent the next generation of IVSS with special requirements for economic assessment. Therefore, in the following we will use the term cooperative safety systems or cooperative systems – consisting of singular applications – to describe the concept of IVSS considered in the socio-economic assessment.

This chapter describes in section 2.1 the general approach of economic assessment of IVSS. However, its application on cooperative systems as well as special research questions of SAFESPOT demand for an adaptation of the assessment methodology on which section 2.2 gives an overview. In section 2.3 the methodology of cost-benefit analysis is provided with respect to the evaluation steps and data requirements. Finally, section 2.4 shows the methodology of economic assessment of IVSS from the point of view of important stakeholders.

2.1. Integrated assessment framework for socio-economic evaluation

2.1.1. Core of socio-economic evaluation

Socio-economic evaluation of the cooperative SAFESPOT applications is about reviewing economic efficiency of cooperative systems from the view of the society. A standard tool for evaluation of the profitability of road safety measures from a societal point of view is cost-benefit-analysis (CBA). CBA evaluates road safety measures by weighing the benefits against the costs of these safety measures.

In policy decision–making, the application of a CBA in an economic assessment can give advice as to whether supporting activities for market introduction of the cooperative systems, for example support via public subsidies, makes sense. Such an evaluation will be necessary especially because experience about success of supporting activities for market introduction of innovative cooperative systems has not yet been gathered. With these supporting activities, public authorities are breaking new ground. Therefore, the socio-economic evaluation will provide information and guidance about whether and how society’s welfare can be increased by the introduction of cooperative systems.




By using a CBA the efficiency rule is applied for public decision making. Due to this rule, a road safety measure like cooperative systems increases welfare, if benefits outweigh the costs of the systems. The relevant costs have to include also eventually arising costs of public supporting activities for diffusion of cooperative systems.

2.1.2. Needs for socio-economic assessment

Cooperative systems are introduced into the market by private industry. Their decision to offer to bring these systems to market depends on private business calculations that involve weighting costs against expected revenues. To increase market penetration of cooperative systems the automotive industry can use marketing instruments, such as rebates.

However, cooperative systems are a sort of “mixed good” in having private and public elements. Clearly, the car driver and his / her passengers will benefit from using cooperative systems by improving the driver ´s control of the car and the driving situations. Additionally, cooperative systems considered in SAFESPOT will detect and communicate safety relevant information to other vehicles. Thus, cooperative systems will reduce the occurrence of dangerous situations and possible collisions from which simultaneously other car drivers and pedestrians will benefit as well. Besides private benefits also public benefits arise which cannot be completely attributed to individual road-users. Regarding public benefits of cooperative systems supply and demand is insufficient since vehicle buying decisions are only driven by individual rationale.

Following economic reasoning from a society’s point of view supporting activities such as standards, informational measures, provision of infrastructure or obligatory implementation may be justified as a means of increasing market penetration of cooperative systems. Thus, it has to be proven whether society’s welfare will increase when market penetration of cooperative systems is increased. In this respect CBA can be used in order to proof profitability of cooperative systems from a society’s point of view.

2.1.3. Needs for an integrated assessment approach

Supply and demand concerning cooperative systems follow private decisions of users and vehicle manufacturers. Insurance companies, public authorities, road operators, telecommunication providers will also have an influence on these decisions. Consequently, the distribution of financial streams connected to the supply, use and market introduction of cooperative systems to the “stakeholders” has an impact on these private decisions. By integrating these distributional effects into a comprehensive evaluation system, central issues can be derived as to whether cooperative systems are valuable for society as a whole and for single stakeholder groups, such as users and car manufactures etc. Allocating costs and benefits according to the different stakeholder groups is called “stakeholder-analysis”.




When implementing cooperative systems, several stakeholders have to be considered, because each of them pursues its own specific interest which is affected by the stakeholder specific costs and benefits of the cooperative systems. In the following, of special interest will be the stakeholder group vehicle users and road operators:

• Drivers aim to achieve private profitability from the cooperative systems installed in the vehicles. It is essential for their purchase decision that the private benefits exceed the private costs. Users’ willingness to pay (for cooperative systems) derives from safer driving, increasing driving comfort, and savings in fuel consumption and so on.

• In the case of cooperative systems road operators (public or private), on the one hand, can be confronted with higher costs due to necessary infrastructure investments. On the other hand, providing cooperative systems can lead to higher revenues, greater reliability of traffic, enhanced traffic flow and increased transport safety. Hence, using the argument of creating improved transport performance, road operators could promote the use of their road infrastructure and private users can increase their revenues.

Thus, due to the demand for information by the public and stakeholders on the positive and negative effects of cooperative systems, an evaluation method is required that integrates cost-benefit and stakeholder analysis. The cost-benefit analysis represents the core methodology and gives a résumé of the efficiency of cooperative systems for society. By analysing stakeholders’ costs and benefits in detail, the calculation can be itemized. On the basis of such an itemized calculation, it can be viewed differently, as to whether profitability of cooperative systems is given for example to users as well as to the OEM and public authorities. In addition, it can be established, whether implementation and market penetration can be expected.

2.1.4. Methodological approaches

The analytical approach of the integrated assessment framework is to account for the objectives of society and stakeholders. For this purpose, three closely-related approaches are applied: cost-benefit-analysis (CBA), break-even analysis and financial assessment.

• The cost-benefit-analysis makes a statement about the efficiency of cooperative systems for society. The cooperative system is said to be efficient, if benefits outweigh costs. Benefits are the result of savings in accident and transport-related costs (vehicle operating costs, time savings) and of environmental benefits (savings of CO2 and pollutant emissions). In contrast, maintenance and operating of cooperative systems and investment in the cooperative systems and the complementary infrastructure generates costs.




• The interests of the end users are dealt with by a break-even-analysis. For the end users, the end market price of a particular system is the relevant price. The end user is interested in the mileage he has to drive per year such that benefits of the system accumulate in a bundle (e.g. additional comfort and safety) and outweigh the market price. The break-even analysis shows from which critical mileage on the system is profitable for the end-user.

• Revenues of road operators and earnings of public authorities from cooperative systems (in-vehicle systems and infrastructure) and connected costs are determined by using a financial assessment. For example private/public road operators can charge in return for investments in infrastructure user fees and the public authorities can acquire additional value-added taxes (VAT) from the purchase of cooperative systems and infrastructure components. By applying a financial analysis, the understanding of the economic interests of road operators and public authorities in cooperative systems will be improved.

2.2. Methodological progress and performance restrictions

2.2.1. Areas of methodological progress

As outlined above, the SAFESPOT BLADE socio-economic impact assessment has been extended to an integrated assessment which makes use of the methodological base of preceding studies. The specific research questions of SAFESPOT call however for a further adaptation of the assessment framework. Several issues are worth mentioning (see also Figure 2): the impact assessment of system bundles, the consideration of different business and service models, the explicit distinction between V2V and V2I communication applications and on the cost side the inclusion of economies of scale.

• The distinction between V2V and V2I communication applications is important for the cost-benefit analysis because the attainable benefits depend on the communication strategy of the system application. For V2V communication, a critical mass within the vehicle fleet has to be reached in order to generate benefits at all. Obviously, using V2V communication the equipped vehicles have to meet other equipped vehicles in order to transmit the data. Thus, the higher the penetration rate of equipped vehicles the higher the probability of transmitting any relevant data. On the other hand, a critical mass of equipped vehicles is not necessary for V2I applications. Once the infra-structure has been equipped with communication tools each equipped vehicle can make use of the application independently from the vehicle penetration.

• SAFESPOT BLADE pays special attention to different business and service models. The basic distinction between business and service




models can be found in charging. Business models in SAFESPOT do not involve user charges, whereas service models can do. A second distinction can be made whether the complete system is ready to use (one-stop service) or integrates various types of information from different providers. The socio-economic perspective of a fixed system bundle will be sequentially explored in different deployment environments. The difference between the business and service models will become apparent where there are different penetration rates (implying different numbers of equipped vehicles) and different equipment costs.

• Socio-economic assessment within BLADE looks at penetration rates which are derived from the assessment of market potential (see chapter 5.1). Other SPs of SAFESPOT use different penetration rates, because they evaluate the technical feasibility of concepts (e. g. SP4). Different penetration rates are therefore not a matter of inconsistency but follow the research purpose.

• Typically, the impact assessment looks at stand-alone systems but bundling of systems will become more common in the near future. As road safety will continuously improve over time, stand-alone solutions will find it more difficult to avoid accidents which are not yet targeted by other systems. This means for deployment that building block components such as ABS/ESC, optical sensors and communication devices form the technological platform for deployment of further developments. A reasonable bundle is characterised by cost synergies on the one hand and ideally independent safety impact channels on the other hand (different accident situations targeted). In addition, a combination of applications can have a higher safety potential than the accordant stand-alone applications [WESTERKAMP 2009].

• On the cost side, typically average costs go down for higher production of units (economies of scale). The process of the cost estimation in BLADE involved several steps, including system specification, initial estimation of unit costs based on publicly available information and refinement of initial estimations. These estimations were verified in a cost estimation workshop organised by the BLADE subproject involving also experts from the technical subprojects in order to align the BLADE estimations to the SAFESPOT technical developments. In this workshop different component costs have been estimated, depending on market penetration. As a result, the economies of scale could be determined. The results verify the bandwidth for economies of scale which can be found in literature [US DOT 1999].

• A spin-off of the assessment is a formula which allows – independent from distinct market penetration rates – the quick and user-friendly calculation of the BCRs for any considerable penetration rate.




2.2.2. Performance restrictions

The assessment of the SAFESPOT systems by WP6.5 had to respect some limitations caused by the project goals and the set up of the SAFESPOT project:

1. The assessment concentrates on two systems which are based on the two “pure” technological concepts of cooperative systems. The first system is based on V2V communication using sensing technologies and ad hoc dynamic networks for information sharing between vehicles. The other system is based on V2I communication by using RSUs and vehicles as nodes for communication sharing. Of course, a lot of further hybrid and mixed solutions using both V2V and V2I are possible. To start thinking about the optimal combination of V2V and V2I solutions a clear understanding and economic assessment of the extreme solutions is necessary. Considering the large number of possible types of the SAFESPOT system, this approach allows to find evidence for the question, on which concept (V2V, V2I) the SAFESPOT system should be based in order to start a feasible deployment. Additionally, the approach meets restrictions of the IP regarding time schedule and budget.

14SAFESPOT SP 6 BLADE, Y1 reviewMarch 27th, Brussels

SMA for BLADE analysis

• SMA functionality is crucial input for BLADE tasks T6.3,T6.4,T6.5,T6.6• Criteria: intersection, highway – local road, equaly devided over V2I and V2V• TNO/CRF proposal adopted end of December • BLADE analysis based upon SMA containing 6 functions:

Road intersection

safety function

Road condition

status:

Slippery road

Speed limitation

and safety distance

Safe speed

and manoeuvre to

approach the intersection warning

Local danger

warning:

Road surface status

Speed alert

V2I scenario (SP5) V2V scenario (SP4)

Figure 4: Description of applications considered in SAFESPOT-BLADE

2. The assessment of the SAFESPOT systems is based on a specification of the systems done by the technical subprojects. But this specification is in some respect preliminary. The work in BLADE and in the technical SPs had to be simultaneously performed so that modifications in the specification resulting from the technical development of the enabling




technologies as well as the testsite results could not be taken up by BLADE.

3. The provision of data identifying the system costs raises a problem concerning confidentiality from the OEMs’ point of view. In order to avoid a generic approach for cost estimation an iterative process was used for providing and validating the cost of the SAFESPOT systems in WP6.5. The process has been agreed upon with experts from the subprojects responsible for the technical specification of the SAFESPOT systems. Based on the specification of the applications the components of the SAFESPOT systems have been fixed in the first step. Then, a first cost estimation based on cost data available from former studies have been performed by WP6.5. In the next steps the cost data have been refined by the involved technical experts and thus adapted to the SAFESPOT systems. Finally, the results have been validated in a joined workshop with the technical experts. Thus the cost of the system was based on the SAFESPOT concepts developed in SAFESPOT, and not on a generic approach.

2.3. Cost-benefit analysis (CBA)

2.3.1. General methodology of CBA

The CBA compares the potential economic benefits of a road safety measure, for example, with all relevant consumption in resources due to the implementation of the road safety measure. All the benefits and the costs are measured in monetary terms by multiplying the physical impact units with the accordant cost-unit rates. The CBA can be used to assess absolute efficiency and relative efficiency. A road safety measure is (absolute) efficient if implementation produces positive net benefits. Are different road safety measures compared to reach a safety goal CBA can provide a ranking according to the level of net benefits generated by the different measures (relative efficiency). The most common indicator of the CBA which is also used here is the BCR. The BCR shows which benefit (measured in EUR) can be achieved by investing one EUR.

CBA is based on standard welfare economics, which is a branch of economic theory mainly considering optimal allocation of resources to increase the welfare of society. From a very theoretical point of view this approach evaluates economic effects due to public actions such as road safety measures by measuring variations in the use of resources. The costs of the measure in question are then confronted with its possible productive effect for the overall economy. To determine the benefits usually a “cost-saving approach” is used such that the benefits are given in terms of saved resources. In reality, however, benefits and costs of technologies are often unequally distributed such that winners and losers arise. Following the standard approach of CBA, public action is justified as efficient if benefits, in principal, could be used at least to compensate the “losers” [BOARDMAN et al. 1996].




The aim of the cooperative systems which are considered in SAFESPOT is to reduce the number of road fatalities. Thus, the evaluation of accidents and their resultant costs savings plays an important role. The potential of cooperative systems is avoiding accidents, avoiding fatalities, and avoiding injured. Avoiding accidents, respectively achieving mitigation of consequences of accidents represents the direct benefits of road safety technologies. Beneath these benefits further benefits are linked to the cooperative systems and have to be taken into account. Due to the avoidance of accidents, congestion is reduced. Linked to this consumption of fuel can be reduced, leading to fewer emissions. Another positive impact of cooperative systems can be the harmonizing of the traffic flow which also leads to a reduction of fuel consumption and to fewer emissions. Hence, relevant resource savings are:

• Time use

• Energy consumption (fuel)

• Vehicle operating and maintenance costs

• Greenhouse gas emissions (measured in carbon dioxide CO2)

• Emissions of various air pollutants (measured in nitrogen oxide equivalents NOX).

2.3.2. CBA process

The CBA process consists of four steps, which can be characterised as follows:

In the first step the relevant alternatives that will be compared within the analysis have to be defined. The relevant alternatives are defined as follows:

• The “with-case”, which means that a road safety technology will be introduced.

• The “without-case”, which assumes that there is no implementation of the technology to be evaluated (reference case).

Within the second step the potential impacts have to be quantified. Conceptually, the main effect of road safety technologies is the reduction of hazardous situations which affect the number and/or the severity of accidents. As a consequence, accident costs can be lowered. The other impacts are a reduction of congestion due to accident avoidance, direct traffic effects due to harmonising the traffic flow and prevented emissions.

Within the third step of the CBA process, the benefits are calculated in monetary terms by valuing the annual physical effects with standardised cost-unit rates. In addition to the monetary valuation of the physical benefits, the costs of the technology have to be determined. The costs of technologies for road safety improvements consist of investment costs for the implementation of the technology (vehicles and infrastructure) as well as operating costs and maintenance costs.




The fourth and last step is comparing economic benefits and costs. For this comparison several measures can be calculated: the most common is the BCR. Generally, the technology is profitable from a socio-economic point of view, if the calculated ratio is greater than 1.

t

tt

C

BBCR = , with

BCR Benefit-cost ratio

t Time horizon defined

Bt Estimated value of benefits for the year t [EUR]

Ct Estimated value of costs for the year t [EUR].

The value of the ratio indicates whether the implementation of the technology is favourable from a socio-economic point of view. A BCR of more than “1” indicates that benefits exceed costs. Thus, the introduction of the technology would be beneficial to society. The BCR expresses the absolute profitability of the technology which can be interpreted as the socio-economic return for every monetary unit invested in the implementation of the technology. For example, a BCR of “4.0” would show that 4.00 EUR can be gained for society for every EUR invested in the technology evaluated. Setting absolute, monetised values of benefits and costs into relation, the BCR is an indicator of efficient resource allocation.

The BCR of the cooperative systems being considered is most important for every decision maker interested in the evaluation of the cooperative system prior to deciding on market introduction, deployment or promotion of the cooperative safety system. That is why the results should be presented in a way that is both comprehensive and coherent. As a consequence, ranges of BCR are given which illustrate the variance of evaluation results. In this context, classes for CBA results are introduced to provide a grading of the results. The following classes are used [ASSING et al. 2006]:

• 0 < BCR < 1: The BCR is rated “poor” showing that socio-economic inefficiency of the cooperative safety system is given.

• 1 ≤ BCR < 3: The BCR is rated “acceptable” meaning that the social benefits associated with the implementation of a safety system exceeds the costs up to almost three-times which can be labelled as an acceptable absolute efficiency.

• BCR ≥ 3: The BCR is at least as high as “3” indicating an “excellent” result of the socio-economic assessment. The system evaluated as excellent should be in first line for market deployment.




2.3.3. Scope of CBA

This section comments briefly on the major methodological choices for the assessment.

• Lifecycle vs. snapshot CBA: snapshot CBA

Cost-benefit analyses can produce different summary measures of performance. On the one hand it is common to calculate the Net Present Value (NPV) by summing up all discounted values of benefits (plus sign) and costs (minus sign) over the lifecycle of the measure. On the other hand it is also common to preselect one or several target years and to calculate snapshot benefit-cost ratios (BCR) for these target years. In the second case, the costs will be transformed to annual values (using the discount rate) and will be compared to the target year benefits.

Both ways are feasible and represent good practice. Which way is selected depends on information needs and to some extent also on “evaluation culture”. Whereas transport appraisal guidelines in the United Kingdom (e.g. WebTAG) prefer the lifecycle analysis, the German guidelines for infrastructure investment planning prefer the snapshot method. When study clients are interested in detailed information on the timeline of market success of a measure the lifecycle analysis has its merits. Since the goal of economic assessment within BLADE is to assess the profitability of the V2V and V2I based SAFESPOT system in general a snapshot CBA analysis is appropriate.

• Time horizon: Target year 2020

Since cooperative systems are not yet or only to a little degree on the market, the time horizon for the socio-economic assessment has to be chosen. In a first step, the target year 2020 was envisaged, because 2020 is not too close after the market introduction which is expected for 2015. Additionally, the selection of the target year has to meet the requirements of forecasting cost-unit rates and accident data. Therefore, in the second step, the target year 2020 could be fixed, as we decided to follow the approach used by eImpact which showed a process how to bridge the gap of missing data in the European accident database and to forecast data for the target year of assessment.

• Geographical scope: EU-25

The geographical scope of the assessment of the SAFESPOT system is EU-25. The geographical scope depends on the availability of reliable figures for accident and traffic data. The most promising approach with regard to available data is the one which is used in the SEiSS study [ABELE et al. 2005] and in eIMPACT [ASSING et al. 2006]. This data cover the EU-25 which therefore has been considered as the geographical scope of SAFESPOT (see also chapter 4.2)

• Scope of damages: Focus on accidents with personal damages




The assessment covers personal damages costs of casualties. Property-damages are also included if they are connected to accidents with casualties. Thus, property-damage-only accidents are not part of the assessment (see Chap. 2.3.4.1). In addition, the assessment includes the cost savings of accident related congestion.

• Target population: Only cars and trucks, no powered two-wheelers

The accident target population which is addressed by the SAFESPOT system are accidents caused by a vehicle with at least four wheels (passenger cars, lorries, heavy goods vehicles, buses/coaches) which had an initial collision with another vehicle or a pedestrian. Accident data (as well as traffic data) for motorcycles are not available on EU-25 level. Thus accidents initially caused by 2-wheelers (mopeds, motorcycles) are not considered (see Chap. 4.2.1.1).

2.3.4. Benefits

The benefits are the monetary value of the physical impacts of the cooperative systems. The physical impact is, for example, a reduction of the number of fatalities or a reduction of fuel consumption. Each of these physical impacts has to have a monetary value attributed to it. This is done by using market prices, if available. Where no market prices are available, the cost-unit rates for the physical impacts have to be estimated. For this estimation two approaches are employed: the objective approach and the subjective approach. The subjective approach is based on individuals; the objective approach is based on society. The cost-benefit analysis is an analysis tool which considers the interests of society. Thus, for the cost-benefit analysis, the objective approach is the relevant one. The evaluation methods are displayed in Figure 5.

Figure 5: Evaluation methods




The objective assessment approach is differentiated into the cost-of-damage approach and the cost-of-avoidance approach. The cost-of-damage approach estimates the cost-unit rate for repairing the damage the physical impact produced. The cost-of-avoidance approach estimates the costs which are necessary to avoid the physical impact. Thus, the cost-of-damage approach can be used if the negative effects for the society of the physical impact can be compensated. The costs for compensating the effects are the relevant cost-unit rates. The cost-of-damage approach can be used for all physical impacts which are reduced due to using the cooperative systems.

The cost-of-avoidance approach estimates the costs of avoiding the physical impacts. This approach has to be used if the damage cannot be repaired or the damage cannot be financially estimated. In these cases a cost-unit rate which is based on the cost-of-damage approach cannot be calculated.

If both approaches are possible the lower cost-unit rate is taken. Generally, the cost-of-damage approach delivers the lower cost-unit rates.

The following two classes of benefits and cost-unit-rates are divided as to safety and traffic benefits.

2.3.4.1. Safety benefits

The safety benefit of a transport safety system is the potential of avoiding accidents, fatalities, severely injured, and slightly injured. If the accident cannot be avoided, it should at least be possible to reduce the accident severity. This means that a former fatality becomes a severely or even a slightly injured.

The cost-unit rates used here include personal damage, property damage and congestion related to accidents (see Table 1).

Table 1: Cost-unit rates for fatalities and injured used in SAFESPOT

Cost-unit rate 2020

Fatality 1.63 Mill. EUR / Casualty

Injured 0.06 Mill. EUR / Casualty

The values shown in Table 1 are based on a proposal of the European Commission [EC (2003)] for personal damage costs (for fatalities 1,000,000 EUR, for seriously injured persons 135,000 EUR and for slightly injured persons 15,000 EUR). The cost-unit rates reflect an average productivity growth of 2.4 % in the EU until the target year 2020 [EC (2008)]. In addition the values proposed by the EC are scaled up to prices of the year 2009 using an inflation rate of 2 % which is considered as price stability by the European Central Bank [SCHELLER (2006)].

In addition to personal damage costs also property damages and congestion costs have to be considered. In eIMPACT the cost-unit rates for property damages are assumed at 12,000 EUR for fatalities and 3,500 EUR for injured (severe and slight).




Crashes on motorways are regularly accomplished by congestion. Congestions lead to time losses, higher fuel consumption, higher air pollution and CO2-emissions. Therefore, these effects have to be considered as additional accident costs. While among the effects time losses dominate the result by far, all congestion effects are represented in a single cost-unit rate of 15,500 EUR for fatalities and 5,000 EUR for injuries.

2.3.4.2. Traffic and environmental benefits

The traffic benefits are saved travel time, saved fuel consumption, saved CO2 emissions and saved NOx-equivalent emissions. With exception of CO2, the cost-unit rates we have used are from eIMPACT which are based on 2005. The cost-unit rate for CO2 emissions is updated. The German Federal Environment Agency suggests a cost-unit rate for CO2 of 70 EUR per ton [ERDMENGER et al. 2007]. All cost-unit rates are compounded to the year 2020.

Table 2: Cost-unit rates for the traffic impacts

Cost-unit rate (year) Cost-unit rate (2020)

Travel time cost

Freight transport 22.33 EUR/t (2005) 30.06 EUR/hour

Passenger transport 14.84 EUR/t (2005) 19.97 EUR/hour

Fuel price

Petrol 288 EUR/t (2005) 387.61 EUR/t

Diesel 368 EUR/t (2005) 495.27 EUR/t

NOX-equivalent

Non-urban roads 1050 EUR/t (2005) 1,413.16 EUR/t

Urban roads 3150 EUR/t (2005) 4,239.49 EUR/t

C02-emission cost 70 EUR/t (2007) 90.55 EUR/t

Using the same cost-unit rates as eIMPACT has the advantage of comparability with the results as in eIMPACT where similar cooperative safety systems are considered as in SAFESPOT.

2.3.5. Costs

The cost estimation within the BLADE subproject has followed a well defined procedure. Several steps have been performed in order to come up with cost estimations for the in-vehicle as well as the infrastructure components (see chapter 2.2.2).

The system costs generally consist of two elements:

• investment costs, which appear only once in a lifetime of a system,

• operating and maintenance costs, which appear several times during the lifetime (e. g. every year).




In the CBA performed in this study the year 2020 is considered. Thus, the investment costs have to be annualised. This means that the costs are distributed over the lifetime of the system. For annualising the costs, information about the discount rate d and the lifetime n is necessary. Given these data, the annuity rate can be determined as follows:

1)1(

)1(*

−+

+=

n

n

d

ddAR , with

AR Annuity rate

d Discount rate

n Lifetime

Thus, the system costs per cooperative system and per year are determined by multiplying the investment costs by the annuity rate, and adding the operating and maintenance costs.

An important determinant of the annuity rate is the lifetime (life expectancy) of the system. In reality there are differences in life expectancies of the various elements of the system like infrastructure, vehicle and on-board units (OBUs), which are illustrated in Figure 6. In general, life expectancy for physical infrastructure components is higher than that of vehicles. On the other hand, communication infrastructure has a lifetime which is suggested even shorter than the lifetime of vehicles. In our calculations we use a common life expectancy for all in-vehicle system components. The same holds true for the infrastructure components (see chapter 7 for details).

Figure 6: Life expectancy of different elements of IVSS (own figure)




2.3.6. Benefit-Cost results

There are different ways to present results of benefit-cost analysis. The most common approach is the BCR. For calculating the BCR the benefits have to be divided by the costs. The benefits in formal terms are given by:

tDT_IVSItFtt DT*c*Ic*FB 1++= , with

B Benefits [EUR]

t Considered year t (2020)

cF, cI Cost-unit rate for fatalities/ injured [EUR]

F Avoided fatalities

I Avoided injured

1IVS_DT Function is 1 if the considered cooperative system has traffic impacts, otherwise 0

DT Direct traffic impact.

The first term of the equation handles the avoided fatalities. A monetary value is attributed to the number of avoided fatalities by multiplying them with the cost-unit rate for fatalities. The second summand deals with the injured. The third summand treats the direct traffic impacts. This summand consists of an indicator function and the direct traffic impacts (see also chapter 6.2). The indicator function is 1 for the cooperative systems which have direct traffic impacts. It is 0 for all other cooperative systems which have no direct traffic impacts. The benefits are determined for all scenarios. The benefits are the numerator of the BCR.

The benefits have to be divided by the total costs. For the in-vehicle system the total costs are determined by multiplying the vehicle stock with the fleet penetration rate and with the annualised costs per system. This is done for all business and service models. In formal terms the costs are given by the function:

tttt CS*FP*VSC = , with

C Total costs [EUR]

t Considered year (2020)

VS Vehicle stock

FP Fleet penetration rate [%]

CS Annualised costs per system (discount rate: 3 %, lifetime: 12 years).

For the infrastructure based system the in-vehicle costs are calculated in the same way as described above. The costs of infrastructure components have also to be included to get the total costs of the infrastructure based system. Obviously, the costs of the infrastructure components do not vary with fleet penetration rate.




Besides using total costs as an input for cost-benefit analysis, this figure can deliver further important information:

• Of course, total costs are of relevance since the decision-maker´s budget for road safety will be limited.

• A rate “costs per avoided fatality” can provide information about the cost-effectiveness of the considered cooperative systems. If the decision-maker´s objective is to reduce fatalities utilising the lowest costs possible, this rate can show the most economical of different cooperative systems with regard to prevented fatalities: The less this rate is the more cost-effective the cooperative systems are.

In the following the CBA will be used to draw conclusions about which technical solution of cooperative systems is more efficient – the vehicle to vehicle communication or the vehicle to infrastructure communication. This will be answered for each possible SAFESPOT bundle. By comparing the BCRs of V2V and V2I we may also get some hints about the optimal sequence in the deployment of a V2V or V2I solution or about an optimal combination of a V2V and V2I solution.

2.4. Stakeholder financial analysis

In addition to the cost-benefit-analysis, which evaluates the benefits and costs of IVS on a national or European level, the stakeholder analysis aims to split up the results onto the particular stakeholder groups. The following list shows a brief overview of the relevant stakeholders in the implementation process of cooperative systems:

• User

• Public authorities

• Road operator

• Car manufacturer

• Automotive supplier

• Insurance company

• Infrastructure supplier

• Telecom industry

• Content provider

• Service provider

• Automobile club

The roles of the stakeholders differ not only in the implementation process because of their input in the value chain of cooperative systems but also due to their influence on the market penetration level of cooperative systems.




As illustrated in Figure 7, three different impact levels can be distinguished. These three impact levels vary in the potential to influence the market introduction and penetration of a certain cooperative safety system. Therefore, level three represents the one which includes stakeholders who are not necessarily needed to provide a functioning cooperative system. Nevertheless, automobile clubs can enhance the market penetration of cooperative systems through their media work and awareness campaigns as well as due to demonstration and testing of specific devices. Also, insurance companies are able to provide financial incentives for potential users by reducing the insurance premiums for equipped vehicles.

On the second impact level several important stakeholders are included which guarantee a proper implementation and operating of cooperative systems. A couple of the stakeholders located here are responsible for the proper equipment and operation of the infrastructure (infrastructure supplier, road management, telecom industry). Next to the provision of the infrastructure, institutions are necessary which process and supply the data or information for the users/drivers (service provider, content provider). The automotive suppliers are located in-between the second and the first level as they are very important for the development and production of cooperative devices but lack a direct contact to the end customer. Research institutions can provide valuable information about the effectiveness of safety devices as well as identify their profitability on a macro- and micro-economic level.

Finally, the stakeholders who can have the biggest impact on the market penetration of cooperative systems are public authorities, car manufacturers and users (level one). Public authorities possess a variety of instruments and measures to influence the market environment of cooperative systems. On the one hand, they set the (legal) framework for the development, introduction and operation of cooperative systems, and on the other hand, they can decide on various enhancement initiatives (e.g. awareness campaigns, financial incentives, basic research etc.). Car manufacturers have also a strong impact on the penetration rate due to their direct contact with the customer. They heavily influence the purchase decision of the customers via their product diversification and equipment of new vehicles as well as via their marketing and pricing strategies. After all, the users or customers decide whether a system or product is successful. They express their willingness-to-pay and their overall interest in the products which are the main factors in the purchase decision. Therefore, the design and functions of a cooperative system have to address the customers’ needs and enable the user to identify his individual benefits.




Automotive Supplier

Cooperative Systems

Level One

Level Two

Level Three

Car Manu-facturer

Automobile

Club

Infra-structure

Supplier

Telecom Industry

Content Provider

Insurance Company

Service

Provider

User

Road

Operator

Public Authorities

Research Institutions

Automotive Supplier

Cooperative Systems

Level One

Level Two

Level Three

Car Manu-facturer

Automobile

Club

Infra-structure

Supplier

Telecom Industry

Content Provider

Insurance Company

Service

Provider

User

Road

Operator

Public Authorities

Research Institutions

Figure 7: Impact level of stakeholders

The stakeholder analysis in BLADE concentrates on the main stakeholders: users, public authorities and road operators. The car manufacturers are not considered in BLADE due to confidentiality reasons.

For calculating the profitability of cooperative systems for the three stakeholders two different approaches are used: A critical mileage is determined by a break-even analysis to show from which level of driven kilometres the technology is profitable for the end user. A financial analysis is performed to show which costs and revenues are connected to operating cooperative systems from the perspective of private and public road operators. Such a financial analysis is also used to show the stream of additional taxes going to public authorities caused by the selling of cooperative systems.

2.4.1. Break-even analysis

The break-even analysis is a method of business administration used to determine from which production output an investment is becoming profitable for the producer. Therefore, revenues and costs, depending on output, are compared to each other. Thereby it is assumed that benefits and end consumer prices are linear to the mileage. Then a critical output is determined where revenues equal costs called the “break-even-point” such that if output is lower, costs are higher than benefits (= losses), and if output is higher, revenues are higher than costs (= profits).




Here the break-even analysis is used in order to determine the potential profitability of cooperative systems from the point of view of the users. Benefits and end consumer prices of cooperative systems are, therefore, examined, depending upon passenger vehicle mileage per year. The result of the break-even analysis is expressed as the critical passenger vehicle kilometres for which costs equal benefits.

The user benefits can arise out of cost savings and additional benefits:

• savings regarding avoided accident costs which are not covered by insurance;

• savings through reduced fuel consumption and vehicle operating costs;

• benefits of increased driving comfort for users.

The user/consumer has to contrast these possible benefits with investment costs of the system bundle like the end-consumer price of the cooperative system and possible additional running costs. The considered system bundle is financed e.g. by credit taking (discount rate: 8 %). This discount rate differs from the discount rate used for the CBA.

To determine the benefit of the cooperative system, the cost-unit rates for the assessment of avoiding an accident are calculated using the willingness-to-pay approach. The willingness-to-pay approach, as a subjective method, questions how much the victim of an accident will pay to be able to avoid the accident, or what compensation amount will be accepted by the victim to accept the damage. Using this approach the benefits of cooperative systems are determined by a subjective valuation, in contrast, to the objective assessment of the cost-of-damage approach considering the economic losses caused by accidents.

Empirical research shows that cost-unit rates based on the willingness-to-pay approach for injured are only slightly above the cost-unit rates of the cost-of-damage approach. In contrast, for fatalities a substantial deviation exists: The cost-unit rate of the willingness-to-pay approach is more than three times higher than the cost-unit rate of the cost-of-damage approach [SEISS 2005].

As shown above the potential benefits of cooperative systems may include not only cost savings but also benefits of increased driving comfort. However, in this study the comfort issue is not considered. In this respect the willingness-to-pay for a cooperative system bundle is defined only by benefits generated by cost saving of accident costs and vehicle operating costs.

The benefit and cost components used in the break-even analysis are partly also present in the cost-benefit analysis. The difference is in the cost-benefit analysis the benefits and costs are included from a social point of view, while the break-even analysis considers the effective monetary savings and expenditure. This means, in particular in the break-even analysis, the flows of benefits and costs are considered including taxes (e.g. by a VAT), while in the cost-benefit analysis taxes are treated as transfer payments and thus are not considered.




The break-even analysis can also deliver some valuable information about the parameters which influence the profitability and attractiveness of cooperative systems from the point of view of the users. Therefore, it can help to detect and overcome barriers which prevent a proper market penetration rate. For instance, if a break-even analysis shows that after a short time or a low number of driven kilometres a cooperative system could be profitable for users, suitable measures for an increased awareness of the prospective benefits could be implemented.

2.4.2. Financial assessment for road operators and public authorities

For the V2I case for cooperative systems, the private or public road operator has to equip the infrastructure with roadside units including communication systems and sensors. For the road operator, costs of equipping the infrastructure arise. The road operator has to pay net market prices for equipping the network. To estimate the market prices of the infrastructure needed for cooperative systems the infrastructure costs are used and multiplied with a correction factor which reflects deviations of costs from market prices for example caused by taxes. The multiplication factor used here is 2.5.1 These costs are then assumed as the net market prices for equipping the infrastructure.

Equipping the infrastructure leads to high investment costs before the SAFESPOT system bundle can be brought onto the market. Given the service lifetime of the infrastructure these investment costs are annualised over the lifetime by using a discount rate. The discount rate shows the opportunity costs given by the possible return if instead of the planned investment the next best investment alternative would be realized. For the private road operator investing in the national electricity network could be considered as an alternative. Thus as discount rate 9.29 % is used which has been published as the maximum equity return of Germany’s national electricity network operators [BÖRSEN-ZEITUNG 2008]. For the public road operator a discount rate of 3 % is used. This discount rate is based on the average of the expected long-term productivity growth in Germany and is also applied in the macroeconomic evaluation methodology for the Federal Transport Infrastructure Plan 2003 in Germany.

Based on the annual infrastructure costs road operators can calculate user fees to generate revenues for using the services provided by the RSUs of the infrastructure. The level of user fees charged will depend on the prevailing and forecasted penetration rates within the vehicle fleet. Further the level of fees will depend on whether the infrastructure costs are partly or completely subsidised by public authorities.

A further important stakeholder considered here are public authorities. For the public authority an additional, annual financial burden can arise if investments in infrastructure for cooperative systems are made or if subsidies for implementing cooperative systems into the car are provided.

1 In an experts workshop the factor 3 was raised, since the value-added tax rate of 20 % has to be considered such that a factor of about 2.5 results.




Public support for the deployment of cooperative systems can be carried out in several ways differing in the intensity of an intervention provided and/or on the required earnings caused by the intervention. Launching of voluntary agreements, setting of standards and codes of practice, letters of commitment or memoranda of understanding shall have a lower intervention level and a relatively low impact on public budgets. In contrast, measures such as legislation, regulation, performance agreements and using economic instruments (e.g. tax incentives) have a relatively high impact on public expenditure and can also have a strong impact on markets of cooperative systems.

Public authorities should be especially interested in projects which have at least neutral fiscal budget effect. Therefore, the aim of the financial public budget analysis is to observe the financial impacts of supporting actions of cooperative systems on the earnings and expenditure of public authorities. Thereby the analysis covers monetary flows which are not considered in the cost-benefit analysis such as transfer payments and tax revenues.

Concerning tax earnings, it is assumed below that for each cooperative system that is sold, earnings for the public authorities arise in the form of VAT. A tax rate of 20 % of the end market price is assumed. Given this assumption, the forecasted earnings will be calculated for the year 2020. In this study we will postulate fiscal neutrality of public investments into the SAFESPOT system: The investment should be not higher than the expected tax revenues.

We will further assume that the on-board units of cooperative systems are not subsidized by public authorities. Only the roadside units in the V2I case shall be considered as the object of public investments and subsidies. If public authorities refrain completely from investments or providing financial support for roadside units, then only private road operators will provide investments in infrastructure and charge the users by way of fees to take part in financing the infrastructure.




3. BLADE oriented specification of the selected applications and dedicated accident scenarios

3.1. Introduction

In the following the applications of the SAFESPOT system considered for the socio-economic assessment are described. The applications have been defined by the technical subprojects of the SAFESPOT IP and selected in cooperation with partners of subproject BLADE. For the V2V based system and the V2I based system similar applications have been selected:

V2V – Vehicle based applications:

1. Lateral Collision – LATC: Road intersection safety 2. Road Departure – RODP: Road condition status/Slippery Road warning 3. Longitudinal Collision – LONC: Speed limitation and safe distance

V2I – Infrastructure based applications:

1. Co-operative Intersection Collision Prevention – IRIS: Basic application 2. Hazard and Incident Warning – H&IW: Reduced friction or visibility 3. Speed Alert – SpA: Legal speed limit

The corresponding vehicle and infrastructure based applications feature nearly the same use cases and services. Both the “V2V Road intersection safety” application and the “V2I Co-operative Intersection Collision Prevention Basic” application aim at preventing dangerous situations at intersections. The “V2V Road condition status – Slippery Road Warning” and the “V2I Reduced friction or visibility” application both provide safety relevant information to the driver. Safety relevant information concerning speed is provided by the “V2V Speed limitation and safe distance” application and the “V2I Legal speed limit” application.

The following presentation of the SAFESPOT applications obeys the methodological approach used in the SAFESPOT technical subprojects for differentiating between use cases and system requirements [see SAFESPOT SP8 WP8.4 (2008), p. 11-14]. The use cases state the goal of the applications and such the boundaries of the functional specification of the applications. But the use cases do not specify system requirements in terms of performance; these are under the domain of the system specifications.

The general Human Machine Interface (HMI) proposal for the SAFESPOT applications envisages the use of suitable earcons (audio sample) in order to draw attention to the warnings. Depending on the urgency, the earcon design of the warning will follow the three stages of “comfort”, “safety” and “critical”. The acoustic warning signal is able to capture the driver’s attention immediately, no matter where the driver is looking or what he/she is doing. Hence, the driver may react more quickly under certain conditions. In addition, visual information can be gathered from the central display of the vehicle/PTW. Optional visual attractors




and haptic elements through system intervention are proposed to make appropriate driver reactions quicker and more intuitive. The details of the SAFESPOT proposal for HMI are given in the SAFESPOT deliverable “Conceptualization of on-board information system and extended HMI” [SAFESPOT SP4 WP 3 (2008)].

Only a short description of the applications (purpose and scenario of the analysed application) and of the environmental aspects of the use case (i. e. typical accident situation) is provided below. The accident situations are described in detail in the SAFESPOT deliverable “Use case and typical accident situation” [see SAFESPOT SP4 WP2 (2006)]. For more details on the applications we refer to SAFESPOT deliverable “Use cases, functional specifications and safety margin” [SAFESPOT SP8 WP8.4 (2008)].

3.2. V2V applications

This section gives a BLADE oriented specification of the selected V2V based applications.

3.2.1. Lateral Collision – LATC: Road intersection safety

Purpose and scenario

The objective of the application is to avoid the risk of lateral collision by providing an early warning to the driver. The application consists of three sub-applications: Road intersection safety, Lane change manoeuvres in generic roads, and Safe overtaking.

BLADE focuses on the application Road intersection safety which addresses use cases at intersections:

1. Accident at intersections

2. Obstructed view at intersections

3. Permission denial to go-ahead

4. Defect traffic signs

5. Other vehicle brakes hard due to red light

6. Approaching Emergency Vehicle Warning

The expected frequency of use case 1 is stated as seldom. Thus case 1 is not considered in the following impact assessment. Two use cases make use of information provided from the infrastructure (traffic lights): use case 4 and 5. These are not purely V2V functionalities and, therefore, are not covered in the socio-economic assessment because of their mixed nature. Nor is use case 6 taken into account, which addresses emergency vehicles. The reason for its exclusion is that emergency vehicles are also omitted from the V2I counter-application. Of the three remaining use cases, only use case 2 and 3 (Permission




denial to go-ahead) are assessed under the application road intersection safety. This use case resembles a left turn assistant.

A

B

B

A

B

B

A

B

Figure 8: Examples of the scenarios for Road Intersection safety [SAFESPOT SP4 WP2 (2006),

p. 17]

For information on the concept of the HMI used see SAFESPOT deliverable “Conceptualization of on-board information system and extended HMI” [SAFESPOT SP4 WP 3 (2008), p. 47-50].

Table 3: Environmental aspects of the LATC use case

Road categories Traffic scenarios Vehicle categories Environmental conditions

Intersections in rural, urban, secondary roads

All kind of intersections

Trucks, motorcycles, cars

Bad visibility conditions (all kinds of weather)

3.2.2. Road Departure – RODP: Road condition status/ Slippery Road


The objective of this application is to share with other vehicles information about road status (e.g. a slippery road), bad weather conditions (e.g. ice, fog, etc.), or other risks – especially on roads bends – that may lead to the risk of a lane departure.

This application consists of two sub-applications: Road condition status/Slippery Road and Curve Warning. For example, given the sub-application “Road condition status/Slippery Road Warning”, a warning of a slippery road status or bad road conditions is broadcasted (Figure 9).

Figure 9: Example of scenario for Slippery Road Condition [SAFESPOT SP4 WP 3 (2008), p. 53]

For information on the HMI see SAFESPOT SP4 WP 3 (2008), p. 63, 64.




Table 4: Environmental aspects of the RODP use case


Any type of road Different kind of road conditions status

Trucks, motorcycles, cars

None in particular

3.2.3. Longitudinal Collision – LONC: Speed limitation and safe distance


The objective of this application is to inform the driver at an early stage about the potential risk of frontal or rear-end collisions, for instance, in the case of reduced speed of the preceding vehicles or, in the case of two-way roads, due to overtaking manoeuvres that the vehicles in the opposite traffic direction have started.

The application consists of four sub-applications: Head-on collision warning, Rear-end collision, Speed limitation and safety distance, and Frontal collision warning. The purpose of the sub-application “Speed limitation and safe distance”, which is selected for impact assessment, for example, is to provide information to forewarn the driver about the speed and the safety margin to maintain in case there are black spots and dangerous situations ahead, such as road works, static obstacles, or other factors that may limit or dynamically change speed and safety distance (Figure 9).

Figure 10: Example of scenario for Speed limitation and safe distance [SAFESPOT SP8 WP8.4 (2008), p. 33]

For information on the HMI see SAFESPOT SP4 WP 3 (2008), p. 59-61.

Table 5: Environmental aspects of the LONC use case


Motorway, urban roads, rural and secondary roads

Speed limitation and different kind of safety margin assistant

Trucks, motorcycles, cars2

None in particular

2 Trucks and cars are foreseen only for one of the three use cases [see SAFESPOT SP4 WP2 (2006), p. 51].




3.2.4. V2V applications bundle

In the SAFESPOT Sub-project SCOVA (SP4) the applications are bundled like a tree, as indicated in the following table:

Table 6: Clusters and Applications developed in SAFESPOT SP8 WP8.4 (2008), p. 26)

Application Cluster

Road Intersection Safety

Lateral Collision (LATC) Lane Change Manoeuvre

Safe Overtaking

Head On Collision Warning

Longitudinal Collision (LONC) Rear End Collision

Speed Limitation and Safety Distance

Frontal Collision Warning

Road Condition Status – Slippery Road Road Departure (RODP)

Curve Warning

Vulnerable Road User Detection and Accident Avoidance Vulnerable Road Users (VURU)

The branches of this tree are the use cases that have been detailed for each application in the deliverable “Use case and typical accident situation” [SAFESPOT SP4 WP2 (2006)]. The last cluster Vulnerable Road User (VURU) focusing on the propagation of information about a vulnerable user is not considered here since some parts of the application are developed outside SAFESPOT.

3.3. V2I applications

This section describes for impact assessment selected V2I based applications. For these applications, the algorithms to determine potentially dangerous situations are implemented at a roadside unit (RSU). At an RSU information is stored in a so-called Local Dynamic Map to build a representation of the road or intersection with dynamic data (e.g. floating car data). Based on this data, conflicts and violations are identified. By means of wireless communication (V2I communication) determined warnings are sent to the vehicles. The in-vehicle HMI is dealt with in deliverable “Conceptualization of on-board information system and extended HMI” [see SAFESPOT SP4 WP 3 (2008)].

3.3.1. Intelligent Cooperative Intersection Safety – IRIS: basic application


The objective of this application is to calculate and predict the trajectories of the road users in the proximity of urban intersections. Based on these trajectories, safety-critical situations will be identified and the decision to provide a warning will be made. The in-vehicle information provided to the driver is visual and acoustic




(haptic and tactile modalities are not taken into account). The objective of the selected sub-application – named the basic application - is

• to identify potential red light violators

• to support drivers turning right to be aware of pedestrians and cyclists

• to assist left turning vehicles without a separate signal stage.

The reader is referred to the deliverable “Specifications for Intelligent Cooperative Intersection Safety” [SAFESPOT SP5 WP3 T5.3.3 (2008)] for more details about the IRIS application.

Table 7: Environmental aspects of the IRIS use case


Urban, rural Different kind of intersection

All None in particular

3.3.2. Hazard and Incident Warning – H&IW: Reduced friction or visibility


The objective of this application is to warn drivers in case of a dangerous event on the road. It is assumed that the in-vehicle information provided to the driver is visual and acoustic. Three use cases are considered under the domain of this application: obstacles, wrong way driving, and abnormal road conditions. The selected events are the most relevant in terms of safety (e.g. accidents, presence of unexpected obstacles on the road, presence of a vehicle driving in the wrong direction, dangerous overtaking and also bad weather conditions like snow, rain or fog).

The sub-application which is selected for impact assessment is able to alert drivers to the presence of a hazard due to abnormal weather conditions (e.g. rain, ice or fog) which results in reduced friction or low visibility. In the case of reduced friction, the objective is to give a general warning about slippery road conditions and adjust braking distance for other H&IW sub-applications. In the case of reduced visibility, the objective is to warn traffic in the zone affected to reduce speed due to low visibility ahead.

The reader is referred to SAFESPOT deliverable “Specifications for Hazard & Incident Warning Applications” [SAFESPOT SP5 WP3 T5.3.2 (2008)] for more details about the H&WI application.

Table 8: Environmental aspects of the H&IW use case


Motorway, rural roads None in particular All Friction and reduced visibility




3.3.3. Speed Alert (SpA): Legal speed limit


This Speed-Alert application in general provides a recommended speed based on the weather status, road surface conditions, road typology, traffic flow speed and other events like road works, traffic jams and deviations. To be more specific, it uses information from other sub-applications, one of which is the Hazard & Incident Warning sub-application, as discussed above.

In the assessment the legal speed limit sub-application is also considered. This sub-application deals with two objectives: the first objective is to provide the driver with continuous knowledge about the legal speed limit; and the second is to allow the infrastructure manager to modify locally the legal speed limit, either temporary or permanently.

It is assumed that the in-vehicle information provided to the driver is visual and acoustic. The following table lists the situations for which the application is developed.

The reader is referred to GLASER et al. (2008) for more details.

Table 9: Environmental aspects of the SpA use case


Motorway, Rural Roads

None in particular All None in particular

3.3.4. V2I applications bundle

The V2I applications, identified previously, will be considered as a bundle. For facilitate this, interactions in the functionalities have to be considered. Since the Intelligent coopeRative Intersection Safety (IRIS) application will function independently of the other V2I applications, the impacts of this application will be added to the impacts of the rest of the bundle. In the contrary, there is a strong dependency between the Hazard & Incident Warning application (H&IW) and the Speed Alert application (SpA): The Hazard and Incident Warning application provides information via the Legal Speed Limit application to the driver. Therefore, only the functionalities of the Speed Alert and IRIS application will be considered for determining the overall impact of the V2I bundle (i.e. the effects of the H&IW application are not included for the bundle to prevent double counting of impacts).




4. General accident and traffic data compilation

According to the methodology of the socio-economic assessment accident and traffic data are needed for the impact estimation and the calculation of the safety benefits. In the following the process of generating the general accident and traffic data is described and the relevant compiled data is shown.

In general, the data which is needed for socio-economic assessment has to fulfil requirements of the applications and the impact estimation process. To get the required data, therefore, the following steps have to be done:

• Synchronisation of categories both used for the description of the SAFESPOT applications and the compilation of accident data;

• Synchronisation of categories both used for the description of impacts (safety, traffic efficiency, fuel consumption) and data compilation;

• Final specification of accident data and traffic data needs for socio-economic assessment;

• Compilation of accident and traffic data for the reference year and the geographical regions considered.

All this data will then be forecasted to 2020 which is the target year of the assessment.

4.1. Approach

The specified SAFESPOT applications (see chapter 3) will help to prevent accidents and/or mitigate the severity of collisions. With regard to the safety impacts the accident data have to be synchronized with the systems´ scope, i.e. the accident types addressed by the systems, and the data requirements of the safety impact estimation. Thus, relevant data categories are obtained for which the accident data have to describe:

• the number of accidents,

• the number of fatalities, and

• the number of severely and slightly injured (distribution of accident severity).

Avoiding or mitigating accidents can reduce traffic congestions, too. Thus, the considered applications can increase the efficiency of the road network. Moreover, the systems will homogenize the traffic flow by reducing the speed variances or by fitting the distances to the vehicle in front to a certain level. The homogenization of traffic flow will affect vehicle throughput and road capacity. Reducing the number of congestions and unsteady driving will also have an impact on fuel consumption and emissions of traffic. For the estimation of the effects on traffic flow and velocity, traffic data is needed to quantify the effects.




Traffic data particularly have to provide information on:

• vehicle mileage,

• distribution of vehicle mileage for the road types (motorway, rural, urban),

• vehicle stock,

• distribution of vehicle categories within vehicle stock, and

• traffic density and speed patterns.

The categories used for the compilation of this traffic data have to be synchronized with the categories which are used for the specification of the applications:

• Road categories;

• Collision type;

• Traffic scenarios (e.g. intersection);

• Vehicle categories;

• Environmental conditions (weather, lighting).

4.2. Accident data

For the calculation of the safety impacts we have to specify the accident data on which the impact estimation will be based. The accident categories used for the description of the applications and the safety impacts have to match the accident classification of the accident data base. To get an adequate classification of accidents the structure of accident data of the CARE database (Community database on Accidents and Roads in Europe) was used. Since the CARE data base is limited additional data were compiled to get actual data on the needed level of detail for the EU-25.

The reference year of the accident data compiled is 2005. But the SAFESPOT system is a future system and the target year of assessment is 2020. Therefore, a forecast on the development of the accident data is needed which takes the target year into account.

The collection and compilation of accident data mainly follows the procedure used in the project “Socio-economic Impact Assessment of Stand-alone and Co-operative IVSS in Europe" (eIMPACT). The eIMPACT project assessed the socio-economic effects of twelve IVSS and their impact on traffic, safety and environment. Here, information was shared with TRACE (Traffic Accident Causation in Europe), a parallel worked Strategic Targeted Research Project (STREP) on accident causation, and with members of the EUCAR working group on safety.

With respect to general accident data about fatalities and injuries the accident statistics of EUROSTAT were used.




4.2.1. Data compilation and synchronisation

4.2.1.1. Data sources and data categorization

The disaggregation of the level of accident data is determined by the methodology which is used for the safety impact assessment. The determination of safety impacts follows mechanisms which take the behaviour of the drivers and the resulting effects on fatalities and injured into account. Given this mechanisms the safety impacts on accidents will be analyzed in detail. For this classification of accidents the structure of the CARE database was used. The environmental and situation-based factors of the accident classification are:

• Road class (motorway, rural road, urban road);

• Vehicle type (car, truck);

• Collision type (animal, chain or rear, frontal, lateral, parked vehicle, pedestrian, single vehicle accident, single vehicle accident with obstacle, single vehicle accident no obstacle, other and unknown);

• Road surface conditions (normal (dry) and bad weather (fog or mist, rain, snow));

• Lighting conditions (dark and light);

• Location (junction, no junction).

This means that quite detailed data for the background variables and accident classes distinguished is needed for the EU-25 countries. However, the CARE database has several limitations:

• The CARE database is limited to EU-14 (EU-15 excluding Germany). In 2005 for the first time Estonia, Hungary, Malta and Poland provided data. But, data were needed for the EU-25;

• The level of completeness and relevance of the data varied between countries included in CARE;

• Several definitions of accident data for example the definition of seriously injured persons differs between the countries, so comparisons between member states are difficult;

• The collision type variable had been removed from the database.

To get these disaggregated data additional data was compiled by eIMPACT in order to meet the problems of a limited database and missing data. For this, an enquiry to national data bases was designed to match the data needs of the safety impact analysis with the available data. The enquiry was then sent to the respective countries via the TRACE project. But the enquiries to the national data bases could not solve the problem of different definitions of accident data between the EU-25 countries.




For efficiency reasons, the countries of the EU-25 were then grouped into three different country clusters with a similar level of road safety performance based on the number of fatalities in 2005. As a result the countries were grouped into three different clusters:

1. Cluster 1 (6 countries): Denmark, Finland, Germany, Sweden, the Netherlands, United Kingdom;

2. Cluster 2 (8 countries): Austria, Belgium, France, Ireland, Italy, Luxemburg, Malta3, Spain;

3. Cluster 3 (11 countries): Cyprus, Czech Republic, Estonia, Greece, Hungary, Latvia, Lithuania, Poland, Portugal, Slovakia, Slovenia.

Representative countries of each cluster were then asked to provide data (Cluster 1: Germany, United Kingdom; Cluster 2: France, Spain; Cluster 3: Czech Republic, Estonia, Greece). These countries were chosen because of their high share of fatalities within the clusters, but also for reasons of data availability and reliability. The data of these representative countries were then transferred to all other countries in the same cluster. With this approach, a high degree of accuracy was retained (see eIMPACT, Deliverable 4, p. 31-36).

The data are subject to the following constraints:

• The data cover the EU-25 countries. Reliable data from Romania and Bulgaria (entry to the EU in 2007) were not available.

• The base year for data analyses is 2005 since up-to-date data were not available from most of the member states.

• As in eIMPACT, vehicle types are included in the accident types, so no further discrimination between passenger cars and truck is used.

• The accident data available from eImpact does not include powered two-wheelers (mopeds and motorcycles). Exploitation of powered two-wheelers is only possible by means of estimation the amount of these vehicles in the data and creating a similar dataset for all powered two-wheelers in the EU-25. SAFESPOT partner Piaggio had been asked to check whether reliable and complete data are available. The feedback by Piaggio showed that there is only little data which might be helpful for the socio-economic assessment. Therefore, it was agreed with the project coordinator not to take powered two-wheelers into account.

• The data should include all relevant background variables about accident types given the specified applications (e.g. collision type variables).

Table 10 shows the resulting variables by which injury accidents showing fatalities, seriously or slightly injured should be analyzed to get estimates of safety impacts of the applications. 3 Road safety performance of Malta is relatively high such that it should be assigned to cluster 1. However, in EU-Project eIMPACT it was assigned to cluster 2 to get countries from the same region in one cluster [see WILMINK et al (2008)].




Table 10: Definition of variables for the data enquiry

Category Definition/ Explanation

Collision type

Collision on the road with pedestrian

Accident involving just one vehicle and at least one pedestrian.

Collision on the road with all other obstacles

Accident involving only one vehicle which collided with a fixed or moving obstacle on the road (obstacles including animals, parked vehicles, trams and trains, load, etc). Cars waiting to go ahead or to turn are not parked cars, they are in two vehicle accidents.

Collision besides the road with pedestrian or obstacle or other single vehicle accidents

All other single vehicle accidents involving one vehicle not included in 1 and 2.

Frontal collision Collision between two moving/waiting vehicles travelling in opposite directions. No pedestrian involved in accident.

Side-by-side collision Collision between two moving/waiting vehicles travelling in the same direction, side-by-side. No pedestrian involved in accident.

Angle collision Collision between two moving/waiting vehicles travelling in different directions, but not opposite. No pedestrian involved in accident (e.g. usually accidents at intersections).

Rear collision Collision between two moving/waiting vehicles, travelling in the same direction, one in front of the other. No pedestrian involved in accident.

Other accidents with two vehicles

Any accident between two vehicles not included in any of the categories above. No pedestrian involved in accident.

All other collisions Any accident that is not included in any of the categories above.

Road type

Urban roads (no motorway)

Motorway

Rural roads (no motorway)

Weather

Adverse Includes fog, mist, rain, snow, sleet, hail

Normal Includes dry, strong wind

Light conditions

Darkness

Daylight or twilight or unknown

Location

At intersection

No intersection


Motorway





4.2.1.2. Results of the data collection

Table 11 shows the resulting shares of fatalities and injured for different background categories in the EU-25 accident data, summarized over all accident classes (disaggregated results for the clusters 1-3 are shown in appendix). This table gives some insight into the potential for reduction of accidents. For instance, if a system only targets side-by-side collisions (collision type 5), the potential of the system in avoiding fatalities can never exceed 2 %, even at high penetration rates and high effectiveness of the system.

Table 11: One-dimensional distribution of background variables over all relevant accident classes, EU-25 (2005)

Injury accidents

Fatalities Seriously

injured Slightly injured

Totals 1,127,057 36,069 282,128 1,206,847

Collision type


11 % 13 % 13 % 8 %


6 % 7 % 6 % 6 %

Collision beside the road with pedestrian or obstacle or other single vehicle accidents

13 % 22 % 16 % 11 %

Frontal collision 8 % 18 % 14 % 9 %

Side-by-side collision 5 % 2 % 3 % 5 %

Angle collision 25 % 15 % 22 % 26 %

Rear collision 13 % 5 % 6 % 14 %

Other accidents with two vehicles

6 % 3 % 4 % 6 %

All other collisions 13 % 14 % 15 % 14 %

Road type


66 % 32 % 51 % 64 %

Motorway 5 % 7 % 6 % 5 %


29 % 61 % 43 % 30 %

Weather

Adverse 13 % 13 % 13 % 14 %

Normal 87 % 87 % 87 % 86 %

Light conditions

Darkness 26 % 39 % 30 % 26 %

Daylight or twilight or unknown

74 % 61 % 70 % 74 %

Location

At intersection 50 % 23 % 38 % 52 %

No intersection 50 % 77 % 62 % 48 %




4.2.1.3. Road safety forecast to year 2020

Since up-to-date forecasts of accidents and/or casualties for the target year 2020 were not available on EU-25 level, it was necessary to perform an own estimation of road safety indicators for the selected time horizon.

BLADE followed the approach used in eIMPACT where the road safety prediction (see Figure 11) was based on the future development of the fatality risk for each country cluster, i. e. the ratio between the total number of fatalities and the total vehicle-km driven. Data on vehicle-km were not available for some countries. In these cases the ratio between fatalities and the vehicle stock, being an indicator determining road safety, was calculated. Given these ratios time series of the annual number of fatalities and vehicle-km respectively vehicle stock per year for the period 1991 to 2005 were obtained for the 25 countries using the CARE and the IRTAD (International Road Traffic and Accident Database) databases. For each year of this period, the fatality risk was calculated. A time series analysis was carried out using exponential regression to extrapolate the trend of fatality rates to the year 2020. In a further step, the number of fatalities in each country cluster was calculated backwards by using values for vehicle-km or numbers of vehicles provided by recent forecasts for the target years.

Figure 11: Methodological Approach for road safety prediction for 2010 and 2020, EU-25

[WILMINK (2008), p. 37]

Being based on accident data for the period 1991 to 2005, the forecast approach takes into account the effects of all measures taken for the improvement of road safety at any level in the EU within this time period. Hence, it is presumed for the

Vehicle mileage and vehicle stock

per “country cluster”

1991 - 2005

Fatality rate per 1 mill. veh .-km / per 1 mill.

vehicles per “country cluster”

Extrapolation

2020

Number of fatalities per “country cluster”

Source: Div. Statistics Number of fatalities per “country cluster”

Source: EC, CARE Database

1991 - 2005

Own calculations

Own calculations

Forecast

2020

Vehicle mileage and vehicle stock

per “country cluster”

Source: ProgTrans AG, European Transport Report 2007/2008

1991 - 2005




road safety prediction that the already implemented measures will still be effective in the future – however, with decreasing effectiveness – and that no additional efforts are made to reduce the number of accidents and casualties.

The method for estimating the number of fatalities allows for fatality data being differentiated by the three country clusters which have been introduced for the use in the safety impact assessment. The results of the road safety prediction for 2020 are summarized in Table 12.

Table 12: Results of the road safety prediction for fatalities for 2020, EU-25 [EUROSTAT (2009); own calculations]

Fatalities in 2005

Fatalities in 2020 (estimated)

Index for 2020 (Year 2005 = 100)

Cluster 1 10,596 5,430 51.2477

Cluster 2 17,505 9,283 53.0334

Cluster 3 13,203 6,077 46.0291

Sum 41,304 20,791 50.3363

4.2.2. Data correction with regard to underreporting

By using the CARE database and national reports various problems can result because of differences in reporting practice between countries. This problem may result in underreporting of fatalities and injured.

Only minor differences in reporting practice may exist for counting fatalities where an injured person dies some time after the accident. There are more variations in the reporting threshold for seriously and slightly injured. The reporting for seriously injured differs according to a minimum time of hospitalization in the Member States and reporting of slightly injured varies according to whether hospitalization or medical treatment is necessary. These variations are known to lead to a large and variable under-counting of crashes and injured.

There are estimates for under-reporting of seriously and slightly injured only for Denmark, Germany, Netherlands, Norway, Sweden, Switzerland and the United Kingdom (see for example HEATCO (2005), ICF Consulting (2003). In HEATCO (2005) the application of correction factors is recommended. For countries for which correction factors are not available average values shall be calculated using the known results.

In a meeting of the BLADE partners it was discussed whether and how to correct for underreporting of accidents. The partners agreed on not implementing corrective measures for underreporting thus following the recommendation of BASt ´s statistical section. The reasons can be summarized as follows:

• One important practical reason is that only a few countries provide data on underreporting. Further, generally accepted specifications for correction factors in the EU do not currently exist.

• From a methodological point of view, using correction factors for clusters would be very imprecise, since there are differences between countries




regarding underreporting and the specification of accident data. Consequently, correction factors would have to be specific for each country. Moreover, underreporting is even different within countries, as it may depend on the accident type. They can also differ according to road and vehicle type.

• One important practical reason is that only few countries provide data on underreporting. Furthermore, there are not generally accepted specifications for correction factors in the EU.

• From a methodological point of view, using correction factors for clusters would be very imprecise, since there are differences between countries regarding underreporting and the specification of accident data. Consequently, correction factors would have to be specific for each country. Moreover, underreporting is even different within countries, as it may depend on the accident type. They also can differ according to road and vehicle type.

• If correction factors were defined, they would also have to consider overreporting (suicides). But only a limited number of data which is of very low accuracy is available.

• Concerning CBA, calculation of accident costs is mainly based on severe accidents with fatalities and severe body injuries. But underreporting predominantly occurs with slight injuries and accidents with vehicle damages only. So using correction factors would only have a very small effect on accident costs.

• Correction of underreporting seems to be quite difficult. By correcting for underreporting one would loose reliability and comparability between corrected and not corrected data, e.g. not corrected data from former studies. For interpretation of accident data one should have in mind that the data without correction factors describe the minimum level of accidents but with a maximum of precision.

4.3. Traffic-related data

The traffic data which are used in SAFESPOT are the same as used in the eIMPACT project [ASSING et al. 2006, BAUM et al. 2008]. The vehicle stock and the vehicle mileage are based on [ProgTrans (2007)], and the distribution of the level of services is based on [INFRAS/ IWW (2004]. The level of service concept gives information about the quality of the traffic on a road where “A” indicates a traffic situation in which every traffic participant can drive with his preferred velocity and “F” indicates congestion.

The vehicle stock is used for determining the costs for equipping a certain share of the vehicle fleet with the SAFESPOT bundle.




Table 13: Vehicle stock in EU 25 for the year 2020 [PROGTRANS (2007), own Calculations]

Transport type Mill. Vehicles

Passenger Transport 263.7

Goods Transport 42.2

Bus 0.8

Total 306.7

The vehicle mileage is relevant for determining the traffic and safety effects:

Table 14: (Estimated) Vehicle mileage for EU 25 for the year 2020 [PROGTRANS (2007), own calculations] Transport Type Bn. vehicle kilometres

Passenger Transport 3,028.2

Goods Transport 659.0

Bus 29.2

Total 3,716.4

The distribution of the mileage as well as the distribution of the level of services is relevant for calculating the traffic effects:

Table 15: (Estimated) Distribution of vehicle mileage for the year 2020 [INFRAS / IWW (2004), own Calculations] Motorway Rural Urban

2020 24.2% 52.4% 23.4% Table 16: (Estimated) Distribution of the level of services on the different road categories [INFRAS /IWW (2004), own calculations]

Road Type

Level of Service

A B C D E F

Motorways 51 % 15 % 10 % 8 % 4 % 12 %

Rural Roads

94 % 2 % 1 % 1 % 0 % 1 %

Urban Roads 99 % 2 %

Other data which are relevant for determining the traffic effects are the distribution of drive configuration:

Table 17: Distribution of drive configuration [BAUM et al. 2008] Passenger car diesel 36.00 %

Passenger car benzene 47.37 %

Goods Vehicle 16.63 % The last relevant traffic data which are needed are the infrastructure length. In the year 2020, the length of the infrastructure is estimated as 6.6 Mill. km. The length of motorways is 78,965 km. The length of the other road types cannot be estimated [ASSING et al. 2006].




5. Market assessment

In the following indication about the market potential of the SAFESPOT applications will be given. This will be done for different business and service models of market introduction of the V2V and the V2I SAFESPOTsystems. In particular, the following distinction between a Business and a Service Model is used. On the one hand, in a Business Model there is at least partial public financing: the SAFESPOT systems are paid completely by the public authority (not by the final user) or at least partially (with a contribution of the user). On the other hand, in the Service Model there is only private funding: the cost of the additional service is paid entirely by the user (direct payment).

We start with explaining the different models and show the corresponding estimated figures for penetration rates in the vehicle fleet in 2020. The estimation of the penetrations rates of the SAFESPOT cooperative system is based on an experts´ survey about the market potential of the cooperative systems in new vehicles. A summary of the survey results is shown below.

Policy instruments, such as a reduced road tax or motor insurance premiums or mandatory/ legislative action which may accelerate market introduction of SAFESPOT applications are not considered here. In addition, we assume that an “After Market” for the SAFESPOT systems does not exist. Thus, the SAFESPOT Business and Service Models considered here have been based on the assumption that the market penetration of the SAFESPOT systems is only market driven by substitution of “old” vehicles with “new” vehicles.

5.1. Business and Service Models4

In addition to the technical specification of the SAFESPOT applications and the benefit and comfort they will provide to users, the prevailing business model will have a strong influence on how fast the applications are disseminated into the vehicle fleet. In particular the Business and Service model will define how the costs and benefits of the applications are distributed between, especially, users, road operators and public authorities, which are considered as the main stakeholders in this study. 9 models have been formulated about this.

Three main criteria are established to structure the different models:

1. Basic business model (SAFESPOT base) and enlarged model (SAFESPOT Plus) with respect to additional service features of the SAFESPOT applications (commercial/ marketing point of view);

2. Business Model (BM) with complete or at least partial public financing and Service Model (SM) with complete private financing (financial point of view);

4 For more details about assumptions on stakeholders and pricing policy see the SAFESPOT delivery “Definition of alternative service and business models” [SP6 WP6.6 (2009)].




3. V2V configuration and V2I configuration or a combination of both configurations (technological point of view).

Table 18 gives an overview of the models that have been considered, based on the above-mentioned criteria. For each model, market penetration rates are estimated based also on the experts´ survey. These penetration rates corresponding to the nine models are shown below in table 19.

Table 18: Overview of the Business and Service Models

No. Model

Criteria

Pene-tration

rate

Business Type (Commercialisa

tion) Financing

V2X-(technologic

al) Basis

BM SM Public Public/ private

Private V2V V2I

1. Required SAFESPOT installation

X X X 6.7 %

2. SAFESPOT public investment

X X X 7.9 %

3. SAFESPOT installation X X X 8.7 %

4. Public subsidy X X X 9.5 %

5. Market driven SAFESPOT

X X X 4.2 %

6.

Market driven SAFESPOT infrastructure based

X X X 5.4 %

7.

Market driven SAFESPOT plus

X X X 6.1 %

8.

Market driven SAFESPOT plus infra-structure based

X X X 7.7 %

9.

Combined V2V / V2I solution

X X 11.2 %




The following cost-benefit analysis and stakeholder analysis will be based on the above penetration rates. A given penetration rate is characterized by a Business or Service model, a financing model and a technical solution (V2V or V2I). The following labels will be used in this study to define the different scenarios belonging to a penetration rate, given the V2V or V2I technical solution of a cooperative system:

Table 19: Estimated penetration rates for the Business and Service models

Models V2V V2I

1. Business model base – Public (1. + 2.) 6.7 % 7.9 %

2. Business model base – Public/Private (3. + 4.) 8.7 % 9.5 %

3. Service model base – Private (5. + 6.) 4.2 % 5.4 %

4. Service model plus – Private (7. + 8.) 6.1 % 7.7 %

5. Combined V2V and V2I based solution (9.) 11.2 % For the estimation of the fleet penetration rates in 2020 the approach from the eIMPACT project was used [see WILMINK (2008)]. There estimates for the penetration rate of new vehicles were concerted to fleet penetration rates in the year 2020. This was done assuming that the vehicle age distribution would remain the same in 2020 as in 2005. Fleet data from EU-25 countries were gathered, and in case of insufficient data, complementary International statistics were used or estimates were done based on neighbouring countries. For the estimates of the fleet penetration rates, the information of the first year of market introduction was needed.

5.2. Methodological Approach of the survey

For data generation the experts’ opinions methodology was used. The survey was directed to groups and individuals engaged in the field of cooperative systems and expected to have considerable influence on the future of those systems and services in order to have a valuable estimation. Thus, well informed people are selected and asked to assign a certain degree of importance and probability to several possible future events. To forecast market potential of the SAFESPOT system in new vehicle, a structured questionnaire was applied. The questionnaire addressed stakeholder typology, optimal cooperative system configuration and business drivers, market penetration rate, and deployment. In the questionnaire the experts were asked to indicate the possible SAFESPOT system penetration rates in new vehicle for the years 2015, 2020. The forecasts were done for different business and service models.

For data generation the questionnaires have been submitted to participants of the SAFESPOT and Watch-Over User Forum in Stuttgart (21st and 22nd January 2008), and to the SAFESPOT partners. In addition, 100 relevant international ITS experts were contacted by an online survey. Finally stakeholders, such as insurance companies, public authorities, road operators, telecom operators, service and content providers have also been contacted. As in the previous step




some categories were not represented properly by the sample. In sum, 90 responses to the questionnaire were collected.

The responses are statistically evaluated in the following way: Firstly, the average values of the forecast for the three years are calculated. Secondly, the response values are grouped together according to three clusters with percentiles 33 % and then average values for each cluster are calculated. The clusters could be interpreted as pessimistic, intermediate and optimistic forecasts of market potential of the SAFESPOT system in new vehicles.

The answers have been classified in steps from 20 to 20 (0-20 %, 20-40 %, 40-60 %, 60-80 %, 80-100 %). The outcome of these clusters shows a coherent and consistent trend (for more details about the sample and more forecasts see Annex 1).

Answers are obtained from 16 European countries. The countries most represented are Germany, Italy and the Netherlands. Figure 12 shows the distribution of stakeholders participating in the survey. The biggest group of respondents are scientists (35 %) followed by automotive industry and car manufacturers (20 %), public authorities (15 %) and insurers (9 %). In the following the average forecasts are shown as tables and figures for the above mentioned 5 Business and Service models.

Stakeholders' Categories

Automotive Club0%

Supplier Automotive Industry

8%

Infrastructure System Supplier

1%

Insurance9%

Telecommunication Industry7%

Academia/University11%

Research Centre24%

Car Manufacturer 12%

Content Provider2%

Service Provider3%

Road Manager/Operator8%

Public Authority15%

Figure 12: Stakeholders’ typology




5.2.1. Survey results

In the following some survey results are shown for three years and different V2X technological solutions of the Business and Service models.

Business model base – Public

In this model it is assumed that the SAFESPOT system (in-vehicle components, infrastructure) is completely financed by the public authority such that the end user gets the in-vehicle system for free and does not pay for infrastructure using.

Table 20: Market penetration forecasts for Business model base - Public

2015 2020 2030

V2V V2I V2V V2I V2V V2I

Average 13 % 14 % 30 % 33 % 55 % 59 %

Pessimistic 2 % 3 % 7 % 9 % 16 % 23 %

Intermediate 7 % 9 % 23 % 29 % 56 % 59 %

Optimistic 30 % 31 % 60 % 60 % 93 % 93 %

Business Model penetration rate

13%

30%

55%

14%

59%

33%

0%

10%

20%

30%

40%

50%

60%

70%

2015 2020 2030

Time Frame

Ma

rke

t P

en

etr

ati

on

V2V

V2I

Figure 13: Market penetration forecast for Business model base – Public




Business model base – Public/ Private

In this model it is assumed that the user participates in financing the SAFESPOT system (in-vehicle components, infrastructure) such that the public authority provides only a subsidy for the in-vehicle system and the infrastructure.

Table 21: Market penetration forecast for Business model base – Public/Private

2015 2020 2030


Average 16 % 17 % 32 % 38 % 49 % 53 %

Pessimistic 5 % 5 % 9 % 14 % 16 % 19 %

Intermediate 14 % 16 % 29 % 36 % 45 % 52 %

Optimistic 30 % 29 % 58 % 64 % 85 % 86 %

V2V Public+Private Reliance average

49%

32%

17%

0%

20%

40%

60%

2015 2020 2030

Time Frame

Ma

rke

t P

en

etr

ati

on

Figure 14: Market penetration forecast for Business model base – Public/Private for V2V solution

Service model base – Private

In this model it is assumed that the user pays completely for the SAFESPOT system (in-vehicle components, infrastructure).

Table 22: Market penetration forecast for Service model – Private

2015 2020 2030


Average 10 % 12 % 23 % 27 % 43 % 46 %

Pessimistic 1 % 2 % 5 % 7 % 11 12 %

Intermediate 5 % 6 % 16 % 21 % 37 % 38 %

Optimistic 24 % 27 % 46 % 52 % 81 % 81 %




Service Model V2V-V2I

10%

23%

43%46%

12%

27%

0%

10%

20%

30%

40%

50%

2015 2020 2030

Time Frame

Ma

rke

t P

en

etr

ati

on

V2V

V2I

Figure 15: Market penetration forecast for Service model – Private for complete sample

Combined V2V and V2I based solution

In this model it is assumed that the SAFESPOT system consists simultaneously on the V2V and the V2I based solution. The combined system is completely public financed.

Table 23: Market penetration forecasts for combined SAFESPOT solution 2015 2020 2030

Pessimistic 5 % 15 % 34 %

Intermediate 16 % 40 % 71 %

Optimistic 55 % 81 % 97 %

V2V+V2I with Public Reliance

34%

16%

40%

71%

55%

81%

97%

15%

5%0%

20%

40%

60%

80%

100%

2015 2020 2030

Time frame

Mark

et

Pen

etr

ati

on

Figure 16: Market penetration forecast for combined SAFESPOT solution for three cluster (pessimistic, intermediate, optimistic)




SAFESPOT Service model plus – Private

In this model it is assumed that the user pays for SAFESPOT system (in-vehicle components, infrastructure). The SAFESPOT system includes also additional service features.

Examples of integration of additional functions into the SAFESPOT system can be:

• Traffic Information;

• Automatic road toll payment;

• Parking reservation.

Market penetration increase due to the

introduction of further applications

21%

28%

18%

6%1%

14%

3% 4%0%

4%

0%5%

10%15%20%25%30%

0-10

%

11-2

0%

21-3

0%

31-4

0%

41-5

0%

51-6

0%

61-7

0%

71-8

0%

81-9

0%

91-1

00%

Market penetration rate

Resp

on

den

ts (

%)

Figure 17: Increase in market penetration, if additional functions are provided

The figure shows the increase in market potential of the SAFESPOT system in new vehicles the experts expect if additional services are provided. 28 % expect an increase in the range between 11 and 20 percent points, 21 % in the range between zero and 10 percent points and 18 % in the range between 21 and 30 percent points and so on. 50 % of the experts expect an increase in the market potential in the range between 21 and 30 percent points. Thus, in summary, only moderate effects of the provision of additional services on attractiveness and market potential of the SAFESPOT cooperative systems are expected. For example, assuming a market potential of 5 % and adding further functions which increase the market potential about 20 percent points will result in a market potential of the SAFESPOT system in new vehicles of 6 %.




5.2.2. Summary of survey results

Table 24: New vehicle market penetration forecasts for the Business and Service Models

Scenario 2015 2020 2030


Business model base – Public

13 % 14 % 30 % 33 % 55 % 59 %

Business model base – Public/ Private

16 % 17 % 32 % 38 % 49 % 53 %

Service model base – Private

10 % 12 % 23 % 27 % 43 % 46 %

Combined V2V and V2I solution

25,33 % 45,29 % 67,27 %

The different business and service models define the scope of services delivered by the SAFESPOT system and the payment structure. It can be seen when users pay the entire price for the SAFESPOT system, market potential of the SAFESPOT system in new vehicles for the three target years is lowest for all models. Surprisingly, given that public authorities completely finance installation of SAFESPOT systems, market potential is lower than if users pay at least partly for the applications themselves. By comparing the V2V technological solution with the V2I solution, one can see that for all years and all models the market potential of the V2I case are on a higher level than for the V2V solution. Interestingly, the highest level of market potential is reached for all three target years for the case when both V2V and V2I systems are installed.




6. Impact estimation

This chapter describes the impact estimation for the selected cooperative SAFESPOT systems which have been composed of bundles of safety applications (see chapter 3). The safety impact assessment is based on the methodology and the results of the eIMPACT project. The estimation of traffic impacts is performed by means of a literature review and expert opinion.

6.1. Safety impacts

The safety impacts are the starting point for calculating the safety benefits.

To determine the safety impacts, the methodology developed in a previous project on socio-economic impact assessments on IVSS was applied [WILMINK et al. (2008)]. The approach used there assumes that, when one element of a system is affected, the consequences may appear in several elements and levels of the system, both immediately and in the long term, due to behavioural modification.

For each system this method considers specific mechanisms acting on driving behaviour, and the types of accidents which affect these mechanisms. The mechanisms are described in section 6.1.1.

Starting point of the description of the accident types in section 4.2.1 were the total number of fatalities and injured that occurred in the year 2005. Then, the relevant fatalities and injured were selected and categorized into one of the nine collision types. Fatalities relating to, for example, motorcycle crashes were not included in the accident base because they are not addressed by the SAFESPOT applications. Collision types collected in category 9 (“All other collisions”) which could not be allocated to one of the other specified collision types are also excluded since for unspecified collision types a safety effect could not be determined. Therefore, only the 8 collision types were considered for the safety impact estimation.

Because the determination of safety benefits in the following chapter is based on a forecast of the total number of fatalities and injured in 2020 (including all collision types and causalities) a correction factor has to be used to recalculate the derived safety impacts. Using this factor the derived avoidance potential of the safety systems can be expressed in relation to the total number of fatalities and injured.

The aim of the safety impact assessment is to provide estimates of safety impacts of SAFESPOT applications on the EU-25 level for the target year 2020. The safety impact estimates are given in percentage terms. Changes are then translated into numerical estimates of fatalities and injured that could be avoided by using the share of “relevant” accidents (see Chapter 4.2). These estimates will provide a central input for the benefit-cost calculations, in which monetary values for these benefits are provided.




6.1.1. Mechanisms of safety effects

The approach used here is based on work that has already been performed (e.g. in eIMPACT) [BENZ et al. 2008, p. 31-36]. The approach aims to capture all possible effects (unintended and intended) of IVSS on driving behaviour and on accidents in a systematic manner. The estimation covers the three main factors of road safety:

1. Exposure to accident during a trip;

2. Risk of collision;

3. Risk of collision to result in death or injury.

These three factors are covered by nine behavioural mechanisms, firstly described by DRASKÓCZY et al. 1998 which will determine the “net” safety effect of a considered system.

The first five mechanisms relate to accident risk:

1. Mechanism 1 considers direct in-car modification of the driving task by giving information, advice, and assistance or taking over part of the task. This is the first step in what the system does to a driver.

2. Mechanism 2 considers direct influence by roadside systems mainly by giving information and advice. Consequently, the impact of this influence is more limited than that of the in-vehicle systems. For the systems we are dealing with here, this mechanism will not apply most of the time.

3. Mechanism 3 considers indirect modification of user behaviour in many, still largely unknown ways. The driver will always adapt to the changing situation. This is often called behavioural adaptation and will often not appear immediately after a change but may show up later and it is hard to predict. Behavioural adaptation may appear in many different ways, for example, by change of usage of the car, by change of headway in a car following situation, or by change of expectation of the behaviour of other road users.

4. Mechanism 4 considers indirect modification of non-user behaviour. This type of behavioural adaptation is even harder to study because it is often secondary. Non-equipped drivers may for example change their behaviour by imitating the behaviour of equipped drivers, for example, driving closer or faster than they should, not having the equipment.

5. Mechanism 5 considers modification of interaction between users and non-users. Intelligent transportation systems (ITS) will change the communication between equipped road users. This change of communication may influence the traditional communication with non-equipped road users. To a large extent this problem may appear in the interaction between drivers and unprotected road users.

The next group of mechanisms relate to exposure:




6. Mechanism 6 considers modification of road user exposure by for example information, recommendation, restrictions, and so on. This is certainly an area where introduction of ITS will have a large impact.

7. Mechanism 7 considers modification of modal choice by for example demand constraints (area access restriction, road pricing, and area parking strategies), supply control by modal interchange and other public transport management measures, and travel information systems. Different travel modes have different accident risks, therefore any measure which influences modal choice also has an impact on traffic safety.

8. Mechanism 8 considers modification of route choice by demand restraints by route diversions, route guidance systems, dynamic route information systems, hazard-warning systems monitoring incidents. There are different parts of the road network, including different categories of roads which have different accident risks. Any measure that influences route choice by diverting traffic to roads of different category, therefore, also has an impact on traffic safety.

The final mechanism relates to the consequence of impact:

9. Mechanism 9 considers modification of accident consequences by intelligent injury reducing systems in the vehicle, by quick and accurate crash reporting and call for rescue, by reduced rescue time.

6.1.2. Effects on driver behaviour for relevant impact mechanisms

The estimates for each mechanism can be determined by different means based on the following considerations:

1. Existing empirical studies, for example, Field Operational Tests or Naturalistic Driving Studies.

2. Determined by so-called behaviour-accident risk models. Input performance indicators for these models are for example the mean speed and the standard variation of the speed (for a review on behaviour-accident risk models see [JANSEN 2001]). This approach has also been addressed in the AIDE project [ALONSO et al. 2005]. For example, if a study shows that drivers change their mean speed when driving with the system (as compared to baseline), we may estimate the resultant effect on accident risk through the Nilsson functions. If they change their speed variability, we may apply the Salusjarvi-function, etc. In all, 6 to 7 behaviour-risk functions would be needed to obtain an exhaustive description. In practice, this will seldom be available, so we will have to work with a limited set.

3. Found by so-called indirect evidence on safety impacts. For systems for which no empirical evidence is available we have to apply theoretical insights, which are based on the following elements:




• The first element is the directness of the device, that is, the degree to which drivers can notice the device is there, based on its apparent presence, and of the frequency with which it acts or gives information to the driver. If the device has a prominent appearance, or the driver has been made very aware that it is there in the background (for example, by promotional material), and it is active in a way that the driver notices, this is a pre-condition for significant forms of behavioural adaptation.

• The second element is the type of change that the system induces. We have to distinguish: (a) changes in behaviour, and (b) changes in a driver’s alertness level. In other words, some systems elicit mainly (or only) a behavioural response, while others will mainly invite drivers to a state of overall inattentiveness or to being less on their guard than previously.

• The size of each of these possible effects is derived from available empirical behavioural studies and from modelling work based on them.

• For behavioural changes, the “worst case” has been suggested to be a discount factor of 85 %, i.e., the accident reduction effect is brought down to 15 % of its initial value because drivers change their behaviour in a counter-productive way. This would apply to highly visible systems, which frequently make their presence known even under relatively normal driving conditions. Other levels of ”discounting” by behavioural modification are 50, 30, 15 and 0 %. Percentages that are more exact are presently impossible to give.

• For alertness changes, studies suggest the ‘worst case’ is a 75 % discount. Other levels are 50, 25 and 0 %.

4. Taken from previously validated safety estimates of future systems undertaken in EU projects such as CODIA and eIMPACT [see e.g. WILMINK et al. 2008; KULMALA et al. 2008].

The fourth point was most frequently used in the assessment.

After through-multiplication of all factors (for all separate mechanisms), a final multiplier remains. We thus get a resulting net percentage remaining of the initial total. Of course, the “real” net effects are a function of penetration rate in the population. For the present discussion, net effects were estimated under the assumption of 100% penetration rate.




6.1.3. Results of safety effects

In this section, the safety effects of the selected sub-applications are described.

6.1.3.1. V2V applications

Lateral Collision – LATC: Road intersection safety

To focus here on the “pure” V2V functionality of the Road intersection safety application only the use case “Obstructed view at intersections” is considered for safety assessment. The considered functionality is similar to a left-turn assistant. The remaining functionalities of this application, which have also potential safety impacts, use either infrastructural information (traffic light status) or focus on emergency vehicles and so are not considered here. Traffic-light information is information provided by the infrastructure and is therefore not accounted for in this “comparison” between V2V and V2I. The V2I associated application (IRIS – basic version) does not address Emergency vehicles. Therefore, they are not taken into account for this application either.

The safety impact estimate of similar left-turn functionality was carried out in the PReVAL project [SCHOLLIERS et al. 2008]. PReVAL also draws on the approach used in the eIMPACT project [BENZ et al. 2008]. So the estimates proposed in this deliverable for the Road intersection safety sub-application are based on the figures of the PReVAL study.

The two most important behavioural mechanisms of this application are:

• Direct in-car modification (M1): The initial effects for the different accident types for country cluster 1 are based on the PReVAL study. For country cluster 2 and 3, the estimated effects are in ratio with cluster 1. The ratio is determined by the quotient between the number of fatalities or injured at intersections in cluster 1 and cluster 2 or 3.

• Indirect modification of user behaviour (M3): The Road intersection safety application causes a small decrease in alertness when approaching intersections. Furthermore, drivers learn to trust and rely on a system that is not 100% reliable. Drivers might change the strategy for approaching an intersection for example by increasing the approach velocity. The PReVAL study is again the basis for Mechanism 3/cluster 1. Clusters 2 and 3 are in ratio with the fatalities and injured accident data at intersections.

Figure 18 shows the resulting safety effects estimates for the LATC – Road intersection safety application on fatalities with 100% fleet penetration for the EU-25. The vertical axis represents the nine mechanisms and the total effect. The horizontal axis indicates the estimates of the effects. A negative number represents a reduction in fatalities and positive number an increase. The safety effect for the EU-25 is found by summing the absolute reduction in the number fatalities (injured) of all the clusters. The resulting safety effect is found by dividing this absolute reduction by the total number of fatalities (injured).




-10% -9% -8% -7% -6% -5% -4% -3% -2% -1% 0% 1% 2%

Total impact

Mechanism 9

Mechanism 8

Mechanism 7

Mechanism 6

Mechanism 5

Mechanism 4

Mechanism 3

Mechanism 2

Mechanism 1

effect on number of fatalities Figure 18: V2V Lateral Collision (LATC) – Road intersection safety effect on fatalities with 100%-

fleet penetration, EU-25.

Table 25 summarizes the effect of the LATC sub-application on fatalities and injured for 100 % penetration. It shows that a reduction of 0.7 % of the fatalities and a reduction 2.2% of the injured are estimated.

Table 25: The safety effect of LATC on fatalities and injured for full penetration

Penetration rate [%] Safety Impact

Fatalities [%] Injured [%]

LATC – Road intersection safety 100 -0.7 -2.2

Road Departure – RODP: Road conditions status/Slippery Road

The purpose of this sub-application is to transmit warning information concerning a slippery road.

A safety impact estimation of a similar function was carried out in the eIMPACT [WILMINK et al. 2008] and CODIA EU projects [KULMALA et al. 2008]: the Wireless Local Danger Warning (WLD). One function of the WLD system gives a warning about reduced friction and visibility (WLD2). The estimates in this deliverable are based on the results of these projects.

Specifically, the RODP considers slippery roads and the WLD considers friction and visibility. Therefore, the figures have to be adapted. This is carried out by using data of the CODIA project [KULMALA et al. 2008]. In the CODIA study, the change in velocity caused by the WLD2 application is estimated by means of the literature. Next, the safety effect is determined by using the Nilsson function




[NILSON 2004] and the assumption of a base velocity of 80 km/h. The effect of an integrated system for both, friction and visibility is found by averaging the two separate effects. From these data multiplication factors are calculated to determine the RODP sub-application effect from the eIMPACT results. Table 26 shows the resulting multiplication factors.

Table 26: Multiplication factors to determine the safety effect for the RODP application

Initial effect for adverse weather conditions

Average friction and fog

[%]

Accident category

Multiplication factor

Friction [%] Fog [%]

-5 -12.1 -8.6 Injured 0.6

-10 -22.8 -16.4 Fatalities 0.6 The relevant safety mechanisms are:

• Direct in-car modification (M1): The initial effects for the different accident types for country cluster 1 are based on the eIMPACT study. For country cluster 2 and 3, the estimated effects are in ratio with cluster 1. For this application, the ratio is determined by the quotient between the number of fatalities or injured for adverse weather at cluster 1 and cluster 2 or 3.

• Indirect modification of user behaviour (M3): The initial benefit will be reduced because of adaptation (i.e. driving faster or with less attention under bad weather conditions). The reason is that the driver learns to trust the systems and relies on the system. Since the system might not always operate perfectly, the trust or rather the delegation of responsibility causes an adverse effect. Besides that, the driver might driver faster with a system. The eIMPACT study is again the basis for Mechanism 3/cluster 1. Cluster 2 and 3 are in ratio with the fatalities and injured accident data for adverse weather conditions.

In the eIMPACT and CODIA studies the interaction between users and non-users is positioned as a relevant mechanism: The non-users (drivers) approaching a hazardous location are forced to slow down when the equipped vehicle in front of them slows down. However, in this deliverable the effect of mechanism 5 is neglected because we assume a 100 percent penetration of equipped vehicles. The resulting safety effect of this application with a penetration rate lower than 100% might therefore be higher.




Figure 19 shows the estimates for this application:

-10% -9% -8% -7% -6% -5% -4% -3% -2% -1% 0% 1% 2%

Total impact

Mechanism 9

Mechanism 8

Mechanism 7

Mechanism 6

Mechanism 5

Mechanism 4

Mechanism 3

Mechanism 2

Mechanism 1

effect on number of fatalities

Figure 19: V2V Road departure (RODP) – Road condition status: slippery road safety effect on fatalities with 100 %-fleet penetration, EU-25

Table 27 summarizes the effect of the RODP – Road condition status/slippery road sub-application on fatalities and injured for 100 % penetration. It shows that a reduction of 1.0% percent of the fatalities and a reduction 0.5% of the injured are estimated.

Table 27: The safety effect of the RODP – Road condition status: slippery road for full penetration



RODP – Road condition status/ Slippery road 100 -1.0 -0.5

Longitudinal Collision – LONC: Speed limitation and safe distance

This V2V application performs “speed limitation” and “keeping a safe distance”. It therefore resembles the V2I Speed Alert application, but not completely. Speed Alert has no safe distance functionality. Speed Alert is discussed in the next section.

A safety impact estimation of a similar function was carried out in the eIMPACT [WILMINK et al. 2008] and CODIA [KULMALA et al. 2008] projects: Speed Alert (SPE). In eIMPACT a distinction is made between two different implementations. The second implementation – indicated in the study by SPE2 – addresses the legal speed, speed limit violation and dynamically recommends speeds in curves, near work zones and schools, on slopes and bridges, and for events, weather




and traffic. The CODIA study accounts for speed limits, weather conditions, obstacles and congestion. These results are rather similar compared to the results of the eIMPACT study. Because the scope of the eIMPACT result is a little bit broader, we base the estimates for the Speed Limitation application in this deliverable on the eIMPACT project.

The Safety Distance functionality will warn a driver if he gets too close to the lead vehicle, which will reduce breaking time or reduce the time to release the accelerator. In the COWI project (see final report: COWI (2006)) an overview is presented on the safety effects of these kinds of collision warning systems. The estimates for the Safe distance functionality in this deliverable, therefore, are based on COWI (2006). There a reduction in collision probability between 8 and 16% for fatalities and 10 and 30% for injured is shown.

The relevant safety mechanisms are:

• Direct in-car modification (M1): The initial effects for the different accident types for all country clusters are based on the eIMPACT study and COWI (2006).

• Direct influence by roadside systems (M2) In the CODIA and eIMPACT projects, mechanism 2 is included for the speed limit functionality. The use of roadside speed enforcement cameras decreases the maximum velocities of the equipped vehicles. Speed enforcement is a trend that is assumed to further increase in the coming years. The forecast is that this enforcement will cause a stronger influence on the safety impact of the application compared to the influence when there is no application installed in the vehicles. The magnitude of the effect is taken from the eIMPACT project.

• Indirect modification of user behaviour (M3): The eIMPACT study argues that mechanism 3 will increase the safety impact. The motivation lies in the fact that drivers will change there behaviour due to strong speed enforcement. The CODIA project gives an analysis for the separate functionalities (i.e. speed limit, adverse weather and obstacles and congestion). For adverse weather and obstacles/congestion, mechanism 3 reduces the initial effect. In the present estimation we assume that there will be a mild form of adaptation in some conditions, leading to some reduction in the initial effect.

• Modification of route choice (M8): The CODIA study predicted that drivers are tempted to use roads with higher speed limits, which is a safety benefit as roads with higher speed limits tend to have higher quality and hence, lower accident rate on average. This effect is also included in the present study.

In the eIMPACT and CODIA studies the interaction between users and non-users is positioned as a relevant mechanism: “Reduced speed by application users will influence drivers behind the equipped vehicle, especially on two-lane urban roads.” However, in this deliverable the effect of mechanism 5 is neglected




because we assume a 100% penetration of equipped vehicles. The resulting safety effect of this application with a penetration rate lower than 100% might therefore be higher.

Figure 20 shows the resulting safety effects:

-10% -9% -8% -7% -6% -5% -4% -3% -2% -1% 0% 1% 2%

Total impact

Mechanism 9

Mechanism 8

Mechanism 7

Mechanism 6

Mechanism 5

Mechanism 4

Mechanism 3

Mechanism 2

Mechanism 1

effect on number of fatalities Figure 20: V2V Longitudinal (LONC) safety effect on fatalities with 100 %-fleet penetration, EU-25

Table 28 summarizes the effect of the LONC sub-application on fatalities and injured for 100%. It shows that reductions of 7.5% percent for the fatalities and of 6.1% for the injured are estimated.

Table 28: The safety effect of the LONC sub-application



LONC - Speed limitation and safe distance 100 -7.5 -6.1

6.1.3.2. V2I applications

Intelligent Cooperative Intersection Safety system – IRIS: Basic application

The objective of the basic version is to identify potential red light violators, support drivers turning right in being aware of pedestrians and cyclists as well as assist left turning vehicles without a separate signal stage. Note that this application has more functionality with respect to the previous described LATC – Road intersection safety application.

As mentioned previously, the safety impact estimate of a left-turn function was carried out in the PReVAL project [SCHOLLIERS et al. 2008]. The two other




functions of the IRIS – Basic application show some similarity with functions studied in the eIMPACT [see WILMINK et al. 2008] and CODIA projects [see KULMALA et al. 2008]. These functions are the traffic light assistance (INS TL) and the right-of-way support (INS RoW) in eIMPACT. The estimate in this deliverable is based on the figures of the eIMPACT study.

To be more precise, the functionality of the traffic light assistance (INS TL) resembles the functionality of the red light violators’ support. In contrast, the right-of-way support (INS RoW) described in the eIMPACT study is not exactly the same as the support for drivers turning right in being aware of pedestrians and cyclist (i.e. the initial safety effect of the IRIS functionality is less since it considers only vulnerable road users). To cope with the difference between eIMPACT and SAFESPOT, the effect of the eIMPACT system on vehicles is not taken into account for the present study.

-10% -9% -8% -7% -6% -5% -4% -3% -2% -1% 0% 1% 2%

Total impact

Mechanism 9

Mechanism 8

Mechanism 7

Mechanism 6

Mechanism 5

Mechanism 4

Mechanism 3

Mechanism 2

Mechanism 1

effect on number of fatalities Figure 21: V2I Intersection Safety effect (IRIS) on fatalities with 100 %-fleet penetration, EU-25

Figure 21 shows the estimates of the IRIS sub-application. The effects for the behavioural mechanisms were estimated as follows:

• Direct in-car modification (M1): The initial effects for the different accident types for country cluster 1 are based on the PReVAL and eIMPACT studies. For country cluster 2 and 3, the estimated effects are in ratio with cluster 1. The ratio is determined by the quotient between the number of fatalities or injured at intersections in cluster 1 and cluster 2 or 3.

• Indirect modification of user behaviour (M3): Similar to the LATC sub-application, the IRIS sub-application causes a small decrease in alertness when approaching intersections. Furthermore, drivers learn to trust and




rely on a system that is not 100 % reliable. Drivers might change the strategy for approaching an intersection for example by increasing the approach velocity. The PReVAL and eIMPACT studies are also the basis for Mechanism 3/cluster 1. Cluster 2 and 3 are in ratio with the fatalities and injured accident data at intersections.

Table 29 summarizes the effect of the IRIS basic-application on fatalities and injured for 100 % penetration. It shows that a reduction of 3.1 % percent for the fatalities and a reduction of 4.8 % for the injured are estimated.

Table 29: The safety effect of the IRIS basic application



IRIS – basis application 100 -3.1 -4.8

Hazard and Incident Warning – H&IW: Abnormal weather conditions

The aim of this sub-application is to be able to alert drivers to the presence of a hazard due to weather conditions (e.g. rain, ice or fog) to result in reduced friction and low visibility.

A safety impact estimation of a similar function was carried out in the eIMPACT [WILMINK et al. 2008] and CODIA [KULMALA et al. 2008] projects: the Wireless Local Danger Warning (WLD). One function of the WLD system gives a warning about reduced friction and visibility (WLD2). Hence, contrary to the RODP application, the HIW sub-application also handles the visibility issue. The estimates in this deliverable are based on both projects.

-10% -9% -8% -7% -6% -5% -4% -3% -2% -1% 0% 1% 2%

Total impact

Mechanism 9

Mechanism 8

Mechanism 7

Mechanism 6

Mechanism 5

Mechanism 4

Mechanism 3

Mechanism 2

Mechanism 1

effect on number of fatalities Figure 22: V2I Hazard and Incident Warning (H&IW) safety effect on fatalities with 100 %-fleet

penetration, EU-25




Figure 22 shows the estimates of the H&IW sub-application. The relevant safety mechanisms are:

• Direct in-car modification (M1): Similar to the RODP sub-application, the initial effects for the different accident types for country cluster 1 are based on the eIMPACT study.

• Indirect modification of user behaviour (M3): Similar to the RODP sub-application, the initial benefit will be reduced because of adaptation (i.e., driving faster or with less attention under bad weather conditions).

Table 30 summarizes the effect of the H&IW sub-application on fatalities and injured for 100 % penetration. It shows that reductions of 1.6% percent for the fatalities and of 0.7% for the injured are estimated. Since this application has more functionalities than the corresponding V2V RODP application, its effect is higher.

Table 30: The safety effect of the HIW sub-application



H&IW – Abnormal weather conditions 100 -1,6 -0.7

V2I Speed Alert: sub-application Legal Speed Limit

The objective is to provide to the driver a continuous knowledge of the legal speed limit and to allow the infrastructure manager to modify the legal speed limit. This implies that the previously discussed H&IW sub-application is taken into account by the Speed Alert application.

As mentioned for the Longitudinal Collision sub-application (Section 6.1.3.1), a safety impact estimation of a similar function was carried out in the eIMPACT [WILMINK et al. 2008] and CODIA [KULMALA et al. 2008] projects: Speed Alert (SPE). The estimates for the Speed Limitation in this deliverable are based on the eIMPACT project.

Estimates for the behavioural mechanisms are obtained as follows: • Direct in-car modification (M1): The initial effects for the different accident

types for all country clusters are based on the eIMPACT study.

• Direct influence by roadside systems (M2): The use of roadside speed-enforcement cameras, decreases the maximum velocities of the equipped vehicles. Furthermore, this enforcement will cause an increase of the safety impact of the SpA application.

• Indirect modification of user behaviour (M3): In the present estimation we assume that there will be a mild form of adaptation in some conditions, leading to some reduction in the initial effect.

• Modification of route choice (M8): The CODIA study predicted that drivers are tempted to use roads with higher speed limits, which is a safety benefit




as roads with higher speed limits tend to have higher quality and hence, lower accident rate on average. This effect is also included in the present study.

Figure 23 shows the resulting safety effect estimates:

-10% -8% -6% -4% -2% 0% 2%

Total impact

Mechanism 9

Mechanism 8

Mechanism 7

Mechanism 6

Mechanism 5

Mechanism 4

Mechanism 3

Mechanism 2

Mechanism 1

effect on number of fatalities Figure 23: V2I Speed Alert (SpA) safety effect on fatalities with 100 %-fleet penetration, EU-25.

Table 31 summarizes the effect of the SpA sub-application on fatalities and injured for 100%. It shows that reductions of 7.1% percent for the fatalities and of 4.9% for the injured are estimated. The estimate of the V2I SpA effect is a bit smaller than that of the V2V LONC variant. This is due to a small difference in the application. The LONC application includes a safe distance advice.

Table 31: The safety effect of the Speed Alert sub-application



Speed Alert – Legal Speed Limit 100 -7.1 -4.9

6.1.3.3. Bundles

Considering the benefit of a bundle of applications the benefit is assumed to be maximal if each application works independently such that different accident types are possibly prevented. This implies that the different systems do not distract, overload, confuse and annoy the driver, thus causing problems that did not exist for the systems in isolation. The reader is referred to the AIDE project, for example, on this issue [BROUWER and HOEDEMAEKER 2004]. In the AIDE project an experiment was performed with two applications examined separately




and with the two applications bundled. The results of the bundle were different compared to the summed effect of the two systems separately.

Considering the example that each of the two systems avoids 100% of head-on collisions, the bundle avoids also 100 % of the head-on collisions. In this case, the combination of the applications has no additional benefit. In the other case, for example, one application avoids 20 % of all head-on collisions and the second application avoids 20 % of all lateral crashes the benefits of both systems can be summed up to reach the whole benefit of the bundle. If there is some overlapping of safety effects concerning prevented accidents of a specified accident type, the bundle will be below the sum of the benefits of the applications. Then, to avoid double counting, only one application according to the specified accident type is considered.

V2V bundle

In this section an estimate of the safety impact of the bundle containing the three V2V applications will be derived.

The safety impacts of the three V2V applications have been estimated in the previous sections. These three applications are the Lateral Collision application (LATC), the Road Departure application (RODP) and the Longitudinal Collision application (LONC). As mentioned, the RODP sub-application is already taken into account by the LONC application. Therefore, the bundle is reduced to two applications. These applications are assumed to be independent because the intersection application will have an effect on other accidents than the speed application.

Figure 24 shows a representation of the V2V applications that resemble the V2V bundle. This figure illustrates that only applications with assumed independent safety impacts are considered such that we can add them up to illustrate the entire effect of the bundle. The complete calculation is not shown here. For doing this, first we have to calculate the safety effect for the different accident types on which the sub-applications have an impact. Then, the different allocations of accident types with regard to the sum of accidents have to be considered as weighting factors to derive the total effect.

Table 32 lists the effects of the V2V bundle.




Longitudinal Collision (LONC)

– Speed limitation and safe distance

Lateral collision (LATC) – Road

intersection safety

Figure 24: Representation of the V2V bundle

Table 32: The estimated safety effect of the V2V bundle

Penetration rate

[%]

Safety Impact


V2V bundle 100 -8.1 -8.3

V2I bundle

Similar to the V2V bundle in the section above, the safety impact of the bundle containing the three V2I applications is shown in this section.

The safety impacts of the three V2I applications have been estimated in the previous sections. These three applications are the Intelligent Cooperative Intersection Safety system (IRIS), Hazard and Incident Warning (H&IW) and the Speed Alert (SpA). The H&IW sub-application is taken into account by the SpA application. Therefore, the bundle is reduced to two applications. These applications are again assumed to be independent because the intersection application will have an effect on other accidents than the speed alert application. Figure 25 illustrates that only applications with assumed independent safety impacts are considered such that we can add them up to illustrate the entire effect of the bundle. Table 33 shows the effect of the V2V bundle.




Speed Alert (SpA) –

Legal speed limit

Intelligent

Cooperative

Intersect Safety (IRIS) – Basic

application

Figure 25: Representation of the V2I bundle

Table 33: The estimated safety effect of the V2I bundle



V2I bundle 100 -10.2 -9.7




6.1.4. Summary

Table 34 shows the safety impacts for each considered application. “Safety impacts” are estimated changes in fatalities or injured in terms of percentage.

Table 34: Estimates for behavioural mechanism effects for selected applications and bundles based on relevant aggregated accident statistics (EU-25)

Application Sub-application Aim

Safety impact

Fatalities [%]

Injured [%]

V2V

Lateral collision (LATC)

Road intersection safety/Obstructed view at intersections

Left-turn assistant -0.7 -2.2

Road departure (RODP)

Road condition status /Slippery road

Warning about slippery road only

-1.0 -0.5

Longitudinal Collision (LONC)

Speed limitation and safe distance

Information about speed limit and keeping safe distance

-7.5 -6.1

V2I

Intelligent Cooperative Intersection Safety system (IRIS)

Basic application

Identify potential red light violators, support drivers turning right and left turn assistant

-3.1 -4.8

Hazard and Incident Warning (H&IW)

Abnormal weather conditions

Warning about slippery road and reduced visibility

-1,6 -0.7

Speed Alert (SpA) Legal Speed Limit Information about speed limit

-7.1 -4.9

Bundles

V2V bundle Road intersection safety and Speed limitation and safe distance

Left-turn assistant and information about speed limit and keeping safe distance

-8.1 -8.3

V2I bundle Intersection Safety system and Speed Limit

Identify potential red light violators, support drivers turning right and left turn assistant and Information about speed limit

-10.2 -9.7

As mentioned above, the estimation of safety impacts that has been made here was based on eight collision types. The unspecified collision type nine (“All other collisions”) was not included in the estimation. However, the effect of this collision type of relevant accidents has also to be taken into account to get a more accurate estimation of the complete safety impact of the SAFESPOT system on the nine collision types. For this we assume that the safety impact on the 9th-collision type corresponds to the mean effect on the other (eight) collision types.




Given this mean effect, we can estimate the safety impact for the two SAFESPOT bundles for all (nine) relevant collision types. The corresponding numbers are shown in Table 34.

Since for the calculation of the safety benefits in 2020 the safety impacts which regard to the total number of fatalities and injured are required a recalculation was made using a correction factor. The factor is given by the ratio of the number of relevant fatalities or injured to the total number of fatalities or injured for the year 2005. Table 35 shows also the resulting “total” safety impacts.

Table 35: Estimates for behavioural mechanism effects for the bundles based on aggregated total accident statistics for EU-25

Bundle

Safety impact for the relevant accidents

Factor Safety impact for the

total of accidents

Fatalities [%]

Injured [%]

Fatalities [%]

Injured [%]

Fatalities [%]

Injured [%]

V2V -8.1 -8.3 0.87 (=36,069/ 41,304)

0.88 (=1,127,057/1,279,554)

-7.1 -7.3

V2I -10.2 -9.7 -8.9 -8.5

6.2. Traffic and environmental impacts

In addition to safety impacts, non-safety impacts may arise because the SAFESPOT system potentially influences traffic flow, fuel consumption and resulting emissions. In the following an overview will be given on the estimated impacts of the SAFESPOT system towards transport efficiency and emissions. These impacts can potentially create additional benefits which need to be taken into account in the cost-benefit analysis. We start with describing the assumptions and background information before the results are provided.

6.2.1. Assumptions

First, the relationship between traffic efficiency and environmental impact will be shown, as there is a direct correlation between the traffic efficiency and environmental impacts. This correlation can be illustrated by a TNO-study that estimates the emissions from vehicles in various driving stages. The table below gives some insight into the correlation between transport efficiency and emissions.




Table 36: Effect of more or less congestion on motorway sections on emissions

CO VOC NOX CO2 PM10

Less congestion

Least congested routes -7.9 % -21.3 % -10.3 % -3.4 % -10.8 %

Most congested routes -21.3 % -44.1 % -20.3 % -9.6 % -21.1 %

All routes -11.5 % -29.0 % -13.6 % -5.0 % -13.9 %

More congestion

Least congested routes 32.1 % 67.5 % 21.4 % 11.6 % 22.4 %

Most congested routes 16.1 % 31.2 % 10.5 % 6.1 % 9.6 %

All routes 24.9 % 53.5 % 16.2 % 8.7 % 14.7 % These figures are merely indicative and only show the relationship between transport efficiency and emissions. Thus, if traffic impacts actually occur, effects on the environment have to be expected. Consequently, to determine possible emission effects of the safety system we have to start with analysing traffic impacts.

The estimation of the traffic impacts was based on available literature since a lot of effort had already been put into estimating traffic impacts of IVSS in general. The following studies were taken into account:

• eIMPACT;

• CODIA;

• WillWarn.

The methodology follows a direct line from the selected applications, which are compared to the applications in the selected studies. With this comparison the limitations of the SAFESPOT system, as it is currently defined, also need to be taken into account. The possible effects on traffic can be derived from these various sources and are discussed in the following chapter.

6.2.2. Results

To get a feeling for potential traffic impacts a comparison is made with comparable applications from the selected projects to identify possible direct traffic impacts of SAFESPOT applications.

Intersection applications are discussed only in eIMPACT and SAFESPOT. Keeping in mind that eIMPACT has no communication module implemented for the various applications and given the current definition of the SAFESPOT system (single hop messages and short range communication)5, the LATC intersection application of SAFESPOT is still very comparable to the eIMPACT intersection application. Also, the Local danger warning application discussed in eIMPACT and CODIA is comparable to the RODP and H&IW applications from

5 Single hop communication indicates only transfers of messages from one node to another. It is not broadcasted or sent on by other vehicles without verification of the message.




SAFESPOT. For the LONC application of SAFESPOT the Full speed range ACC is used and the SpeedAlert function is discussed in eIMPACT and CODIA as well.

The results of the comparison are discussed in the table:

Table 37: Comparison of direct traffic impacts in SAFESPOT, eIMPACT and CODIA

SAFESPOT applications eIMPACT results CODIA results

LATC: Road intersection safety

Expected, but not proven, therefore assumed none

N/A

RODP: Road condition status - Slippery Road warning

Only very local effects can be expected, overall none

N/A

LONC: Speed limitation and safe distance

No real effects to be expected, depends strongly on the traffic volumes estimated

Increase in travel times of 0.2 % - 2.4 % on motorways strongly depending on level of service (traffic volume)

IRIS: Basic application Expected, but not proven, therefore assumed none

N/A

H&IW: Reduced friction or visibility

Only very local effects can be expected, overall none

Increase in travel times of 0.1% - 5.5% but only for motorways with Level of Service of E/F

SpA: Legal speed limit Assumption is that the overall effects are negligible, but the effects depend on traffic volume

N/A

Overall it can be concluded that the applications of eIMPACT and CODIA have little or no direct traffic impact. Based on these studies, it is assumed that for the BLADE-selected SAFESPOT applications also no significant traffic impacts will be obtained. Therefore, environmental impacts are estimated to be non-existent for the SAFESPOT applications.

Up to now we have not considered the effects which the SAFESPOT applications, RODP and H&IW, will have on improving in-car information on routing. This effect is not present in the eIMPACT applications. For the current SAFESPOT system only small scale validation and evaluation activities are carried out in the test sites, addressing almost exclusively the safety impact of the cooperative systems. Larger scale evaluations should be performed by means of specific Field Operational Tests. Therefore no traffic efficiency impacts are to be expected for the SAFESPOT test sites. However, the WILLWARN project can give some indication on which traffic efficiency effects can be reached. These effects are mainly valid for the H&IW application of the V2I based system and the RODP application of the V2V based system. However, from the WILLWARN results it can be seen that traffic impacts of improving in-car communication are slightly negative since travel time increases when taking into account rerouting. These impacts however are relatively small (under 3%) compared to the reference scenarios where no WILLWARN functionality is available [see NÖCKER (2007)].




Further, in terms of traffic impacts, WILLWARN has a slightly negative effect on the throughout measures of travel time and delay. This effect is caused by the braking action of WILLWARN vehicles before the hazard. This increases the standard deviation of the speeds on the road segment before the hazard. The potential accident reduction, as well as the mitigation in accident severity, was not taken into account. They are, however, expected to be significant as a result of decreased hard accelerations before and during the hazard.

In conclusion, it can be said that the traffic efficiency effects from SAFESPOT are estimated to be zero. Therefore, the effects on emissions are also estimated to be zero. This result, however, strongly relates to the current level of implementation and testing of the SAFESPOT system.




7. Results of the Cost-Benefit Analysis

The cost-benefit analysis (CBA) is performed for the SAFESPOT system bundles V2V and V2I. For each of the bundles BCRs are determined. In addition, the CBA is completed by a sensitivity analysis. Compared to other recent assessment approaches this study analytically connects all impacts in a single formula. In this formula the penetration rate represents the only independent variable.

7.1. Results of the V2V applications bundle

The following penetration rates are connected to the different business and service models developed in the market assessment and are used for the CBA (see chapter 5): 4.2 %, 6.1 %, 6.7 % and 8.7 %. These numbers represent the market penetration of the V2V systems in the entire vehicle fleet in the year 2020. The penetration rates have been estimated under the assumption that the deployment in new vehicles starts in the year 2015 [see SP6 WP6.6 (2009)].

7.1.1. Benefits

The benefits of the V2V co-operative systems depend on the probability of matching vehicles that are also equipped with such a system. This probability depends on the share of the vehicle fleet equipped with IVSS (penetration rate). Further, the probability of meeting another V2V vehicle will be higher if the annual mileage of these vehicles is above average. Thus, the share of annual mileage driven by vehicles equipped with cooperative systems also determines the matching probability. Additional influencing factors are the reception area of the communication tool used and the traffic flow.

The influence of the performance of traffic flow (average velocity, traffic density and traffic volume) on the matching probability and the benefits of the SAFESPOT system can be described, for example, by the concept of Level of Service (LOS) [FGSV 2001]. The LOS describes the quality of driving on the road section. If there is congestion, the level of service is bad (F). If the road is not crowded and traffic flows at or above the posted speed limit, the level of service is very good (A). However, the LOS simultaneously has an influence on probability of communication between vehicles and on occurrence of critical incidents, and thus on safety benefits. On the one hand, in congestion, the warning of an oil spill is not relevant because the vehicles’ speed is too low to get into a critical situation. So, the benefit of the received message is not very high because of very low speeds. On the other hand, in dense traffic there are enough vehicles so that the probability of meeting another equipped vehicle and so receiving a relevant message is high. In case of an undisturbed LOS the relevance of a message is higher, but the probability of receiving this message is lower. There is less traffic and hence the probability of meeting another equipped vehicle is marginal for example for LOS A or B for low penetration rates. Studies such as (HERRTWICH (2003)) show that V2V systems can only work properly for




penetration rates above 5 %. Below 5 % the probability of meeting another equipped vehicle is too low.

Thus, inter-relations between the effectiveness of the communication between the vehicles and safety benefits of the SAFESPOT systems in different traffic situations for different levels of services are fairly complex. However, since a LOS with a stable, not congested traffic situation is much more frequent than a congested LOS [see INFRAS/IWW 2004] we will not consider this aspect further for calculations. In the following it is assumed that the matching probability and thus benefits depend only on the share of the vehicles equipped with the SAFESPOT system under consideration. Further it is assumed that no benefits are achieved for penetration rates below 5 %. Above or equal 5 % the share of benefits shall equal the penetration rate. So, stated as a formula the factor which determines the share of benefits realizable is given as follows:

<

=elsewise )( rate npenetratio

5% rate npenetratio for 0

pfB

.

The factor Bf is multiplied with the safety effect of the bundle assuming full penetration and with the number of fatalities/injured. The resulting number of avoided fatalities and injured is then multiplied with the corresponding cost-unit rates to get the safety benefit of the V2V bundle. Because casualties can only be avoided in whole numbers, the floor function is used. The floor function rounds the number of avoided casualties down. The calculation formula for the safety benefits is given as follows:

( ) ( )( )IIFFBS c*I*ec*F*e*fB += , with (1)

BS Safety benefits [EUR]

fB Penetration rate (share of safety benefits) [%]

eF, eI effectiveness in avoiding fatalities / injured [%]

cF, cI Cost-unit rate for fatalities / injured [EUR]

F Total number of fatalities

I Total number of injured.

In addition to the safety benefits, traffic benefits can also arise out of the SAFESPOT system bundle. However, as was shown above (see chapter 6.2) traffic benefits are considered as 0 for the bundle. Thus, the above formula is valid for determining the complete benefits.

The values of the single parameters can be seen in the following table:




Table 38: Safety effects of the V2V bundle

Fatalities Injured

eF F cF eI I cI

Effective-ness rate

[%]

Number of fatalities [1,000]

Cost-unit rate

[Mill. EUR]

Effective-ness rate

[%]

Number of Injured [1,000]

Cost-unit rate

[Mill. EUR]

7.1 20.8 1.63 7.3 870 0.06 The numerical values of the effectiveness rate provided in Table 38 are used to calculate the safety impact for the total of accidents. Table 39 shows the number of avoided fatalities and injured per penetration rate. For penetration rate 4.2 % the values have been set to 0 as discussed above.The monetary value of the total safety benefits sums up the monetary values of the avoided fatalities and injured. The number of avoided injured is about forty times higher than the number of avoided fatalities, and in monetary values the benefit of avoiding injured is about 60 % higher than the benefit of avoiding fatalities.

Table 39: Safety benefits of the V2V bundle

Penetration rate

Avoided number of Fatalities [Mill. EUR]

Injured [Mill. EUR]

Total benefit [Mill. EUR] Fatalities Injured

4.2 % 0 0 0 0 0

6.1 % 89 3,862 145 231 376

6.7 % 98 4,253 159 255 414

8.7 % 128 5,541 209 332 541

7.1.2. Costs

The cost of the SAFESPOT V2V system bundle is given by the sum of the component costs. The components of the system bundle contain a V2V communication tool (including antenna and cables) and a display. These components will be assumed to be in series in the year 2020 and thus will not lead to extra costs. Further, the system bundle needs a dual frequency GPS, digital maps including intersections, a warning module and (as a reference sensing option) a LRR (Long-Range Radar) front.6

The costs of the single components were estimated by experts out of the project SAFESPOT from a project internal expert enquiry. In total, three loops were done until the final cost data was estimated. The cost estimation was based on two fleet penetration rates (4.2 % and 11.3 %) to cover the possible bandwidth of equipped vehicles. The penetration rate of 4.2 % is the lowest penetration rate which is considered within this project, and 11.3 % is the highest one. To include implementation costs the sum of component costs is multiplied by the factor 1.05 [BAUM et al. 2008].

6 The specification of the systems is done by the technical sub-projects. The inclusion of the LRR in the V2V bundle is made to align the System under Analysis with the current implementations available in the SAFESPOT test sites. The process of specification and the limitations for economic assessment arising out of it are described in section 2.2.2.




Table 40: Total component costs and system costs per penetration rate

Pene-tration rate

Component Costs [EUR] Sum [EUR]

System costs incl.

Implementation [EUR]

Dual freq. GPS

Digital maps

Warning module

LRR front

4.2 % 20.00 30.00 10.00 84.00 144.00 151.20

11.3 % 5.00 20.00 10.00 84.00 119.00 124.95 The estimation for the dual frequency GPS was 20 EUR for the penetration rate of 4.2 % and 5 EUR for 11.3 %. The range for digital maps was between 20 EUR and 30 EUR. The last two components – warning module and long-range radar front – were considered as independent of the penetration rate. The warning module costs 10 EUR and the long-range radar 84 EUR.

Figure 26 shows the cost composition for the V2V bundle. More than half of the costs are associated with the LRR front, followed by the digital maps which account for approximately 20 %. Dual freq.-GPS plus warning module account for less than 20 % of the total costs.

LRR front

Digital maps

Dual freq.-GPS

Warning module

Figure 26: Cost composition for the V2V bundle (without implementation costs)

It is assumed that the cooperative system is used over the complete lifetime of the vehicle. The average economic lifetime of a vehicle in EU-25 is 12 years (BAUM et al. (2008)). Because the socio-economic assessment will be done for the target year 2020 yearly, costs have to be determined. So the costs for the V2V system bundle given by the sum of the used components have to be multiplied with the annuity rate to get the yearly average costs. The annuity rate is determined by the following formula:




1004601031

03103011

112

12

,,

),(*,

)d(

)d(*dAR

T

T

=−

=−+

+= , with

AR annuity rate

d discount rate (3 %)

T service time (12 years).

The result is the average annual costs over 12 years for equipping the vehicle with the bundle. In addition, the digital maps require annual updates which costs about 20 % of the component costs. The updating costs are added to the yearly costs for the system. The result is the annual total costs for equipping and operating the vehicle with the system bundle.

The component costs of the dual frequency GPS and the digital maps depend on economies of scale on the part of producers: By doubling the production level, total production costs increase by a smaller amount such that average unit component costs decrease with the level of production. Consequently, for the dual frequency GPS and the digital maps the base costs (penetration rate: 1 %) have to be calculated to estimate the component costs for all other production and corresponding penetration rates. With economies of scale the component costs (cc) which are linked to the penetration rate (p) are determined by the degression rate (d) and the base costs. The degression rate is the cost reduction in percentage terms for doubling the output. Using the exponential function, the component costs can be written by the formula:

( )

( )21

costs base ln

dln

p*)p(cc

−

= , with (2)

cc component costs [EUR]

p penetration rate [%]

d degression rate [%].

In the following this function shall be used to estimate the base costs and the degression rate to derive an estimation function for the costs of the system bundle only as a function of the penetration rate. The calculation proceeds as follows:

Two values for the component costs and the corresponding penetration rates are used. If the output is doubled, the costs per unit are decreasing by d. This relationship is shown in the above function by the term ln (1-d) / ln (2). With respect to the above penetration rates, the change to the higher penetration rate leads to an increase of the output by the rate 11.3 % / 4.2 % (ph / pl). The change in costs following the change in the penetration rate, equals the term (1 – d) and is given by the rate of the lower component costs above higher component costs (ccl / cch). Given this calculations and using ccp = cch and p = ph the base costs are determined as follows:




=

l

h

h

l

p

pln

cc

ccln

h

h

p

cccosts base , with

cc component cost (low/high) [EUR]

p penetration rate (low/high) [%].

Using the values for component costs and corresponding penetration rates the base costs can be calculated: for a penetration rate of 4.2 % the dual frequency GPS costs 20 EUR, and for a penetration rate of 11.3 % only 5 EUR. Thus, if the output is scaled up with the factor 2.69 (11.3 % / 4.2 %), the costs are reduced to 25 % of the costs for the lower penetration rate (5 / 20). Thus, the exponent in the denominator of the last equation is -1.4 (ln (0.25) / ln (2.69)). As a result, the base cost for the dual frequency GPS is 149.30 EUR and the base cost for the digital maps is 54.01 EUR.

Table 41: Base component costs and base system costs

Penetration rate

Component Costs (base) [EUR]

Sum [EUR]

System costs incl.

Implementation [EUR]

Dual freq. GPS

Digital maps

Warning module

LRR front

1.0 % 149.30 54.01 10.00 84.00 297.31 312.18 For both penetration rates the average annual costs are determined. For the (low) penetration rate (pl) 4.2 % the component costs (cch) are 21.19 EUR. This is the sum of the system costs including implementation times, the annuity rate and the annual costs for updating the digital maps (= 151.20 EUR * 0.10 + 20 % * 30 EUR). The component costs (ccl) for the high penetration rate (cch) are 16.55 EUR (= 124.95 * 0.10 + 20 % * 20 EUR). With these two average yearly costs for both penetration rates, the base costs for the bundle are determined. They are 30.32 EUR per year.

Finally, the degression rate for component costs is determined. Therefore the exponents of the latter two formulas have to be equalized:

( ) ( )

%eed

ln*

.

.ln

.

.ln

ln*

p

pln

cc

ccln

l

h

h

l

1611

2

2431119215516

2

=−=−=

Thus, the costs can be calculated as follows:

( )( )

( )2161

3230 ln

%ln

p*.pcosts component

−

= [EUR] (3)

To determine the complete costs for the equipped vehicles, the component costs have to be multiplied with penetration rate p and with the estimated vehicle stock of the year 2020 (307 Mill. vehicles) [PROGTRANS (2007)]:




( )( )

( ) 7502161

2430893073230 .ln

%ln

p*.,*p*p*.pC ==

−

[Mill. EUR]. (4)

For the considered penetration rates the system costs and the total costs are displayed in the following table.

Table 42: System costs and total costs for the considered penetration rates

Penetration rate

Equipped vehicles [Mill.]

Costs per system [EUR]

Operating costs per

system [EUR p.a.]

Total costs [Mill. EUR p.a.]

4.2 % 13.0 151.20 6.00 276

6.1 % 18.6 140.71 5.15 359

6.7 % 20.5 138.19 4.96 386

8.7 % 26.7 131.41 4.45 471

7.1.3. Benefit-Cost-Ratio

The BCR compares the benefits with the costs. The benefits are determined by a function of the share of mileage driven which depends on the penetration rate. The costs are determined by the penetration rate. Thus, for all penetration rates below 5 % the BCR is not available.

( )( )

[ ] [ ]( )750243089 .

IIFFB

t

Btt

p*.,

c*I*ec*F*e*f

pC

fBBCR

+==

(5)

[ ] [ ]( )

+

<

=else

243089

5% rate npenetratio a for n.a.250

,..

c*I*ec*F*e*p IIFF

. .

The parameters can be set into the formula, so that the only unknown figure is the penetration rate p (valid for penetration rates above 5 %):

( )[ ] [ ]( )

243089

250

.,

c*I*ec*F*e*ppBCR IIFF

.+

=

[ ] [ ]( )

670243089

92326

Euro mill. 243089Euro mill, 06069587337Euro mill. 6317912017

250250

250

.*p.,

.,*p

.,

.*,*%..*,*%.*p

..

.

==

+=

To calculate the BCR, the penetration rates have to be set in the equation above. The lowest considered penetration rate is below 5 %, i.e. the BCR cannot be calculated. For the other penetration rates, the BCR is between 1.0 and 1.1 (see Table 43). The derived benefits outweigh the costs of the cooperative systems for the V2V case.




Table 43: Benefit-cost-ratios for the considered V2V business and service models

Model Penetration rate BCR

Service model base – Private 4.2 % -

Service model plus – Private 6.1 % 1.0

Business model base– Public 6.7 % 1.1

Business model – Public/Private 8.7 % 1.1

7.2. Results of V2I applications bundle

For the V2I bundle the cost-benefit-analysis (CBA) deviates in some respect from that of the V2V bundle, but is based on the same methodological foundations.

For determination of the benefits the above results of safety and traffic impact analysis are used as input on the benefit side. On the cost side several assumptions have to be made about the equipment rate in order to be able to estimate the costs of infrastructure.

For the V2I bundle the following penetration rates were considered: 5.4 %, 7.7 %, 9.5 % and 11.3 %. These penetration rates are based on the assumption that the system will be deployed from the year 2015 onwards in only new vehicles. The penetration rates themselves are derived form the market assessment. The high and low numbers are derived from the various business and service models that are available for the V2I scenario [see SP6 WP6.6 (2009)].

7.2.1. Benefits

The SAFESPOT V2I applications are situated both at intersections and stretches of the road. In contrast to the V2V case discussed above, the main advantage of equipping the infrastructure is the fact that an equipped vehicle can experience the benefits from equipped infrastructure independently of the penetration rate of other vehicles. Therefore there is no minimal penetration rate for the V2I bundle in order to generate benefits, in contrast to the V2V bundle where a penetration rate of at least 5 % is required. In the case study some more attention will be given to this assumption (see section 7.3.3).

Given the safety effects of the V2I system bundle the benefits are determined, on the one hand, by the number of intersections and sections of highways that are equipped and, on the other hand, by the number of equipped vehicles passing the equipped infrastructure. The number of passing vehicles depends on the number of equipped vehicles which is determined by the fleet penetration rate and by the mileage driven by these vehicles. Thus, if we assume that the amount of equipped infrastructure is fixed, the benefits will depend on the mileage driven by equipped cars.

In the following we use a regression function for the estimation of the inter-relationship between the fleet penetration rate and mileage driven by the fleet. The idea of this function is that new vehicles have a higher mileage than older ones and that people who have a higher-than-average mileage will equip their




vehicle with the cooperative system rather than the user on average mileage [see ABELE et al. 2005]. In the eIMPACT project an empirical foundation of the function was developed [see BAUM et al. 2008]. For a few countries the distribution of the mileage, depending on the year of registration of vehicles, was used and scaled up for EU 25.

The following function provides an estimation of the inter-relation between the fleet penetration rate and mileage driven by the fleet. The inter-relation depends on the age structure of the fleet, the share of mileage driven by cars of different ages and the year of first installation of a cooperative system in new cars. The relevant share of mileage driven by equipped vehicles is then calculated by a regression function using model parameters a, b, c:

( )2p*cp*bafB ++= , with

fB Mileage driven by equipped vehicles (share of benefits) [%]

a -0.029

b 1.400

c -0.003

p Fleet penetration rate [%].

Since not all vehicles are equipped with an SAFESPOT in-vehicle system only the share Bf of the possible safety benefits can be generated by the V2I bundle. Multiplying the safety effects of the bundle with the accident base for the number of fatalities and injured one gets the number of avoided fatalities and injured. Using the cost-unit rates for the avoided fatalities and injured and multiplying with the factor fB the safety benefits of the V2I bundle are given. In formal terms the safety benefits are calculated as follows:

[ ] [ ]( )IIFFBS c*I*ec*F*e*fB += , with (1´)

BS Safety benefits [EUR]

fB Mileage driven by equipped vehicles (share of benefits) [%]

eF, eI effectiveness in avoiding fatalities / injured [%]

cF, cI Cost-unit rate for fatalities / injured [EUR]

F Total number of fatalities,

I Total number of injured.

The values for the parameters are shown in the following table:




Table 44: Parameters for safety benefits

Fatalities Injured

eF F cF eI I cI

Effectiveness rate [%]

Number of fatalities [1,000]

Cost-unit rate

[Mill. EUR]

Effectiveness rate [%]

Number of injured [1,000]

Cost-unit rate

[Mill. EUR]

8.9 20.8 1.63 8.5 870 0.06 The effectiveness rates were determined in chapter 6. The following table shows the benefits of the V2I bundle.

Table 45: Safety benefits of the V2I bundle

Penetration rate

Relevant share

Avoided number of Fatalities

[Mill. EUR]

Injured [Mill. EUR]

Total benefits

[Mill. EUR]

Fatalities Injured

5.4% 7.4% 137 5,491 222.69 329.46 552.15

7.7% 10.5% 194 7,788 315.35 467.28 782.63

7.9% 10.8% 199 7,987 323.48 479.22 802.70

9.5% 12.9% 239 9,565 388.50 573.90 962.40 So far, the calculation of the safety benefits is based on the assumption that the infrastructure is fully equipped. In the following chapters it is assumed that full equipment of infrastructure is not necessary in order to cover almost all of the accidents which would be covered in case of a fully equipped infrastructure. This assumption will be discussed in chapter 7.2.2.2 and taken into account for the calculation of the infrastructure costs.

7.2.2. Costs

The cost for the SAFESPOT V2I system bundle is the sum of the costs for equipping the vehicles and the infrastructure.

7.2.2.1. Vehicle cost

The vehicle cost for the V2V bundle has been elaborated in Section 7.1.2. Because the V2I vehicles use the same basic equipment as the V2V vehicles, the V2V bundle is taken as the basis. Based on the specifications of the systems done by the technical SPs, two components are not required for the V2I implementation compared to V2V: the digital map and the Long Range Radar front. These components are implemented at the roadside unit instead. Hence two components remain: a dual frequency GPS and a warning module.

Similar to the V2V case, the cost for the components as a function of the penetration rate has to be determined with respect to economies of scale in production. Only the cost for the dual GPS depends on the penetration rate. The cost of the warning model is assumed to be fixed (i.e. independent of the penetration rate). By using equation (2) from section 7.1.2, the GPS component




cost as a function of the penetration rate can be described. The two parameters for this function – degression rate d and base costs for 100 % penetration – are identified by using the data from Table 40. This table shows the estimated costs for the GPS for two penetration rates. The cost for the GPS (component costs (cc)) as a function of the penetration rate p[-] is then given by Equation 6:

41

26201

240

240.

)ln(

).ln(

GPS

p*.

p*.)p(cc

−

−

=

= (6)

Next, the vehicle cost per year is calculated. Equation (7) shows the result. Firstly, the costs for the components are added. Secondly, these component costs are multiplied by a factor 1.05 to include the implementation cost. Finally, the costs are multiplied by the annuity rate (lifetime of 12 years and discount rate of 3%). In formal terms the component costs for year t are given by

{ { {

rateannuity

costtionimplementa

ncludemodulewarningGPS

41 10005110240 .*.*p.)p(cci

.

year

+=−

43421

(7)

7.2.2.2. Infrastructure cost

The equipment of infrastructure represents typically a major cost block of the total V2I system because the infrastructure must be equipped irrespective of the number of equipped vehicles. It is however questionable whether the entire road network would have to be equipped in order to realise the bigger part of the benefits. Actually, the accident risk on roads is not uniform. Some roads are more dangerous than others. Recently, much effort was done in order to categorise roads according to their accident risk by the Euro Road Assessment Programme (EuroRAP). Figure 27 displays exemplarily the results for Spain. Similar risk maps became recently available also for other countries.




Figure 27: Risk rating of the Spanish road network (EuroRAP)

Risk mapping provides the key to come up with working hypotheses about the relation between infrastructure equipment and coverage of accidents. The present study assumes that an infrastructure equipment rate of 50% will cover almost all of the accidents in the EU-25. This means, that an infrastructure equipment rate of 50% will achieve the same total benefits as a fully equipped infrastructure. The assumption implies a deployment strategy where the most critical sections are equipped first and subsequently the following category of roads. The less critical roads will remain unequipped. In the sensitivity and scenario analysis the case of an infrastructure equipment rate of 50% will be used as a reference case (section 7.3). Although the working assumption seems to be quite optimistic, recent data from the United Kingdom (half of the fatalities are concentrated on one tenth of the road network, according to ROAD SAFETY FOUNDATION, 2010) suggests that the concentration of casualties on a number of black spots is even higher.

The cost estimations for the infrastructure components reflect the findings from literature, external experts statements and the views of the BLADE cost estimation team (see chapter 2 for methodology). Table 46 shows the infrastructure component cost estimations for the three V2I sub-applications. These estimates are valid for large quantities and are assumed to be independent of the penetration rate. Besides these components, the application uses components that are already available in the target year 2020.




The lifetime of the infrastructure components was assumed with ten years. This is considerably shorter than the lifetime of other road infrastructure elements such as road surface, bridges and tunnels which have got an average lifetime between 20 and 50 years [FGSV (1997)]. On the other hand, the ten year period is rather long compared to other IT equipment. For instance, US sources assume for some elements of the information infrastructure only a lifetime of five years. Moreover, ten years is also quite close to the average vehicle lifetime which amounts to 12 years. Weighing the arguments for longer and shorter lifetimes and depreciation periods, we decided to base the calculations on a ten year lifetime which is applied to all infrastructure components.

Table 46: Infrastructure costs

Component Costs (incl. installation)

[EUR]

Operation and main-tenance

costs [EUR]

Lifetime [Years]

Annuity rate [-]

Yearly costs [EUR/year]

Roadside unit incl. antenna system 6,500 130 10 0.117 1,549

Existing CCTV Video Cameras 1,200 1,139 10 0.117 1,280

Automatic Ice Detection System 8,275 441 10 0.117 1,411

Laser Scanner 3,515 441 10 0.117 853

Digital Maps incl. Intersections 12 2 10 0.117 3 Sources: Hessisches Landesamt für Strassen- und Verkehrswesen; Federal Highway

Administration, USA (exchange rate dollar/EUR: 0.74); SAFESPOT SP2 internal document about hardware costs; SAFESPOT expert judgment and workshop (assumed penetration 0.5)

Intelligent Cooperative Intersection Safety – IRIS: basic application

This section provides the cost estimation for equipping 50 % of the existing intersections with the IRIS basic application. Table 47 shows the resulting costs. Approach and driving factors for cost estimation are explained in the following:

• Starting point for the determination of quantities is the number of signalised intersections in the EU-25. Coherent to the eIMPACT approach, a high level ratio of 1 signalised intersection per 1,500 inhabitants was applied. The ratio varies of course from country to country but various sources (FRIEDRICH (2009), HUBACHER ALLENBACH (2002)) provide a range between 1,200 and 2,500 inhabitants per signalised intersection. Hence, the number of signalised intersections in the EU-25 was assumed to be 300,000. The target number for the equipment is then 150,000 (50% of 300,000).

• Before 2020, a limited part of the road network will be already equipped with intelligent infrastructure, in particular Roadside Units and Laser scanner. This equipment process will start as soon as the first cooperative systems will be deployed. From a 2020 perspective, the infrastructure is already partly equipped and the investment before




2020 can be regarded as sunk costs which are not relevant for the 2020 cost-benefit assessment.

• Based on expert estimation (by a roadside equipment supplier) it is assumed that 15% of the roads are already equipped with Roadside Units. The remaining investment hence adds up to 105,000 (35% of 300,000).

• Following the same argument, it is assumed that 10% of the intersections are already equipped with laser scanners in 2020. These laser scanners will replace the magnetic road sensors which are currently in use. In terms of quantity, four laser scanners are needed per intersection. 120,000 intersections (40% of 300,000) still need to be equipped, each with four laser scanners.

• Digital maps must be available and regularly updated at each signalised intersection. Because of a missing deployment before 2020, the quantity amounts to 150,000 units (50% of 300,000).

Table 47: Infrastructure costs for the IRIS basic application for equipping 50% of the intersections

Component Quantity per intersection

Costs per intersection [EUR/year]

Quantity for EU-25 (for 50%)

Costs EU-25 [Mill.

EUR/year]

Roadside unit incl. antenna system 1 892 105,000 94

Laserscanner 4 3,412 120,000 409

Digital Maps incl. intersections 1 3 150,000 0.5

504

Hazard and Incident Warning – H&IW: Reduced friction or visibility

This section provides cost estimation for equipping 50 % of the total road length with the H&IW sub-application. Table 48 shows the components. Approach and driving factors for the cost assessment are explained as follows:

• Starting point for the determination of quantities is the length of the road network in the EU-25. Referring to INTERNATIONAL ROAD FEDERATION (2008) and eIMPACT, the road network in the year 2020 amounts to approximately 6.6 Mill. km.

• Roadside Units are required each 1 to 2 km, on average each 1.5 km, on a road stretch. 15% of the roads will be, in line with the above mentioned estimation, already equipped with Roadside Units. Another 105,000 Roadside Units are already accounted for in the IRIS application. Thus, the number of RSU amounts to 1,435,000 (which is 6.6 Mill. km divided by 1.5 km, multiplied with 35% equipment and finally subtracted 105,000 RSU at intersections). Since the Speed Alert application makes use of the same RSU, the costs are allocated to the Speed Alert table. The cost allocation to single applications does not




matter in a bundle perspective. It is only for making sure that each (shared) component is only once accounted for.

• The existing CCTV video cameras are normally used for monitoring the traffic flow (predominately at motorways) and surveillance. However, these cameras can be simultaneously used to detect fog, visibility and rain. In the situations where no cameras are available local weather-station information is utilized. It is assumed that the cost for the additions to make the existing cameras useful for SAFESPOT is 5% of the yearly camera costs, and that the average distribution of the detection equipment equals the distribution of the RSUs.

• For ice detection a special CCTV camera system is developed. Based on an assessment of the climatic conditions (Northern Europe, mountain regions) we assume that 25% of the roads in the EU-25 require an ice detection system.

Table 48: Infrastructure costs for H&IW sub-application for equipping 50% of the road net

Component Quan-tity per

RSU

Costs per RSU

[EUR/Year]


Costs EU-25 [Mill.

EUR/Year]

Roadside unit incl. antenna system 1 1,549 See SpA

Add-On to existing CCTV Video Cameras 1 64 1,435,000 92

Automatic Ice Detection System 1 1,411 385,000 543

Digital Maps incl. intersections 1 4 See SpA

635

Speed Alert (SpA): Legal speed limit

The Speed Alert application uses the same number of Roadside Units (RSU) as the H&IW application. Digital maps represent the second cost element of this application. For both elements a maximum quantity of 1,540,000 is relevant whereby for RSU the 105,000 units at intersections have to be taken into account. Table 49 shows the results of the cost assessment.

Table 49: Infrastructure costs for SpA sub-application for equipping 50% of the road net

Component Quantity per RSU

Costs per RSU

[EUR/Year]


Costs EU-25

[Mill. EUR/Year]

Roadside unit incl. antenna system 1 892 1,435,000 1,280

Digital Maps incl. Intersections 1 3 1,540,000 5

1,285

V2I bundle costs

The previously derived vehicle and infrastructure costs are added to get the total cost of the V2I bundle. As it becomes obvious from the following table, the costs




are dominated by the large block of infrastructure costs. The total bundle costs vary between 2.6 and 2.7 Mill. EUR per year.

Table 50: V2I bundle costs for considered penetration rates and infrastructure equipment of 50 %

Fleet penetration

rate [%]

Equipped vehicles

[Mill.]

Total vehicle costs [Mill. EUR

/year]

Infra-structure

equipment rate

Infrastruc-ture costs [Mill. EUR/

year]

Bundle costs

[Mill. EUR/ year]

5.4% 16.7 191.5 50% 2,424 2,615

7.7% 23.5 227.3 50% 2,424 2,651

7.9% 24.2 236.2 50% 2,424 2,660

9.5% 29.1 257.5 50% 2,424 2,681

Figure 28 represents the cost composition for the V2I bundle. It shows that more than half of the costs are associated with the Roadside units, followed by the ice detection system which accounts for one fourth of the total costs. The in-vehicle costs account for less than 10%.

Roadside Units

Laser scanner

Automatic Ice Detection System

Add-on Exsting CCTV Video Cameras

Digital maps incl. Intersections

In-vehicle costs

Figure 28: Cost composition for the V2I bundle (for penetration rate 7.9%)

7.2.3. Benefit-Cost Ratio

The BCR of the V2I bundle can be derived by using equation (1), the safety benefits (table 45) and the costs (table 50). This information is shown in table 51.

Table 51: Benefit-cost ratios for considered fleet penetration rates and 50% infrastructure equipment

Model Penetra-tion rate

Benefits [Mill. EUR/year]

Costs [Mill. EUR/year]

Benefit-cost ratio

Service model base – Private 5.4 % 552 2,615 0.21

Service model plus – Private 7.7 % 783 2,651 0.29

Business model base – Public 7.9 % 803 2,660 0.30

Business model base – Public/Private 9.5 % 962 2,681 0.36




The BCR of the V2I bundle shows fairly low values. Under the introduced assumptions for the conditions foreseen for 2020, the results range between 0.2 and 0.4. Benefit-cost ratios for which 0 < BCR ≤ 1 holds are regarded as inefficient from an economic point of view: For one Euro invested the pay off for the society is less than one Euro. These results can be explained in a number of ways, especially due to the current system design, the assumptions concerning the large-scale equipment of infrastructure and the strong relationship with the required level of vehicle penetration. Moreover, the two elements which determine the costs, i.e. quantity and costs per unit, should be also commented.

1. The current system relies on short range communication (with a maximum coverage of 1,500 m). As a consequence, a large number of Roadside units (RSU) will be necessary to cover the road network, even at a 50% equipment target..

2. The assumption of a large-scale equipment of 50 % of the infrastructure to receive benefits seems too strong. In the case study this assumption will be relaxed and the effects on costs of lower infrastructure equipment rates will be shown. Probably using an optimized, “smarter” equipment programme with nearly similar safety benefits will result in less infrastructure costs.

3. The assumption that the large-scale of equipment of 50 % of infrastructure is independent of the vehicle penetration rate seems questionable. The safety benefits are estimated using different vehicle penetration rates, whereas the infrastructure equipment rate is assumed as fixed on one scale. There is also an inter-relationship between investments in infrastructure and the prevailing vehicle penetration rate. However, it is difficult to estimate the impacts that different extents of infrastructure equipment will have on the safety effects, since the relation between the extent of infrastructure equipped and the number of accidents prevented is unclear. In the case study this problem is addressed and a theoretical approach shows the potential impact on safety benefits of changing this assumption.

4. The costs for each infrastructure component are appraised as the product of the quantity and the unit costs. Reflecting on quantity, it may be possible to establish a leaner SAFESPOT architecture in the future. This points towards the four laser scanner which are necessary to cover one intersection. In the same line of argumentation it is also advisable to look at possibilities to integrate the SAFESPOT concept into a broader communication strategy (making also use of cellular communication).

5. The estimated unit costs are too high in the light of the attainable benefits. The costs for the year 2020 were forecasted to the best available knowledge of the involved experts. However, in the context of high technology goods considerable uncertainties and also remarkable economies of scale are prevalent so that the cost perspective will look




fairly different in a couple of years. In order to become mature for deployment, the unit costs have to be pushed downwards significantly.

6. Referring again to a leaner and smarter equipment strategy, the UK results of the risk mapping in the road network (ROAD SAFETY FOUNDATION, 2010) can be used for exploring the costs and benefits of a Black Spot oriented deployment concept (see Figure 29). The underlying base parameters are delivered by the concentration of half of the fatal collisions and one third of those resulting in serious injuries on only 10% of the road network. For simplification we assume that half of the fatalities and a third of all injuries in the EU-25 are concentrated on 10% of the roads. In terms of benefits, the Black Spot concept would reduce the benefits approximately factor 3 towards a range between 220 and 400 Mill. EUR. Because of the leaner equipment, the total costs could be cut by factor 4 (infrastructure costs below 600 Mill. EUR). As a result, this helps to improve the benefit-cost ratio up to 0.5. It represents an important step in the right direction. However, the result is still considerably lower than the threshold of 1. This again reinforces the argumentation above towards a leaner system with lower unit costs.

0

500

1000

1500

2000

2500

3000

Large scale deployment Black spot concept

Benefits

Costs

Figure 29: Large scale versus black spot deployment of the V2I system

7.3. Sensitivity and scenario analysis

In order to study effects occurring with the variation of relevant parameters used for the calculations of the BCR a sensitivity analysis, a scenario analysis and a case study have been performed. These analyses aim at the following goals: First, they will provide proof of the reliability of the results of the cost-benefit analysis. Second, since the sensitivity analysis will reveal the parameters that have the greatest influence on the BCRs, hints on parameters which are critical for the deployment of the systems are provided. This can be interesting for public authorities and other stakeholders, such as OEMs and road operators, for developing strategies to promote deployment. Third, alternative scenarios based on different infrastructure equipment rates can be compared with the baseline V2I case.




Since traffic impacts have been estimated to be approximately zero, traffic effects are not taken into account within these analyses.

7.3.1. Sensitivity analysis

The sensitivity analysis will discuss the responses of BCR when the values of relevant parameters are changed. Both the V2V and the V2I bundle are considered here.

Analogous to cost-benefit analysis, a parametrical approach is used for the sensitivity analysis. To get an overview of the possible parameters which can be varied in the sensitivity analysis we present the BCR in formal terms:

[ ] [ ]( )

t,InfratS,tt

IIFFB

t

tt

Ccc*p*VS

c*I*ec*F*e*f

C

BBCR

+

+== , with

Bt Benefit

Ct Cost

fB Share of benefits

e Effectiveness in avoiding fatalities / injured (safety effect)

F Number of fatalities

I Number of injured

c Cost-unit rate for fatalities/ injured

t Considered year t (2020)

S Considered scenario (low or high)

VS Vehicle stock

p Penetration rate

cct Annualized component costs per system (discount rate: 3 %, lifetime: 12 years)

CInfra,t Annualized cost of infrastructure (discount rate: 3 %, lifetime: 10 years, fixed equipment rate 50 %)

In the following some of the parameters will be selected for variation. These parameters will be shown below.

7.3.1.1. Parameters of sensitivity analysis

Given the parameters of the BCR calculation formula the following factors can be expected to influence the benefits and costs of the CBA:

• Demand data: Market penetration of cooperative systems and vehicle stock;

• Estimated safety impacts of SAFESPOT applications;




• Cost rates: cost figures of the SAFESPOT system (in-vehicle system costs, infrastructure costs) and cost unit rates for fatalities and injured;

• Exogenous model parameters: time horizon, vehicle lifetime, discount rate.

Thus, the following parameters are chosen for the sensitivity analysis:

• Effectiveness in avoiding fatalities: eF

• Effectiveness in avoiding injured: eI

• Number of fatalities: F

• Number of injured: I

• Cost-unit rate for fatalities:cF

• Cost-unit rate for injured: cI

• Vehicle stock: VS

• In-vehicle system costs: cc

• Infrastructure costs for V2I: CInfra

The analysis focuses on the effects caused by changing a single variable parameter and fixing the other parameters. The variation of a parameter will be done on a very small scale, adding and subtracting 10% of the value of a parameter. Because the penetration rate is critical for realizing benefits of cooperative safety systems a comparison of the results will be made for the following penetration rates: 2 %, 5 % and 10 %.

7.3.1.2. Results of the sensitivity analysis of the V2V case

In the first row of Table 52 the calculated BCR for V2V is shown for penetration rates 2%, 5% and 10%. The rows below show the effects of variation within the range of -10 % to +10 % for every parameter. Then, the cells of the table show the relative change in BCR caused by a slight variation of one parameter.




Table 52: Results of sensitivity analysis for the V2V bundle

Parameter Parameter variation

Effect on BCR in case of shifting the parameter value by -10% / +10%

BCR V2V

0.8 1.0 1.2

Penetration rate

2% 5% 10%

Effectiveness in avoiding fatalities -10% - -3.9% -3.9%

+10% - 3.9% 3.9%

Effectiveness in avoiding injured -10% - -6.1% -6.1%

+10% - 6.1% 6.1%

Fatalities (accident data) -10% - -3.9% -3.9%

+10% - 3.9% 3.9%

Injured (accident data) -10% - -6.1% -6.1%

+10% - 6.1% 6.1%

Cost-unit rate for fatalities -10% - -3.9% -3.9%

+10% - 3.9% 3.9%

Cost-unit rate for injured -10% - -6.1% -6.1%

+10% - 6.1% 6.1%

Vehicle stock -10% - 11.1% 11.1%

+10% - -9.1% -9.1%

In-vehicle system costs -10% - 11.1% 11.1%

+10% - -9.1% -9.1%

Life expectancy -10% - -6.3% -6.6%

+10% - 5.7% 6.0% The results of the sensitivity analysis for the V2V bundle show same values for several parameters. At first glance this seems surprising. However, the reason for this outcome can be found in the algorithm of the BCR equation shown above, on which the sensitivity analysis has been based. According to the composition of the equation there are four groups of parameters which differ by the intensity of reaction to variation of +/-10 %:

• Parameters which show benefits in avoiding fatalities, i.e. effectiveness in avoiding fatalities, number of fatalities, cost unit rate fatalities.

• Parameters which show benefits in avoiding injured, i.e. effectiveness in avoiding injured, number of injured, cost unit rate injured.

• Parameters which are used to calculate vehicle cost, i.e. vehicle stock, in-vehicle costs, life expectancy.

• Infrastructure costs (which are not applicable for the V2V case).




The highest sensitivity appears with the vehicle cost parameters, followed by the injury benefit parameters. The lowest sensitivity is caused by the variation of the fatality parameters. However, variation of the penetration rate has no major effect on the sensitivity of the BCR.

Moreover, variations (Plus and Minus) in the numerator of the equation result in equivalent relative changes of the effect on BCR, whereas variations in the denominator cause different relative changes in the result.

7.3.1.3. Results of the sensitivity analysis of the V2I case

For the V2I scenario the same calculations have been made, although some variables have changed due to the nature of the V2I bundle (e.g. estimated impacts and infrastructure costs).

Table 53: Results of sensitivity analysis for V2I bundle

Parameter Parameter variation

Effect on BCR in case of shifting the parameter value by -10% / +10%

BCR V2V

0.08 0.2 0.37

Penetration rate

2% 5% 10%

Effectiveness in avoiding fatalities -10% -4.1% -4.1% -4.1%

+10% 4.1% 4.1% 4.1%

Effectiveness in avoiding injured -10% -5.9% -5.9% -5.9%

+10% 5.9% 5.9% 5.9%

Fatalities (accident data) -10% -4.1% -4.1% -4.1%

+10% 4.1% 4.1% 4.1%

Injured (accident data) -10% -5.9% -5.9% -5.9%

+10% 5.9% 5.9% 5.9%

Cost-unit rate for fatalities -10% -4.1% -4.1% -4.1%

+10% 4.1% 4.1% 4.1%

Cost-unit rate for injured -10% -5.9% -5.9% -5.9%

+10% 5.9% 5.9% 5.9%

Vehicle stock -10% 0.5% 0.7% 1.0%

+10% -0.5% -0.7% -1.0%

In-vehicle system costs -10% 0.5% 0.7% 1.0%

+10% -0.5% -0.7% -1.0%

Life expectancy -10% -0.3% -0.3% -0.5%

+10% 0.2% 0.3% 0.4%

Infrastructure costs -10% 10.5% 10.3% 9.9%

+10% -8.7% -8.5% -8.3%




The results of the sensitivity analysis for the V2I bundle show the same groups of parameters which differ by the intensity of reaction to variation of +/-10 % as in the V2V case. However, for the V2I bundle, the highest sensitivity appears obviously with the infrastructure costs, whereas the injury and fatality benefit parameters show a similar reaction pattern to parameter changes as in the V2V case. The lowest sensitivity is caused by the variation of the vehicle cost parameters. The different figures of the V2V case and the V2I case depend exclusively on the different effectiveness in avoiding fatalities and injured. Similar to the V2V case, the variation of the penetration rate has no major effect on the sensitivity of the BCR.

7.3.2. Scenario analysis

The current assumption of large-scale infrastructure equipment for the baseline V2I case (50 % of the road net, equipping every 1.5 km stretch of road) has rather strong impact on the level of costs for the required infrastructure. However, when relaxing this assumption and taking infrastructure equipment rates as a variable, the estimation of infrastructure costs and resulting safety impacts is complex. The scenario analysis discusses some conceptual ideas for doing this variation and provides the possible effects on costs and safety benefits. This scenario approach will be further elaborated inside a case study for the Netherlands.

To get a better understanding of the influence of infrastructure costs on the BCR four different scenarios were designed. The scenarios create insight into the effect of the current high infrastructure investment on BCRs and show how, by using “smarter”, more effective investments compared to equipping at random, this ratio can be improved.

The following scenarios will be considered:

• Scenario 1 is based on the 80:20 rule (also known as the Pareto principle) such that 80 % of the safety effects will be reached with only 20 % of the costs. This scenario is more an optimistic thought experiment which assumes that infrastructure is installed on black spots with high accident frequency where the highest level of safety impacts can be realized. However, identification of the most effective points of infrastructure investment is difficult. In table 54 the change in safety effects is shown given the number of equipped intersections and stretches of the road. The total costs of this scenario are 864 Mill. EUR. All safety applications selected for the V2I bundle will be taken into account in this scenario.

• Scenario 2 assumes that only the highways will be equipped with the necessary RSUs. Consequently, only the hazard and incident warning application (H&IW) and the SpeedAlert application will be applied and create safety effects. The resulting total costs for this scenario are 70.5 Mill. Euro which are based on the assumption that only 8 % of the estimated safety impacts can be realised. This assumption takes the number of fatalities on highways and the ratio of kilometres of highways compared to the rest of the road network into account.




• Scenario 3 purely focuses on intersections. The equipment of all intersections will result in costs of 1,879.7 Mill. Euro and in a reduction of the safety effects to 50 %. This assumption is based on current accident data where it is shown that in the EU, approximately 50 % of the accidents happen on intersections. Consequently, only the Intelligent Cooperative Intersection Safety (IRIS) sub-application will be applied.

• Scenario 4 is based on the assumption that only the European trunk road network and the urban area intersections are equipped. Thus, 25 % of the whole number of intersections is equipped. This scenario will result in costs of 1436.9 Mill. Euro and in a reduction of the safety impacts to 35 % (based on the number of fatalities occurring in urban areas). All safety applications selected for the V2I bundle will be taken into account in this scenario.

For all these scenarios the lower infrastructure equipment rate will result in reduced equipment costs and, however, in reduced safety benefits. The estimations for infrastructure costs and resulting reductions of safety impacts are given in the following table. For the last three scenarios the reductions in safety impacts are based on available statistical data from the SafetyNet project.

The estimation of the safety impacts are based on the following figures:

• The average number of road accident fatalities in EU-14 occurring on motorways is 8 % of the total number of accidents [SAFETYNET 2008a]. Therefore, in scenario 2 safety impacts are reduced to 8%.

• The average number of road accident fatalities in EU-14 occurring on junctions is 20 % of the total number of fatalities [SAFETYNET 2008c]. Since in scenario 3 all relevant intersections will be equipped (larger intersections and intersections with traffic lights), it is assumed that 50% of the fatalities at junctions can be reduced.

• The number of fatalities in urban areas is 35 % of the total number of fatalities [SAFETYNET 2008b]. In scenario 4, therefore, 35% of the safety impacts can be realised, which represents the part in urban areas.

The table below gives an overview of the four scenarios showing the costs of the V2I bundle when equipping the infrastructure at lower levels than with the baseline V2I bundle, and the corresponding safety benefits which also decrease due to the reduced infrastructure equipment.




Table 54: Input parameters for validation scenarios

Scenario Road Type Km road or number of

RSU

Bundle costs [Mill. Euro]

Safety impacts compared to baseline V2I case[%]

Scenario 1 Highways (km) 46,390

864 80% Total of safety

impacts Intersections (# RSU) 61,818

Scenario 2

Highways (km) 78,965

70.5 8% Only highways Intersections (# RSU) 0

Scenario 3

Highways (km) 0

1879.7 50% Only intersections Intersections (# RSU) 309,093

Scenario 4

Trunk roads (km) 425,000

1,436.9 35% 25% of intersections,

only major trunk roads Intersections

(# RSU) 77,273 These scenarios were used as input for the calculation of BCRs for the V2I based system bundle. The effects which these scenarios will have on the BCR are shown for various penetration rates. In the following table only the relative changes in the BCR compared to the calculations made for the baseline case are shown:

Table 55: Results of the relative change of BCRs for different infrastructure scenarios

Penetration rate Scenario 1 Scenario 2 Scenario 3 Scenario 4

2 % 171% 67% -17% 4%

5 % 158% 15% -19% 0%

10 % 145% -9% -20% -5%

15 % 137% -22% -21% -7%

20 % 131% -29% -22% -9%

25 % 127% -33% -22% -10% In Figure 30 the resulting effect on the BCR of the scenarios is shown in terms of the relative difference compared to the baseline V2I case. Scenario 1 has a large positive effect on the BCR for the different penetration rates. The BCR of scenario 1 is clearly larger than that of the baseline V2I case. Scenario 2 and 4 show a positive effect on the BCR only for low penetration rates. For scenario 3 only a negative effect on the BCR can be observed. Of course, these results are hypothetical, but they show that an investment in infrastructure, which is “smarter” than the large scale equipment of the baseline case, can result in a considerable increase in the BCR.




-50

0

50

100

150

200

2 5 10 15 20 25

Rel

ativ

e d

iffer

ence

of B

CR

, in

%

Fleet penetration rate, in %

Scenario 1

Scenario 2

Scenario 3

Scenario 4

Figure 30: Effect on BCR of the scenarios 1 to 4 (difference relating to the baseline V2I case)

Scenario 2, 3 and 4 are based on different assumptions concerning which part of the infrastructure is equipped. In scenario 3 with an equipment of intersections only, the corresponding reduction of safety benefits outweighs the reduction of infrastructure costs for all vehicle penetration rates.

Only for the low penetration rates in scenarios 2 and 4 a positive effect on the BCR results: Equipment of RSU only at highways in scenario 2 and at trunk roads and urban intersections in scenario 4 decrease infrastructure costs and will generate safety benefits which are higher than the costs. So, for low penetration rates a positive effect compared to the baseline V2I case results. Of course, this positive effect will not arise, if the penetration rate in the vehicle fleet is too low (< 1 %). At higher penetration rates the increase of the in-vehicle component costs is stronger than the increase of the safety benefits such that a negative impact on BCR results.

The results allow the conclusion that, in addition to infrastructure costs, the vehicle penetration rate still has an impact on the BCR since the costs are apparently not only driven by high infrastructure costs but also by the penetration rate of vehicles. Therefore, concerning deployment of a V2I based system, it seems necessary to discuss a smart combination of equipment of vehicles and of infrastructure such that high safety benefits are captured with fewer costs.

7.3.3. Case study of the Netherlands

To get a better understanding of the necessary investment in infrastructure to capture the safety benefits of the V2I based system in a cost efficient way a case study is performed. The case study focuses on the Netherlands and will provide




first hints about the relation between the safety impacts and the location and level of equipment of the infrastructure to indicate also directions for further research.

Looking at the results from the SafetyNet project (2008a) the relative share of accidents for various types of roads are very comparable to one another. For example a comparison between rural roads and highways, between urban and non-urban roads and between intersections and stretches of roads showed an approximate relationship of 50:50 of the two compared types. In other words, a distinction based on accident data about different road types on an EU-level does not seem to give a clear indication for effective investments on different road types in the Netherlands.

Therefore, in order to define the minimal necessary amount of “equipped” infrastructure to realize sufficient safety benefits the analysis performed in the case study used a more theoretical approach compared to the one used in the scenario analysis in chapter 7.3.2, where four scenarios were run to identify the potential effect of a less than 50 % equipment of infrastructure. For the scenarios which are based on the aggregated accident data from the SafetyNet project the assumption is made (with exception of the first scenario) that the safety impacts behave linear with the equipment rate of infrastructure.

For the case study, the Dutch situation is used by taking detailed data of accidents for the past 20 years into account. In the following the assumption is made that the location of black spots of accidents is perfectly predicted and that by equipping these locations with RSU all accidents could by avoided.

The following scenarios are defined in order to get a better understanding of the infrastructure equipment and correlated infrastructure costs:

1. Equipment of the accident black spots on the highways and intersections where an accident occurred in 2008

2. Equipment of the same accident black spots as in the year 2008, but now preventing accidents for a period of 20 years

3. Equipping accident black spots at important roads and intersections of the past 20 years, and assuming preventing accidents for the next 20 years.

Scenarios 1, 2 and 3 are purely theoretical calculations for which the following parameters are used:

Table 56: Scenario input for the case study for the Netherlands

Scenario Road Type Km road or number of

intersections

Bundle costs [Mill. EUR]

Estimated avoided prevented fatalities

Scenario 1 Highways 4,881

3.1

630

Intersection 4,016 8,267

Scenario 2 Highways 4,881

3.1

12,600


Scenario 3 Important roads 97,620

122

12,600





For all scenarios the number of prevented fatalities is adapted, as well as the vehicle stock (7.2 mill. vehicles). The number of equipped roads is based on the number of fatalities (630) and the number of equipped intersections on the number of hospitalised fatalities (8267) for 2008. The results can be found in the figure below, where the baseline V2I case is the reference scenario with large-scale infrastructure equipment which has already been presented in section 0.

0,0

0,5

1,0

1,5

2,0

2,5

3,0

0 10 20 30 40 50 60 70 80 90

Benefit-cost rate BCR

Fleet penetration rate, in %

Baseline V2I case

Scenario 1

Scenario 2

Scenario 3

Figure 31: BCR for different scenarios and penetration rates (case study)

The above Figure 31 describes the BCR of the different scenarios and the baseline V2I case. All the scenarios show higher BCRs than calculated in the baseline V2I case. However, the result we derive here is based on the assumption of perfectly predicting all black spots of accidents and preventing all these accidents. These assumptions are, of course, idealistic, but still show the impact of a smarter concept of equipping the infrastructure compared to large-scale equipment of the infrastructure.

Scenario 3 shows the highest BCR values which are already from a penetration rate of 2 % on above the BCR of 1. It is based on the assumption of an encompassing and effective equipment rate of infrastructure for the next 20 years. Scenario 1 and 2 are less efficient. Here, a BCR of 1 will be exceeded with penetration rates above 44% for scenario 1, and respectively above 36% for




scenario 2. In the baseline scenario a BCR of 1 will be reached not before a penetration rate of 62%.

To sum up, it can be seen that a partially equipped infrastructure can also create large benefits. Thus the scenarios underline the importance of the identification of accidents black spots. This result can also have some implication for the deployment strategy of cooperative systems and the question whether to start with a V2V based system or a V2I based system. If the vehicles are already equipped with V2V based systems and can communicate with one another, the additional safety impacts of subsequent infrastructure equipment will be small. The results at least indicate that given a share of equipped vehicles and equipping the infrastructure at black spots can increase the BCR of cooperative systems based on V2V and V2I further.




8. Stakeholder financial analysis

8.1. Break-even analysis

The profitability of the SAFESPOT system from the point of view of the user is addressed by a break-even analysis.

The following data are used for the calculations:

Table 57: Data for the break-even analysis

Traffic data (year = 2020) (see chapter 4.3)

Vehicle mileage bn. Km 3,716

Vehicle stock Mill. vehicle 307

Annual mileage per passenger vehicle Km 11,484

Annual mileage per goods vehicle

Km 16,005

Accident data without SAFESPOT (year = 2020) (see chapter 4.2)

Fatalities 1,000 20.8

Injured 1,000 873.7

Safety data with SAFESPOT (see chapter 6.1.3.3)

V2V: effectiveness rate (fatalities)

% 7.1

V2V: effectiveness rate (Injured)

% 7.3

V2I: effectiveness rate (fatalities)

% 8.9

V2I: effectiveness rate (Injured) % 8.5

Cost-unit rates (year = 2020)

Fatality 1,000 EUR 2,005

Injured 1,000 EUR 92

Annuity rate (discount rate 8 %, lifetime 12 years) (see chapter 4.2.1.2) 0,13

The cost-unit rates for personal losses caused by accidents are based on the willingness-to-pay approach (WTP). Conceptually, this approach differs from the damage costs-approach which was applied for the social cost-benefit analysis. Willingness-to-pay however represents the appropriate approach for user-based break-even analyses. The WTP reflects how individual users value road safety improvements. It is clear that risk averse users have a higher willingness-to-pay than the damage cost-based value indicates because the risk value is quantified by the WTP approach.

The used WTP values are from BICKEL et al. (2005) and were updated to the year 2020. In contrast to BICKEL et al., in SAFESPOT, however, no distinction between severely and slightly injured was made. Thus, the ratio of slightly injured




to severely injured is used to calculate the cost-unit rate for injured. The number of slightly and severely injured for the year 2020 was estimated by the project eIMPACT where it was shown that 18.9 % of all injured are severely injured, and the remaining 81.1 % are slightly injured. Using those data, the cost-unit rates for the year 2005 for fatalities and injured were calculated. These cost-unit rates are updated with an inflation rate of 2 % to the year 2020: Thus the cost-unit rates used here are for fatalities 2,004,508 EUR and for injured 92,005 EUR.7

Since, traffic effects and, correspondingly, effects on fuel consumption are estimated as zero in SAFESPOT, the benefits are only based on unit-cost rates for avoided fatalities and injured.

The break-even analysis is done in two steps. In the first step the safety benefits of the SAFESPOT system per driven kilometre per year are calculated and in the second step the end-market price is estimated and a critical mileage is calculated.

If every vehicle were equipped with the V2V SAFESPOT system, 7.1 % of all fatalities could be avoided (= 1,470) and 7.3 % of all injured (= 63,324). These values are divided by the total vehicle mileage in 2020 (= 3,716.4 bn. km) and then multiplied with the corresponding cost-unit rates. The result is the benefit for driving 1 km with the safety effect of the V2V system.

In formal terms, the safety benefits per 1,000 vehicle kilometres are defined as follows:

00010001 ,*c*vm

I*e,*c*

vm

F*eB /

IF

Fkm,S += , with

BS,km Safety benefits per 1,000 km

e Effectiveness rate in avoiding fatalities/ injured

F Number of fatalities

I Number of injured

c Cost-unit rate for fatalities/ injured

vm Total vehicle mileage in 2020

Using the above data, the average driver equipped with the SAFESPOT V2V system would spend 0.79 EUR per 1,000 km for avoiding fatalities and another 1.58 EUR per 1,000 km for avoiding injured. In sum, the driver would spend 2.37 EUR per 1,000 km driven. Similar calculations show that for the V2I bundle, the user on average would spend 2.84 EUR for driving 1,000 km with the bundle.

The critical mileage of the SAFESPOT system a market price is determined on an estimated market price for three different equipment rates: 5 %, 10 % and 20 %. The price estimation is based on the costs of the cooperative systems and it is assumed that market prices equal three times cost. The cost estimation of the

7 Since these cost-unit rates are based on subjective estimations they can differ between the end-users. So they are assumed to be average values.




V2V and V2I bundle then follows the cost-benefit calculation with respect to economies of scale (see 7.1.2 and 7.2.2). For the V2I bundle only the in-vehicle system is considered. The estimated market prices for the V2V solution are then in a range between 325 EUR (linked with a penetration rate of 20 %) and 440 EUR (linked with a penetration rate of 5 %). The prices for the V2I bundle are in the range between 165 EUR (5 % market penetration) and 90 EUR (20 % market penetration).

For determining the critical mileage, the market price is multiplied with the annuity rate. The result is the annual payment for equipping the vehicle with the system. A mark-up of 15 EUR for the digital maps update is included. This sum is then divided by the benefit for driving 1,000 km with the system. The result is the critical mileage.

Using the safety benefit per 1,000 kilometre mileage per year and the market price of the cooperative system, the critical mileage is given by the formula

km,S

Update

Updatekm,S

B

MAPSAR*P,*km

MAPSAR*P,*km*B

+=

⇔+=

0001

0001

, (8)

with

BS,km Safety benefits per 1,000 km

P Market Price for the V2V and V2I solution

AR Annuity rate

MAPSupdate Annual up date of digital maps

km Critical vehicle mileage driven.

Used data and the calculation results are displayed in Table 58.

Table 58: Break-even analysis: calculation parameters and results

V2V V2I

Safety benefits per 1,000 km

Effectiveness rate fatalities 7.1 % 8.9 %

Potential avoided fatalities 1,470 1,851

Effectiveness rate injured 7.3 % 8.5 %

Potential avoided injured 63,324 74006

Safety benefits per user [EUR/1,000 km] 2.39 2.85

Critical mileage per year

Critical mileage (5 % penetration rate) 30,964 12,991

Estimated market price of in-vehicle system [EUR] 440.00 165.00








For the V2V system bundle the critical mileage is in the range between 24.5 and 31.0 thousand kilometre per year. For the V2I system bundle only the in-vehicle system bundle was considered. Up to now, there is no information about the infrastructure costs which have to be borne by the user. Thus, it is assumed that there are no infrastructure costs borne by the end users. Given this assumption the critical mileages are relatively low. They range between 9.5 and 13.0 thousand kilometre per year.

With respect to the EUROBAROMETER study the average user has an annual mileage of about 15,000 km [EUROBAROMETER 2006]. In the study it is indicated that only 6 % of the drivers drive more than 30,000 km per year and about 15 % more than 20,000 km per year. Comparing this information to the calculated critical mileage for the SAFESPOT cooperative system one gets some hints about which share of the drivers will find the system worthwhile. Of course, we have to assume that driver behaviour is relatively constant in the next decade.

Using this assumption the V2V bundle would be worthwhile for about a share of 6 % of the drivers for a penetration rate of 5 %. For the higher penetration rates of 10 and 20 % this share would be in the range between 6 and nearly 15 %. For the V2I case the share of drivers would be much larger: For the critical mileage calculated for three considered penetration rates the system would be worthwhile for about 65 % of the drivers.

So, in contrast to the V2V case, for the V2I case a considerably market potential seems to exist. The reason for this may be relatively low in-vehicle system costs and the assumption of using the infrastructure for free. Having this market potential in mind for the V2I case, a funding model can be envisaged in which, for example, end users of cooperative systems participate in funding the equipment of infrastructure by, the payment of fees. How this can work is shown in the next chapter.

8.2. Financial assessment for road operators / public authorities

8.2.1. Road operators

8.2.1.1. Private road operator

In the case of a private road operator, a discount rate of 9.29 % is used instead of 3 % such that the annuity rate changes from 0.10 to 0.16 (see chapter 2.4.2). To calculate the net-market price of roadside units in the case of private road operators the factor 2.5 is used. Given the annual net costs of equipping the infrastructure with sensors and communication tools and using the factor 2.5, one calculates 7,580 Mill. EUR as annual costs for the private road operator.




If we assume that the annual net costs per vehicle equals a fee paid by the drivers for using the infrastructure, then, for the assumed penetration rates (5 %, 10 %, 20 %), the fee per vehicle in formal terms is given as:

VF*p

CFee Infra= , with (9)

Fee Annual fee for using the infrastructure in the V2I case

CInfra Annual infrastructure costs (7,580 Mill. EUR)

VF Vehicle fleet (307 mill.)

p Fleet penetration rate (5 %, 10 %, 20 %)

Table 59 shows the annual fees per vehicle.

Using equation (8) and adding the fee to the in-vehicle system market price (e.g. 90 EUR for a penetration rate of 20 %) we can calculate the critical mileage in the case of the V2I bundle in formal terms by

km,S

Update

Updatekm,S

B

MAPSAR*PFee,*km

MAPSAR*PFee,*km*B

++=

⇔++=

0001

0001

(8´)

with

BS,km Safety benefits per 1,000 km (= 2.85 EUR)

P Market Price for the V2I case

AR Annuity rate

MAPSupdate Annual up date of digital maps (= 15 EUR)

km Critical vehicle mileage driven.

Table 59 shows the annual mileage for the V2I case for three penetration rates.

Table 59: Critical mileage in case of complete private (user) funding of private infrastructure

Penetration rate Market price of

in-vehicle system [EUR]

Annual fee [EUR] Critical mileage

[km]

5 % 165.00 494.00 186,908

10 % 120.00 247.00 97,816

20 % 90.00 123.00 52,926

8.2.1.2. Public road operator

If the infrastructure equipment is operated by public authorities, the expected return on investment is not as high as for private operators (3 % instead of 9.29 %). Thus, the infrastructure costs per year are less than for the private road operator. The costs can be determined, again, by multiplying the annual




infrastructure costs (see chapter 7.2.2) with the factor 2.5. This results in a net-cost per year of 6,060 Mill. EUR.

Using equation (9) the average costs per vehicle and the corresponding fee is given. Inserting in equation (8) the critical mileage can be calculated for the case that the infrastructure is operated publicly but financed by the users. See table 60 for the results.

Given the annual vehicle costs of 27 EUR (=12 + 15) and an annual fee of 99 EUR for a penetration rate of 20 %, the total costs of using a cooperative system in the V2I case are 126 EUR for the user. Dividing these costs by safety benefits per 1,000 km, as calculated above (2.84 EUR/1,000 km), the critical mileage amounts to 43,263 km per year.

Table 60: Critical mileage with complete private (user) funding of public infrastructure



Annual fee [EUR] Critical mileage

[km]

5 % 165.00 395.00 150,298

10 % 120.00 197.00 79,122

20 % 90.00 99.00 43,263 Comparing the results with an estimated annual mileage of about 15,000 km, for the average driver given by the EUROBAROMETER study, it seems fairly unrealistically that the infrastructure equipment could be operated economically, independently, whether the infrastructure equipment is operated by private or public road operators.

To sum up, given the estimated market prices for the in-vehicle systems and fees for using the infrastructure the V2I bundle does not seem to generate enough benefits on the part of the average driver to make it attractive for a large enough fraction of the drivers.

8.2.2. Public authorities

The public authorities can profit from VATs which are generated by equipping vehicles with the in-vehicle systems and the infrastructure with sensors and communication tools. In the following the additional VAT is estimated for the year 2020 by deployment of cooperative systems.

For the ongoing calculations the vehicle stock in 2020 is estimated as 307 Mill. vehicles, the rate of VAT is set at 20 %, the discount rate for public authorities 3 %, and the lifetime for vehicles is assumed as 12 years, resulting in an annuity rate of 0.10046.




8.2.2.1. V2V bundle

In the case of a V2V bundle only the in-vehicle systems are relevant. The VAT which can be realised for the estimated market prices (see chapter 8.1) are calculated for the penetration rates 5 %, 10 % and 20 %.

Using the above data the VAT in formal terms is given by

( ) VF*p*AR*MAPAR*P*taxVAT += , with

VAT Value added tax

tax Tax rate (= 20%)

AR Annuity rate (= 0.10046)

P In-vehicle market price

MAPSupdate Annual update of digital maps (= 15 EUR)

VF Vehicle-fleet (307 Mill. vehicles)

p Penetration rate

If the public authorities plan to support the V2V bundle by respecting fiscal neutrality, they can spend up to 585.14 Mill. EUR, given a penetration rate of 20 %, for example. If they spend more, the subsidies will not be fiscally neutral. However, part of the subsidy will finance itself since by providing some rebate on the vehicle tax, for example, the penetration rate will increase resulting in higher VAT revenues given the tax rate. Thus for determining a fiscally neutral level of funding for supporting actions, the interaction between penetration rate and VAT on in-vehicle cooperative systems has to be considered.

8.2.2.2. V2I bundle

For the V2I bundle two scenarios were considered. In the first one, the complete infrastructure costs are borne by the public authorities (i.e. completely subsidized). In the second scenario, there is only a partial subsidy of the infrastructure equipment.

In the first scenario, only the in-vehicle costs are relevant. This scenario is calculated for a penetration rate of 5 % and the corresponding market prices of the V2I bundle (e.g. 5 % penetration rate; market price: 165 EUR). The calculations are done similar to the above V2V case. See table 53 for the results on VAT earnings.

In the second scenario there are only partial subsidies for the infrastructure equipment. The infrastructure equipment is financed and operated by a private road operator. Part of the equipment costs is borne by the user and, thus, the VAT is linked to the equipment costs of both in-vehicle systems and infrastructure equipment usage. The other part which is not borne by the end-users is subsidized by public authorities. This part is not VAT-relevant.




This scenario is considered here because it seems in some respects realistic that the public authorities will subsidize part of the infrastructure for example for goal achievement in the field of road safety but also respecting constraints on the public budgets.

Thus for example, it is assumed, that for the usage of the infrastructure equipment a lump sum expense of 230 EUR has to be paid by the user. The lump sum expense is below the above calculated fee for a complete private financing of infrastructure and above the in-vehicle market price of the V2I bundle, which seems realistic.

To calculate the sum of VAT earnings the additional VAT earnings generated by the fee have to be added to the tax earnings based on the in-vehicle market price of the V2I bundle. So, firstly, the usage expense has to be annualised and then the additional VAT per user can be calculated:

0.10046*Euro 23020Euro 604 *%, = .

Given this additional tax earning per vehicle for a penetration rate of 5 %, the additional VAT earnings amount to 70.61 Mill. EUR, for 10 % to 141.22 Mill. EUR, and for 20 % to 282.44 Mill. EUR.

The sum of the VAT earnings based on the in-vehicle cooperative system market price and on the usage lump sum expense borne by the end-user is also shown in the following table 61:

Table 61: VAT earnings (in Mill. EUR) for different scenarios

Scenario

Penetration rate

5 % 10 % 20 %

V2V

VAT earnings [Mill. EUR] 181.75 300.59 585.14

In-vehicle market prices [EUR] 440.00 338.00 325.00

V2I

VAT earnings [Mill. EUR] –completely subsidized infrastructure 50,89 74,02 111,03


VAT earnings [Mill. EUR] – part of infrastructure financed by users 121,83 215,89 394,78

User fee [EUR] 230.00 230.00 230.00 The VAT earnings are calculated for a given level of market penetration. Of course, as a result of subsidies on market prices or expenses for infrastructure usage the penetration rate will change. Thus for determining the level of tax earnings, again, the interaction between penetration rate and VAT on in-vehicle cooperative systems has to be considered.




9. Conclusions

WP6.5 provides a socio-economic assessment of the SAFESPOT cooperative systems. Given the goals of the sub-project BLADE, the assessment concentrated on two extreme solutions of the SAFESPOT system based on the main technological concepts, V2V and V2I communication.

The results of the socio-economic assessment can be summarized by the following main findings:

1. The methodology for socio-economic impact assessment of IVSS was successfully adapted to cooperative safety systems. The core of the used assessment methodology is a cost-benefit analysis (CBA). In addition to the society ´s view on the profitability of cooperative systems the profitability is also viewed from the point of view of vehicle drivers, road operators and public authorities (stakeholder analysis).

Some adaptations and further development of the methodology were done because of the special requirements of the project. These developments particularly concerned the assessment of system bundles in contrast to single safety applications. For the assessment, cost synergies and possible inter-relationships and complementarities in the safety effects had to be identified and assessed.

The benefits to the users of cooperative systems which are based on V2V communication depend critically on the existence of a sufficient number of other users. Thus, a reliable estimation of market penetration of cooperative systems for the year 2020 was undertaken as an important input for socio-economic assessment of the SAFESPOT bundles. This estimation was also made with respect to different business and service models concerning the financing of the systems and services provided.

Since cooperative systems are not on the market or installed only to a very small scale in the vehicle fleet we have to expect economies of scale in costs of the applications. This important point for the cost-benefit analysis was addressed by a process of cost estimation (including a workshop with experts from other SAFESPOT sub-projects). Depending on a forecast of the market penetration of these systems, economies of scale were estimated for the single components of the systems (on-board units, sensors, navigations tools, etc.).

In addition, the approach succeeded to meet some restrictions on the availability of data with regard to technical specification of applications, test site results, and cost data. The functional design of the applications considered in the socio-economic assessment may slightly deviate from that of the final specification by the technical sub-projects.

2. The impact assessment has found significant safety effects. Road safety is improved by the considered SAFESPOT systems.




The approach used for safety impact assessment is based on work that has been performed e.g. in eIMPACT [BENZ et al. 2008, p. 31-36]. The approach aims to capture all possible effects – including unintended and intended behavioural adaptations – of cooperative systems on driving behaviour and on accidents in a systematic manner.

Since in SAFESPOT safety impacts of bundles of applications were assessed the problem of overlapping of safety effects concerning prevented accidents of a specified accident type can arise. Thus, to avoid double counting, only one application per specified accident type could be considered. In this respect our impact estimation was careful, and safety effects achievable in real traffic may be higher.

For the V2V bundle the safety impacts derived from the three applications are independent and can be summed up to get the total safety impact: Lateral Collision application (LATC), the Road Departure application (RODP) and the Longitudinal Collision application (LONC). The following table shows the total safety impact, assuming full penetration of the V2V system in the vehicle fleet:

Table 62: Estimates for behavioural mechanism effects of the V2V bundle

Fleet penetration

rate [%]

Safety Impact


V2V bundle 100 -7.1 -7.3 For the V2I bundle the safety impacts are also based on three applications, but two of the applications have been identified as being not independent with regard to safety impacts. Therefore only the two independent applications have been considered for the impact estimation: Intelligent Cooperative Intersection Safety system (IRIS), and the Speed Alert (SpA). The H&IW sub-application is included in the SpA application. The total safety impact of the V2I system shown in Table 63 is based on the assumption of full penetration in the vehicle fleet and full equipment of the infrastructure.

Table 63: Estimates for behavioural mechanism effects of the V2I bundle

Fleet penetration

rate [%]

Safety Impact


V2I bundle 100 -8.9 -8.5 Besides safety impacts, no significant traffic and environmental impacts can be shown for the SAFESPOT systems at penetration rates estimated for 2020. This finding is in line with the impact results achieved in the projects eIMPACT and CODIA.

However, non-safety impacts in terms of traffic flow, fuel consumption and resulting emissions may arise for more advanced penetration rates and especially for systems which do not have an exclusive focus on safety




improvements. For example, the CVIS project has found such traffic effects for their applications which have been primarily designed for traffic optimization.

3. The CBA leads to acceptable BCRs for the V2V bundle, whereas the efficiency of the V2I bundle could not be proved.

These results were derived using the safety impacts and the estimations for the cost of the vehicle system components and the costs of the infrastructure equipment (roadside units). The benefits are calculated using the safety impacts and the expected number of avoided fatalities and injured valued with well-accepted cost unit rates.

Then, the benefits and costs are determined by a function which depends on the penetration rate. The penetration rate is a result of the underlying business and service models which have been developed in cooperation with the BLADE work package “Business models”.

The following penetration rates were taken from the underlying business and service models and used in the cost-benefit assessment of the V2V bundle: 4.2 %, 6.1 %, 6.7 %, and 8.7 %. These numbers represent the estimated penetration rates of the system in the entire vehicle fleet in the year 2020, assuming that the deployment in new vehicles starts in the year 2015.

Table 64: Benefit-cost-ratios for the considered business/service models of the V2V bundle

V2V bundle

Business/service model Penetration rate BCR

Service model base – private 4.2 % -

Business model – public/private 8.7 % 1.1

Service model plus – private 6.1 % 1.0

Business model base – public 6.7 % 1.1 The lowest considered penetration rate is below the critical value 5 % for which no benefits of the V2V bundle arise. Former studies showed that V2V systems can only work properly for penetration rates above 5 %. Below 5 % the probability of meeting another equipped vehicle is too low.

For the penetration rates higher than 5 %, the BCR is between 1.0 and 1.1. Thus, from a social point of view, investing in the V2V based SAFESPOT system is profitable.

For the cost-benefit estimation it is assumed that an infrastructure equipment rate of 50% will be sufficient to cover almost all of the relevant accidents. As can be seen from Table 65, the BCR of the V2I bundle calculated for the selected business and service models indicate that the V2I based SAFESPOT system is not efficient under the given assumptions. The main cause of this finding is the high level of




infrastructure costs resulting from a large scale equipment of the infrastructure.

Table 65: Benefit-cost-ratios for the considered business/service models of the V2I bundle (infrastructure equipment rate 50 %)

V2I bundle

Business/service model Penetration rate BCR

Service model base – private 5.4 % 0.21

Service model plus – private 7.7 % 0.29

Business model base – public 7.9 % 0.30

Business model – public/ private 9.5 % 0.36 The low numbers for the four selected service and business models can be put in a broader perspective to be able to create a more positive benefit cost ratio. Positive numbers may be found, if a combination is made with Value Added Services which potentially have a better business case, and where the safety benefits will create a positive addition. Moreover, non-safety impacts in terms of traffic flow, fuel consumption and resulting emissions may be intensified for V2I systems which also include applications designed for traffic optimization. However, this needs further investigation and will be more elaborated in BLADE work package 6.7 deployment program.

4. The sensitivity of the BCR was extensively tested. In general, the BCR of the considered systems is not very sensitive to small variations of the calculation parameters. The sensitivity analysis was performed for the following parameters: effectiveness in avoiding fatalities/ injured, estimated trend of fatalities and injured, cost-unit rate for fatalities and injured, vehicle stock, in-vehicle system costs, life expectancy of in-vehicle system, and Infrastructure costs. The highest sensitivity appears with the dominating cost parameter (V2V: vehicle cost, V2I: infrastructure costs), followed by the injury and fatality benefit parameters. However, variation of the penetration rate has no major effect on the sensitivity of the BCR.

5. The infrastructure cost outweighed the safety benefits of the V2I based system. A separate scenario analysis was made relaxing the assumption of a large-scale equipment of the infrastructure. Four scenarios have been considered assuming that (1) 80 % of the safety effects will be reached with only 20 % of the costs, (2) only highways, (3) only intersections, (4) only European trunk roads and urban intersections would be equipped. Besides a reduction in infrastructure costs the scenarios also entail a reduction in safety benefits compared the baseline scenario of a large-scale equipment of infrastructure. Interestingly, the positive effect on BCR caused by a reduction of the level of infrastructure equipment was rather low for high fleet penetration rates, but high for low fleet penetration rates. Thus the calculations reveal that the fleet penetration rate may have a considerable impact on realized safety benefits and system costs.




Furthermore, the scenario analysis reveals that the profitability of the V2I system from a society point of view could be increased, if the equipment of infrastructure is done on a smaller scale concentrating on accident black spots. The case study of the Netherlands underlines this result showing that a partially equipped infrastructure can also create large benefits.

6. The V2I based system pays off for private drivers with an average mileage, whereas the V2V based system pays off for drivers with an annual mileage above the average. This was shown by the break-even analysis which calculated a critical mileage (minimum mileage) for using the SAFESPOT systems.

For the V2V system bundle the critical mileage is in the range between 24 and 31 thousand kilometres per year. For the V2I system bundle the critical mileage is more than halved compared to the V2V case ranging between 9 and 13 thousand kilometres per year. It was assumed that the end user only pays for the in-vehicle system but does not take a share in financing the infrastructure. This is certainly a critical assumption, which was relaxed later in the financial analysis for road operators.

7. From the results of the financial analysis of costs and revenue flows for road operators it can be assumed that the large-scale infrastructure equipment of the V2I based SAFESPOT system is too expensive for private or public road operators so that the system can not be operated economically. Charging cost prices to the private users will largely increase their financial burden. As a result the critical mileage of the drivers becomes much larger than the aforementioned one and thus may decrease the acceptance of the system. The following table shows the results for private and public road operators:

Table 66: Critical mileage with complete private (user) funding of private infrastructure



Annual fee [EUR]

Critical mileage [km]

5 % 165.00 494.00 186,908

10 % 120.00 247.00 97,816

20 % 90.00 123.00 52,926




Table 67: Critical mileage with complete private (private) funding of public infrastructure



Annual fee [EUR]

Critical mileage [km]

5 % 165.00 395.00 150,298

10 % 120.00 197.00 79,122

20 % 90.00 99.00 43,263

On the contrary, yield from VAT will increase with rising private enterprise, e.g. production and supply of in-vehicle systems, sensors and communication tools for infrastructure, such that the state will profit from that. This effect should be taken into account when financial incentives for fostering deployment are considered by public authorities. In this case, the VAT yield increase partly compensates the initial amount of the incentives (similar to the effects in the recent scrappage scheme in the automotive sector).

The analysis was done for the V2V case and the V2I case. In the V2V case, the VAT was based on the in-vehicle market price of the cooperative system. In the V2I case, the VAT was based on the in-vehicle market price plus a lump sum payment of the driver for the infrastructure. It was assumed that this payment does not completely cover the expenses for the infrastructure. This assumption reflects that public authorities support road safety goals by supporting deployment of cooperative systems besides budget constraints.

The table summarizes the VAT earnings in the different scenarios:

Table 68: VAT earnings (in Mill. EUR) for different scenarios

Scenario

Penetration rate

5 % 10 % 20 %

V2V

VAT earnings [Mill. EUR] 181.75 300.59 585.14


V2I

VAT earnings [Mill. EUR] –completely subsidized infrastructure 50,89 74,02 111,03


VAT earnings [Mill. EUR] – part of infrastructure financed by users 121,83 215,89 394,78

User fee [EUR] 230.00 230.00 230.00




The VAT earnings are calculated for a given level of market penetration. By providing tax rebates to subsidize market prices, public authorities can increase the attractiveness of cooperative systems for the drivers. Such supporting actions will increase the penetration rate of these systems. Of course, these interactions between policy supporting actions, penetration rates, and VAT earnings on in-vehicle cooperative systems have to be taken into account.

Considering different models of financial support for cooperative systems and their estimated impact on penetration rates, and BCRs could be a promising field for future research.

8. It is a question, whether the deployment process should start with the V2V or the V2I based SAFESPOT system. In detail this question will be discussed in BLADE WP7 “Deployment program”. Looking at both systems as independent and alternatives (not as combined solution) the BCR and the stakeholder analysis indicate that it seems more successful to start deployment rather with the V2V solution than with the V2I solution. However it needs to be realised that the first step for V2V is to overcome the 5% penetration barrier, before effects will arise, so that the V2I solution could be very important in this respect (see conclusion no. 9 below). Although the V2I based SAFESPOT system is an effective solution for improving road safety, the considered configuration of this system (i.e. large-scale infrastructure equipment) is not efficient from a socio-economic point of view.

On the other hand, if public authorities opt for a V2I based solution they can stimulate the fleet penetration by providing the infrastructure equipment thus relieving the financial burden for the private user. This effect is indicated by the lower critical mileage of the V2I based solution.

When considering the investments on the roadside necessary for V2I systems, one may also address the question, whether the introduction of cooperative systems might be an alternative solution for investing in the traditional road infrastructure. Since the most cost-effective measures are already taken in part of the European countries, further substantial improvement of road safety, traffic efficiency and environmental friendliness of road traffic, by means of investing in the road infrastructure in the traditional way, will be difficult to achieve in these countries. Instead, investing in cooperative systems might be considered. These considerations could not deeply be discussed in this study but may be taken into account in further research projects.

9. The economic assessment provides some indication for deployment strategies of combining the V2V based and the V2I based SAFESPOT system. It seems that added value in terms of increased safety benefits can be achieved, if in addition to a V2V based solution a V2I based solution is implemented. This requires a “smart” equipment of the infrastructure concentrating on a limited number of black spots with high




accident numbers. Recently published data by EuroRAP support this argument.

Both V2V system and V2I system use the same hardware for the in-vehicle device. Once V2V is implemented, no additional costs for the in-vehicle device arise in order to put V2I into effect for the SAFESPOT applications. This underlines our recommendation on the deployment strategy: It seems to be beneficial 1) to start with the V2V system implementation after having some black spots of the infrastructure equipped, 2) then strengthen the market penetration of the V2V system in order to exceed the critical penetration rate as fast as possible, and 3), if needed, add further infrastructure equipment at other black spots.

10. Of course, the results of the socio-economic assessment of the SAFESPOT system are based on simulations, studies, desk research and expert estimations. To improve the reliability and validity of results about the traffic and safety impacts, further empirical research programs in particular (such as field operational tests) are necessary.




10. References

[1] ABELE, J. et al. (2005), Exploratory Study on the potential socio-economic impact of the introduction of Intelligent Safety Systems in Road Vehicles, SEiSS-study, Teltow/ Köln

[2] ALONSO, M. et al. (2005), Literature review of behavioural effects, Deliverable D1.2.1, AIDE Project: Adaptive Integrated Driver-Vehicle Interface

[3] ASSING, K. et al. (2006), Methodological framework and database for socio-economic evaluation of Intelligent Vehicle Safety Systems, Deliverable D3, eIMPACT Project: Socio-economic Impact Assessment of Stand-alone and Co-operative Intelligent Vehicle Safety Systems (IVSS) in Europe, Köln

[4] BAUM, H. et al. (2008), Cost-Benefit Analyses for standalone and co-operative Intelligent Vehicle Safety Systems, Deliverable D6, eIMPACT Project: Socio-economic Impact Assessment of Stand-alone and Co-operative Intelligent Vehicle Safety Systems (IVSS) in Europe, Delft

[5] BAUM, H. and GRAWENHOFF, S. (2007), Cost-Benefit-Analysis of the Electronic Stability Program (ESC), Cologne

[6] BENZ et al. (2008), Impact Assessment of Intelligent Vehicle Safety Systems, Deliverable D4, eIMPACT Project: Socio-economic Impact Assessment of Stand-alone and Co-operative Intelligent Vehicle Safety Systems (IVSS) in Europe, Delft

[7] BICKEL, P. et al. (2006): Proposal for Harmonised Guidelines, Deliverable 5, HEATCO Project: Developing Harmonised European Approaches for Transport Costing and Project Assessment, Stuttgart

[8] BICKEL, P., et al. (2005), State-of-the-art in project assessment, Deliverable 2, HEATCO Project: Developing Harmonised European Approaches for Transport Costing and Project Assessment, Stuttgart

[9] BOARDMAN et al. (1996), Cost-Benefit Analysis: Concepts and Practice, New Jersey

[10] BROUWER, R.F.T. and HOEDEMAEKER, D.M. (Eds.) (2004), Driver support and information systems: Experiments on learning, appropriation and effects of adaptiveness, Deliverable D1.2.3, AIDE Project: Adapted Integrated Driver-Vehicle Interface

[11] COWI (2006), Cost-benefit assessment and prioritization of vehicle safety technologies, Final report, European Commission Directorate-General for Energy and Transport

[12] ICF CONSULTING Ltd. (2003), Costs-Benefit Analysis of Road Safety Improvements, Final Report

[13] DEPARTMENT FOR TRANSPORT IN THE U.K. (2004), The CVHS Business Case, Deliverable 6, Development Study for Cooperative Vehicle Highway Systems, London




[14] DONNER, E. et al. (2004), ADAS – Market Introduction Scenarios and Proper Realisation, Deliverable 1, RESPONSE 2 Project: Advanced Driver Assistance Systems: From Introduction Scenarios towards a Code of Practice for Development and Testing, Köln

[15] DRASKOCZY, M., CARSTEN, O. M. J. and KULMALA, R. (1998), Road Safety Guidelines, Deliverable B5.2, CODE Project: Telematics Application Programme

[16] EUROPEAN COMMISSION (2003), Proposal for a Directive of the European Parliament and of the Council amending Directive 1999/62/EC on the charging of Heavy Goods Vehicles for the use of certain infrastructure, Brüssel

[17] EUROPEAN COMMISSION (2008), European Energy and Transport, Trends to 2030 – update 2007, Brüssel

[18] EUROSTAT (2009), Victims of road accidents, download from http://epp.eurostat.ec.europa.eu/tgm/table.do?tab=table&init=1&language=de&pcode=tsdtr420&plugin=1

[19] ERDMENGER, C. et al. (2007), Climate change – Climate protection in Germany: 40 % reduction of CO2 emissions by 2020 compared to 1990, Dessau-Roßlau

[20] EUROBAROMETER (2006a), Users’ attitudes towards Electronic Active safety systems in Vehicles, Qualitative Study, Brüssel

[21] EUROBAROMETER (2006b): European Commission, Use of Intelligent Systems in Vehicles, Quantitative Study, Brüssel

[22] FGSV (2001), Forschungsgruppe für Straßen- und Verkehrswesen, Handbuch für die Bemessung von Straßenverkehrsanlagen, Köln

[23] FGSV (1997), Forschungsgesellschaft für Straßen- und Verkehrswesen, Arbeitsgruppe Verkehrsplanung, Empfehlungen für Wirtschaftlichkeits-untersuchungen an Straßen (EWS), Köln

[24] FRIEDRICH, M. (2009), AMONES, AMONES Symposium, Berlin

[25] GLASER, S. (2007), Excessive Speed Alert / Status and Progress, Presentation SAFESPOT SP5 WP3 Meeting, Paris, 21. September 2007

[26] HERRTWICH, R. (2003), E-Cars, Communication on the Road, Presentation IFIP Conference I3E, Sao Paulo

[27] HUBACHER, M., ALLENBACH, R: (2002), Lichtsignalanlagen – Anlagespezifische Untersuchung sicherheitsrelevanter Aspekte von vierarmigen Kreuzungen im Innerortsbereich, bfu-report R48, Bern

[28] INTERNATIONAL ROAD FEDERATION (2008), World Road Statistics 2008, Geneva.




[29] JANSSEN, W. (2001), The most important are: speed, speed variation and time headway, Presentation ITS World Congress, Sydney

[30] KULMALA, R. et al. (2008), Final study report, deliverable D5, CODIA Project: Co-Operative systems Deployment Impact Assessment, Espoo

[31] MALONE, K. et al. (2007), Traffic effects of WILLWARN, Internal Report PReVENT subproject WILLWARN, PReVENT Project: Preventive and Active Safety Applications

[32] NASA (2004), NASA Cost Estimating Handbook, Washington

[33] Nöcker, G. (2007), WILLWARN final report, SP Deliverable, Prevent Project: Preventive and Active Safety Applications, Sindelfingen

[34] ProgTrans (2007), European Transport Report 2007/2008, Basel

[35] ROAD SAFETY FOUNDATION (2010), Saving lives for less - EuroRAP 2010 results, Basingstoke

[36] SAFETYNET (2008a), Traffic Safety Basic Facts 2007 – Motorways, Brüssel

[37] SAFETYNET (2008b), Traffic Safety Basic Facts 2007 – Urban Areas, Brüssel

[38] SAFETNET (2008c), Traffic Safety Basic Facts 2007 – Junctions, Brüssel

[39] SCHELLER, H.-K. (2006), The European Central Bank – History, role and functions, European Central Bank, second edition, Frankfurt

[40] SCHOENEBECK, S. (2008), Accident Data Compilation – Challenges and Solutions, Presentation eIMPACT Meeting, Paris

[41] SCHOLLIERS, J. et al. (2008), Project final report and recommendations for future assessments, PReVENT Project: Preventive and Active Safety Applications, Tampere

[42] SAFESPOT SP4 WP2 (2006), Use case and typical accident situation, Deliverable D4.2.3, SAFESPOT subproject 4 (“SCOVA – Cooperative Systems Applications Vehicle Based), workpackage 2

[43] SAFESPOT SP4 WP 3 (2008), Conceptualization of on-board information system and extended, Deliverable D4.3.4, SAFESPOT subproject 4 (“SCOVA – Cooperative Systems Applications Vehicle Based), workpackage 3

[44] SAFESPOT SP6 WP 3 (2008), Preliminary Organisational Architecture, Deliverable D6.3.1, SAFESPOT subproject 6 (“BLADE – Business Models, Legal Aspects, and Deployment”), work package 3

[45] SAFESPOT SP5 WP2 (2007), Application Scenarios and System Requirements, Deliverable D5.2.3, SAFESPOT subproject 5 (“COSSIB – Cooperative safety systems infrastructure based”), work package 2

[46] SAFESPOT SP5 WP3 T1 (2008), Specifications for Speed Alert, Deliverable D5.3.1, SAFESPOT subproject 5 (“COSSIB – Cooperative safety systems infrastructure based”), work package 3, task 1




[47] SAFESPOT SP5 WP3 T5.3.2 (2008), Specifications for Hazard & Incident Warning Applications, Deliverable D5.3.2, SAFESPOT subproject 5 (“COSSIB – Cooperative safety systems infrastructure based”), work package 3, task 5.3.2

[48] SAFESPOT SP5 WP3 T5.3.3 (2008), Specifications for Co-operative Intersection Collision Prevention System, deliverable D5.3.3, SAFESPOT subproject 5 (“COSSIB – Cooperative safety systems infrastructure based”), work package 3, task 5.3.3

[49] SAFESPOT SP6 WP6.6 (2009), Service and Business Model Definition, Deliverable D6.6.1, SAFESPOT sub-project 6 (“BLADE – Business models, Legal Aspects, and DEployment”), work package 6

[50] SAFESPOT SP8 WP8.4 (2008), Use cases, functional specifications and safety margin, Deliverable D8.4.4, SAFESPOT subproject 8 (“HOLA – Horizontal Activities”), work package 8.4

[51] US DOT (1999), Evaluation of the Intelligent Cruise Control Systems, Volume 1, Study Results, Washington

[52] VAN AREM, B. (2007), Cooperative vehicle-infrastructure systems: an intelligent way forward? TNO Report, Delft

[53] WESTERKAMP, U. (2009), Ökonomische Bewertung von Systembündeln in der Fahrzeugsicherheit – Methodik und Bewertung am Beispiel ausgewählter Systeme, Köln

[54] WILMINK, I., et al. (2008), Impact Assessment of Intelligent Vehicle Safety Systems, Deliverable D4, eIMPACT Project: Socio-economic Impact Assessment of Stand-alone and Co-operative Intelligent Vehicle Safety Systems (IVSS) in Europe




11. Annex/es

11.1. Accident data (2005, EU-25)

Cluster 1 (6 countries):

Denmark, Finland, Germany, Sweden, The Netherlands, United Kingdom

Injury accidents

Fatalities Seriously injured

Slightly injured

Totals 542.555 9.480 109.629 592.723 Collision type


10,1 % 12,7 % 13,1 % 7,1 %


5,3 % 2,0 % 3,2 % 5,3 %


14,6 % 27,9 % 20,6 % 12,9 %

Frontal collision 8,4 % 17,0 % 12,5 % 9,7 % Side-by-side collision 4,5 % 2,0 % 2,9 % 4,5 % Angle collision 24,3 % 14,1 % 21,4 % 24,9 % Rear collision 14,1 % 3,7 % 5,3 % 16,0 % Other accidents with two vehicles

7,6 % 6,0 % 6,2 % 7,6 %

All other collisions 11,0 % 14,7 % 14,9 % 11,9 % Road type


69,0 % 32,0 % 54,0 % 67,5 %

Motorway 5,3 % 9,3 % 6,3 % 6,1 % Rural roads (no motorway)

25,7 % 58,7 % 39,7 % 26,5 %

Weather

Adverse 14,9 % 11,7 % 12,7 % 15,9 % Normal 85,1 % 88,3 % 87,3 % 84,1 %

Light conditions

Darkness 24,0 % 38,6 % 29,4 % 24,1 % Daylight or twilight or unknown

76,0 % 61,4 % 70,6 % 75,9 %

Location

At intersection 69,9 % 41,7 % 60,8 % 70,6 % No intersection 30,1 % 58,3 % 39,2 % 29,4 %





Austria, Belgium, France, Ireland, Italy, Luxemburg, Malta, Spain

Injury accidents


Slightly injured



11,6 % 11,0 % 11,5 % 7,6 %


4,4 % 5,8 % 5,7 % 3,6 %


11,1 % 21,8 % 14,4 % 9,5 %


4,8 % 3,5 % 3,8 % 5,0 %



61,7 % 24,5 % 48,7 % 60,1 %


33,1 % 69,7 % 45,6 % 34,4 %

Weather

Adverse 11,2 % 13,3 % 12,6 % 11,3 % Normal 88,8 % 86,7 % 87,4 % 88,7 %

Light conditions


73,3 % 61,9 % 70,0 % 72,8 %

Location






Cyprus, Czech Republic, Estonia, Greece, Hungary, Latvia, Lithuania, Poland, Portugal, Slovakia, Slovenia

Injury accidents


Slightly injured



14,3 % 16,6 % 16,4 % 9,6 %


13,6 % 13,3 % 16,1 % 14,0 %


10,5 % 17,3 % 11,1 % 10,6 %


0,9 % 1,3 % 0,9 % 1,0 %



67,7 % 41,3 % 54,0 % 65,2 %


30,1 % 53,0 % 43,2 % 32,4 %

Weather

Adverse 14,0 % 15,0 % 15,7 % 14,8 % Normal 86,0 % 85,0 % 84,3 % 85,2 %

Light conditions


70,7 % 59,2 % 65,7 % 70,4 %

Location





11.2. Market assessment: The sample

Market Penetration V2I Public Reliance_2030

7% 6%

11%9%

17%

0%2%4%

6%8%

10%12%14%

16%18%20%

0-20%

20-40%

40-60%

60-80%

80-100%

P e ne tra tion (% )

c

Res

pond

ents

(%

)

0-20% 20-40% 40-60% 60-80% 80-100%

Penetration (%)

Market Penetration V2V Public Reliance_2030

9% 9%6%

9%

16%

0%

5%

10%

15%

20%

0-20%

20-40%

40-60

%

60-80%

80-10

0%

Penetration (%)

Re

sp

on

de

nts

(%

)R

espo

nden

ts (

%)

0-20% 20-40% 40-60% 60-80% 80-100%

Penetration (%)

M arket Penetration V2V Public Re liance_2015

38%

9%

1% 0% 1%

0%5%

10%15%20%25%30%35%40%

0-20

%

20-4

0%

40-6

0%

60-8

0%

80-1

00%

P en etration ( % )

Res

pond

ents

(%

)

0-20% 20-40% 40-60% 60-80% 80-100%

Penetration (%)


27%

10%

3% 5% 5%

0%

5%

10%

15%

20%

25%

30%

0-20

%

20-4

0%

40-6

0%

60-8

0%

80-1

00%

Penetration (%)

Re

sp

on

de

nts

(%

)R

espo

nden

ts (

%)

0-20% 20-40% 40-60% 60-80% 80-100%

Penetration (%)


17%

13%

11%

6%

3%

0%2%4%6%8%

10%12%14%16%18%20%

0-20%

20-40%

40-60%

60-80%

80-100%

P e n e tra t io n (% )

Res

pond

ents

(%

)

0-20% 20-40% 40-60% 60-80% 80-100%

Penetration (%)

Market Penetration V2V Public Reliance_2020

20%

10%12%

4% 3%

0%

5%

10%

15%

20%

25%

0-20%

20-4

0%

40-6

0%

60-8

0%

80-1

00%

Penetration (%)

Re

sp

on

de

nts

(%

)R

espo

nden

ts (

%)

0-20% 20-40% 40-60% 60-80% 80-100%

Penetration (%)




Market Penetration V2I Private Reliance_2015

61%

12% 19%1% 4% 1% 0% 0% 0% 1%

0%20%40%60%80%

0-10%

11 -20%

21 -30%

31 -40%

41-50%

51-60%

61 -70%

71 -80%

81 -90%

91 -100%

Penetration (%)

Re

sp

on

de

nts

(%

)

Market Penetration V2V Private Reliance_2030

13%11%

9%5%

11%

0%

5%

10%

15%

0-20%

20 -40%

40 -60%

60-80%

80-100%

Penetration (%)

Re

sp

on

de

nts

(%

)


41%

5%1% 1% 1%

0%10%20%

30%40%50%

0-20%

20 -40%

40 -60%

60 -80%

80 -100%

Penetration(%)

Re

sp

on

de

nts

(%

)


27%

12%7%

3% 1%0%5%

10%15%20%25%30%

0-20%

20-40%

40-60%

60-80%

80-100%

Penetration (%)

Re

sp

on

de

nts

(%

)


23%

13%9%

3% 1%0%5%

10%15%20%25%

0-20%

20-40%

40-60%

60 -80%

80 -100%

Penetration (%)

Re

sp

on

de

nts

(%

)


13%11%

7% 6%

12%

0%

5%

10%

15%

0-20%

20 -40%

40 -60%

60 -80%

80-100%

Penetration (%)

Re

sp

on

de

nts

(%

)




Market Penetration Public+Private Reliance_2015

29%

12%

3% 2% 4%

0%5%

10%15%20%25%30%35%

0-20%

20-40%

40-60%

60-80%

80-100%

Penetration (%)

Re

sp

on

de

nts

(%

)

Market Penetration Public+Private Reliance_2030

5% 4%10% 8%

23%

0%5%

10%15%20%25%

0-20%

20-40%

40-60%

60-80%

80-100%

Penetration (%)

Re

sp

on

de

nts

(%

)

Market Penetration Public+Private

Reliance_2020

11% 13% 14%

6% 7%

0%5%

10%15%

0-20%

20 -40%

40 -60%

60 -80%

80-100%

Penetration (%)

Re

sp

on

de

nts

(%

)

Documents

Report on socio-economic, market and financial assessment€¦Report on socio-economic, market and financial assessment