Upload
others
View
1
Download
0
Embed Size (px)
Citation preview
COMPARATIVE EVALUATION OF INSYNC AND TIME-OF-DAY SIGNAL TIMING PLANS
UNDER NORMAL AND VARIED TRAFFIC CONDITIONS
Draft Final Report
Aleksandar Stevanovic, PhD, PE
Milan Zlatkovic, PhD
Date of Research: February 2012 – December 2012
ACKNOWLEDGMENTS
The authors want to thank Rhythm Engineering staff for their help in the development and fine-
tuning of the InSync configurations. The authors are also thankful to traffic engineers from
Volusia County and the Florida Department of Transportation (FDOT) for their help with field
data collection and explanations of current field operations.
EXECUTIVE SUMMARY
InSync, an Adaptive Traffic Control System (ATCS) developed and supported by Rhythm
Engineering, is one of the youngest and fastest-growing ATCSs on the market. However, its
innovative software and hardware features and a strong marketing campaign led to dozens of
deployed InSync systems around the country. As with any other new technology, early
evaluations of a new ATCS are important in order to understand the system’s operations, observe
its advantages and disadvantages and compare the system with other benchmarking systems (e.g.
conventional Time-Of-Day [TOD] traffic control). ATCSs are usually evaluated in the following
two ways: in the field, usually through a before and after study; and in microsimulation, where
field-like adaptive logics are evaluated in a virtual reality that resembles field conditions. While
both approaches have advantages and disadvantages, in the virtual reality of microsimulation,
traffic conditions can be tightly controlled and stochastically replicated and varied. Incidents and
event-based traffic conditions can be carefully constructed and tested. Furthermore, optimal
installation can be simulated with high quality detector placement and rigorous management of
global and local control parameters, such as timing constraints. However, the major concern
when evaluating ATCSs and other traffic control regimes in microsimulation is the model’s
ability to reliably replicate all intricacies of real-world traffic conditions.
This study evaluates and compares the performance of InSync and TOD signal timing plans for
SR-421, a 12-intersection corridor in Port Orange, Volusia County, FL through microsimulation
under different operational scenarios. In order to realistically replicate field traffic conditions, a
12-intersection VISSIM model was comprehensively calibrated and validated based on multiple
sets of field traffic data. The operational scenarios used to compare InSync and TOD signal
timing performances cover both current (regular) traffic conditions and a set of irregular traffic
conditions when sudden events alter traffic flows. The use of irregular conditions is critical to
evaluate the adaptive benefit of the InSync system. Examples of the sudden events considered in
this study are: 1. A freeway incident where a significant portion of traffic from the freeway is
diverted to the evaluated corridor, 2. Inclement weather conditions in Florida (e.g. flooded streets
after a heavy rain) when travel speeds and saturation flow rates are reduced, 3. Frequent rail
preemption calls for a freight railway system that intersects the evaluated corridor and 4. A
sudden surge of traffic (on all roads; e.g. similar to evacuation conditions), which oversaturates a
number of turning movements at most of the intersections on the corridor.
To properly cover a variety of approaches traffic signal professionals would use to retime signals
on SR-421(where traffic demand levels [and v/c ratios] vary significantly at various
intersections) three TOD scenarios were tested. An optimized Single Section (SS) approach was
used to provide good coordination and fewer stops by putting all of the intersections under the
same cycle length. Meanwhile, dividing the entire SR 421 segment into optimized Multiple
Sections (MS) with different cycle lengths was an approach selected to accommodate the local
intersections’ traffic demands and thus reduce delays. The third tested TOD scenario represented
signal timings currently operating in the field.
The regular field traffic operations are evaluated under the three TOD signal timing scenarios
and InSync for all three peak time periods (AM, MD and PM peaks). The evaluations are
performed on the intersection, corridor and network-wide levels, as well as on main-street vs.
side-street performance. Also, the traffic signal regimes are evaluated for their environmental
and safety performance. The following conclusions are derived from experiments with regular
traffic conditions:
InSync outperforms TOD signal timings in terms of traffic efficiency. For this corridor
and its traffic demand, InSync is undoubtedly better than the existing signal timings and
the two optimal signal timing plans in terms of overall network performances (delay,
stops, average speed, etc.), corridor travel times, intersection delays and stops, and main
street delays. The only aspect where InSync is not clearly better than any other TOD
signal timings is side street delay, where InSync is neither best nor worst performing of
the evaluated regimes.
InSync outperforms TOD signal timings in terms of fuel efficiency and most of the
emission outputs. Benefits from InSync are not large although they are statistically
significant: InSync saves between 2-4% when compared to other timing plans. It is
speculated that the relatively low savings in fuel consumption are due to the nonlinear
relationship between fuel consumption and traffic efficiency metrics. Thus, InSync (and
other signal timing regimes) sometimes get penalized in higher fuel consumption for
providing a better level of service and higher traveling speeds.
InSync outperforms TOD signal timings in terms of reducing the total number of
vehicular conflicts – a surrogate safety measure reported by the Surrogate Safety
Assessment Model (SSAM). InSync is better in terms of total, rear-end and lane-
changing conflicts. InSync yields a higher number of crossing conflicts than the TOD
scenarios. Further research is needed to investigate the relationship between InSync and
TOD operations and the number of various conflicts reported by SSAM.
Irregular traffic operations are evaluated under the three TOD signal timing scenarios and InSync
only for PM peak conditions (when traffic demand is highest). The evaluations are again
performed on the various levels (intersection, corridor, etc.) as in the previous case. To gauge
performance under “shock” traffic conditions, the various regimes are evaluated using the same
configurations and timing plans as with the regular traffic conditions. The following conclusions
are derived from experiments with irregular traffic conditions:
InSync outperforms TOD signal timings when extra traffic is diverted from a freeway
due to a traffic incident. For this corridor and its traffic demand, InSync is better than all
of the TOD plans in terms of overall network performances, corridor travel times,
intersection delays and stops, and main street delays. The only aspect where InSync is not
clearly better than any other TOD signal timings is side street delay, where InSync is
neither best nor worst performing of the evaluated regimes.
InSync is better than TOD signal timings when a heavy rain reduces traffic speed and
saturation (discharge) flow rates. Again, InSync is better than all of the TOD plans in
terms of overall network performances, corridor travel times, intersection delays and
stops, and main-street delays. Traffic operations on side-streets are the only aspect where
InSync is outperformed by MS (and in some cases SS) optimized signal timings.
InSync also outperforms TOD signal timings when arterial traffic is interrupted by
relatively frequent train preemption calls. Again, InSync is better than TOD plans on all
levels except that the MS scenario (and sometimes the SS scenario) is better in terms of
side street delay.
Finally, InSync outperforms TOD signal timings when the overall traffic demand causes
significant oversaturation in the network. Similar to previous scenarios, InSync is better
than TOD plans in all aspects of operations at each level (intersections, corridor and
network) with the only weak point being side street delays when some of the TOD
scenarios are sometimes better.
Overall, the results of experiments performed in this study show that InSync is a versatile
and adaptable system that can outperform TOD plans (on the studied corridor) for both
regular and irregular traffic conditions (with unexpected changes in traffic).
TABLE OF CONTENTS
COMPARATIVE EVALUATION OF INSYNC AND TIME-OF-DAY SIGNAL TIMING
PLANS UNDER NORMAL AND VARIED TRAFFIC CONDITIONS .................................. 1
LIST OF TABLES .......................................................................................................................... 9
LIST OF FIGURES ...................................................................................................................... 10
1. Introduction ............................................................................................................................. 1
1.1. Overview of InSync ......................................................................................................... 2
1.1.1 Traffic Progression.................................................................................................... 3
1.1.2 Period Length Evaluation ......................................................................................... 3
1.1.3 Local Adaptive Logic ............................................................................................... 3
1.1.4 Digital Signal Control Concepts - Finite Number of Signal States .......................... 4
1.1.5 Scheduling of States .................................................................................................. 4
1.1.6 Review of Previous InSync Evaluations ................................................................... 5
2. MethodOLOGY ...................................................................................................................... 6
2.1. Project Description ........................................................................................................... 6
2.2. VISSIM Model Development .......................................................................................... 8
2.3. VISSIM Models Calibration and Validation .................................................................. 10
2.4. Modeling of InSync ATCS in VISSIM Models ............................................................. 12
2.5. Scenario Development ................................................................................................... 15
2.5.1 Existing Field Traffic Operations ........................................................................... 17
2.5.2 Traffic Operations for an I-95 Freeway Incident Situation .................................... 17
2.5.3 Traffic Operations under Inclement Weather Conditions ....................................... 18
2.5.4 Traffic Operations under Frequent Rail Preemption Calls ..................................... 19
2.5.5 Traffic Operations under Oversaturated Conditions ............................................... 19
3. Results and discussion .......................................................................................................... 20
3.1. Existing Field Traffic Operations ................................................................................... 20
3.1.1 Intersection Performance ........................................................................................ 20
3.1.2 Corridor Travel Times ............................................................................................ 23
3.1.3 Main-Street vs. Side-Street Performance ................................................................ 23
3.1.4 Network Performance ............................................................................................. 26
3.1.5 Environmental Performance Measures ................................................................... 26
3.1.6 Safety Performance Measures ................................................................................. 29
3.2. Traffic Operations for an I-95 Freeway Incident Situation ............................................ 31
3.2.1 Intersection Performance ........................................................................................ 31
3.2.2 Corridor Travel Times ............................................................................................ 34
3.2.3 Main-Street vs. Side-Street Performance ................................................................ 34
3.2.4 Network Performance ............................................................................................. 38
3.3. Traffic Operations under Inclement Weather Conditions .............................................. 41
3.3.1 Intersection Performance ........................................................................................ 41
3.3.2 Corridor Travel Times ............................................................................................ 44
3.3.3 Main-Street vs. Side-Street Performance ................................................................ 44
3.3.4 Network Performance ............................................................................................. 48
3.4. Traffic Operations under Frequent Rail Preemption Calls ............................................. 51
3.4.1 Intersection Performance ........................................................................................ 51
3.4.2 Corridor Travel Times ............................................................................................ 53
3.4.3 Main-Street vs. Side-Street Performance ................................................................ 54
3.4.4 Network Performance ............................................................................................. 58
3.5. Traffic Operations under Oversaturated Conditions ...................................................... 61
3.5.1 Intersection Performance ........................................................................................ 61
3.5.2 Corridor Travel Times ............................................................................................ 64
3.5.3 Main-Street vs. Side-Street Performance ................................................................ 64
3.5.4 Network Performance ............................................................................................. 68
4. Conclusions ........................................................................................................................... 71
5. References ............................................................................................................................. 73
LIST OF TABLES
Table 2.1: Corridor Speeds and LOS .............................................................................................. 8
Table 2.2: Corridor Field Cycle Lengths Generated by Various TOD Signal Timing Plans ....... 16
Table 2.3: Traffic Performance under Inclement Weather Reported by Literature ...................... 19
Table 3.1: Average Intersection Delay and Level of Service ....................................................... 21
Table 3.2: Average Number of Stops per Vehicle ........................................................................ 22
Table 3.3: Corridor (SR 421) Travel Times .................................................................................. 24
Table 3.4: Fuel Consumption and Emission Estimates ................................................................ 29
Table 3.5: Number of Estimated Safety Conflicts Between Vehicles .......................................... 30
Table 3.6: LOS Comparison: Freeway Incident Scenario ............................................................ 33
Table 3.7: Travel Speed and LOS Comparison: Freeway Incident Scenario ............................... 36
Table 3.8: Regular vs. Incident Intersection Performance: Freeway Incident Scenario ............... 40
Table 3.9: Regular vs. Incident Corridor Performance: Freeway Incident Scenario .................... 40
Table 3.10: Regular vs. Incident Network Performance: Freeway Incident Scenario .................. 40
Table 3.11: LOS Comparison: Inclement Weather ....................................................................... 43
Table 3.12: Travel Speed and LOS Comparison: Inclement Weather .......................................... 46
Table 3.13: Regular vs. Incident Intersection Performance: Inclement Weather ......................... 50
Table 3.14: Regular vs. Incident Corridor Performance: Inclement Weather .............................. 50
Table 3.15: Regular vs. Incident Network Performance: Inclement Weather .............................. 50
Table 3.16: LOS Comparison: Rail Preemption ........................................................................... 53
Table 3.17: Travel Speed and LOS Comparison: Rail Preemption .............................................. 56
Table 3.18: Regular vs. Incident Intersection Performance: Rail Preemption ............................. 60
Table 3.19: Regular vs. Incident Corridor Performance: Rail Preemption ................................... 60
Table 3.20: Regular vs. Incident Network Performance: Rail Preemption .................................. 60
Table 3.21: LOS Comparison: Oversaturation ............................................................................. 63
Table 3.22: Travel Speed and LOS Comparison: Oversaturation ................................................ 66
Table 3.23: Regular vs. Incident Intersection Performance: Oversaturation ................................ 70
Table 3.24: Regular vs. Incident Corridor Performance: Oversaturation ..................................... 70
Table 3.25: Regular vs. Incident Network Performance: Oversaturation ..................................... 70
LIST OF FIGURES
Figure 2.1: Test-Bed Network: Dunlawton Avenue, Port Orange, FL ........................................... 7
Figure 2.2: Historic AADT Data for SR-421 .................................................................................. 7
Figure 2.3: VISSIM Models Calibration and Validation .............................................................. 11
Figure 2.4: InSync-VISSIM Interface ........................................................................................... 13
Figure 2.5: Freeway Incident and Diversion Routes .................................................................... 18
Figure 3.1: Main-Street vs. Side-Street Average Vehicle Delays ................................................ 25
Figure 3.2: Total Network Delays ................................................................................................ 28
Figure 3.3: Comparison of Intersection Delays: Freeway Incident Scenario ............................... 32
Figure 3.4: Comparison of Number of Stops: Freeway Incident Scenario ................................... 32
Figure 3.5: Comparison of Average Queues: Freeway Incident Scenario .................................... 33
Figure 3.6: Comparison of EB Segment Travel Times: Freeway Incident Scenario .................... 35
Figure 3.7: Comparison of WB Segment Travel Times: Freeway Incident Scenario .................. 35
Figure 3.8: Comparison of Corridor Travel Times: Freeway Incident Scenario .......................... 36
Figure 3.9: Main-Street vs. Side-Street Movement Delays: Freeway Incident Scenario ............. 37
Figure 3.10: Main-Street vs. Side-Street: Aggregate Delays: Freeway Incident Scenario ........... 37
Figure 3.11: Comparison of Total Network Delay: Freeway Incident Scenario .......................... 38
Figure 3.12: Comparison of Total Network Number of Stops: Freeway Incident Scenario ........ 39
Figure 3.13: Comparison of Total Network Travel Times: Freeway Incident Scenario .............. 39
Figure 3.14: Incident Scenario Comparison on Different Levels: Freeway Incident Scenario .... 40
Figure 3.15: Comparison of Intersection Delays: Inclement Weather ......................................... 42
Figure 3.16: Comparison of Number of Stops: Inclement Weather ............................................. 42
Figure 3.17: Comparison of Average Queues: Inclement Weather .............................................. 43
Figure 3.18: Comparison of EB Segment Travel Times: Inclement Weather .............................. 45
Figure 3.19: Comparison of WB Segment Travel Times: Inclement Weather ............................. 45
Figure 3.20: Comparison of Corridor Travel Times: Inclement Weather .................................... 46
Figure 3.21: Main-Street vs. Side-Street Movement Delays: Inclement Weather ....................... 47
Figure 3.22: Main-Street vs. Side-Street: Aggregate Delays: Inclement Weather ....................... 47
Figure 3.23: Comparison of Total Network Delay: Inclement Weather ...................................... 48
Figure 3.24: Comparison of Total Network Number of Stops: Inclement Weather ..................... 49
Figure 3.25: Comparison of Total Network Travel Times: Inclement Weather ........................... 49
Figure 3.26: Incident Scenario Comparison on Different Levels: Inclement Weather ................ 50
Figure 3.27: Comparison of Intersection Delays: Rail Preemption .............................................. 52
Figure 3.28: Comparison of Number of Stops: Rail Preemption.................................................. 52
Figure 3.29: Comparison of Average Queues: Rail Preemption .................................................. 53
Figure 3.30: Comparison of EB Segment Travel Times: Rail Preemption .................................. 54
Figure 3.31: Comparison of WB Segment Travel Times: Rail Preemption ................................. 55
Figure 3.32: Comparison of Corridor Travel Times: Rail Preemption ......................................... 55
Figure 3.33: Main-Street vs. Side-Street Movement Delays: Rail Preemption ............................ 57
Figure 3.34: Main-Street vs. Side-Street: Aggregate Delays: Rail Preemption ........................... 57
Figure 3.35: Comparison of Total Network Delay: Rail Preemption ........................................... 58
Figure 3.36: Comparison of Total Network Number of Stops: Rail Preemption ......................... 59
Figure 3.37: Comparison of Total Network Travel Times: Rail Preemption ............................... 59
Figure 3.38: Incident Scenario Comparison on Different Levels: Rail Preemption ..................... 60
Figure 3.39: Comparison of Intersection Delays: Oversaturation ................................................ 62
Figure 3.40: Comparison of Number of Stops: Oversaturation .................................................... 62
Figure 3.41: Comparison of Average Queues: Oversaturation ..................................................... 63
Figure 3.42: Comparison of EB Segment Travel Times: Oversaturation ..................................... 65
Figure 3.43: Comparison of WB Segment Travel Times: Oversaturation ................................... 65
Figure 3.44: Comparison of Corridor Travel Times: Oversaturation ........................................... 66
Figure 3.45: Main-Street vs. Side-Street Movement Delays: Oversaturation .............................. 67
Figure 3.46: Main-Street vs. Side-Street: Aggregate Delays: Oversaturation .............................. 67
Figure 3.47: Comparison of Total Network Delay: Oversaturation ............................................. 68
Figure 3.48: Comparison of Total Network Number of Stops: Oversaturation ........................... 69
Figure 3.49: Comparison of Total Network Travel Times: Oversaturation ................................. 69
Figure 3.50: Incident Scenario Comparison on Different Levels: Oversaturation ....................... 70
1. INTRODUCTION
InSync, an Adaptive Traffic Control System (ATCS) developed and supported by Rhythm
Engineering, is one of the youngest ATCSs on the market. However, its innovative software and
hardware features and a strong marketing campaign led to dozens of deployed InSync systems
around the country (1). As with any other new technology, early evaluations of a new ATCS are
important in order to understand the system’s operations, observe its advantages and
disadvantages and compare the system with other benchmarking systems (e.g. conventional
Time-Of-Day [TOD] traffic control). ATCSs are usually evaluated in two ways: in the field,
usually through a before and after study and in microsimulation, where field-like adaptive logics
are evaluated in a virtual reality that resembles field conditions. While both approaches have
advantages and disadvantages, in the virtual reality of microsimulation, traffic conditions can be
tightly controlled and stochastically replicated and varied. Incidents and event-based traffic
conditions can be constructed and tested carefully. Installation can be simulated to be optimal
with high quality detector placement and rigorous management of global and local control
parameters, such as timing constraints. However, the major concern when evaluating ATCSs and
other traffic control regimes in microsimulation is the model’s ability to reliably replicate all
intricacies of real-world traffic conditions.
This study evaluates and compares the performance of InSync and TOD signal timing plans for
SR-421, a 12-intersection corridor in Port Orange, Volusia County, FL through microsimulation
under different operational scenarios. In order to realistically replicate field traffic conditions, a
12-intersection VISSIM model was comprehensively calibrated and validated based on multiple
sets of field traffic data. The operational scenarios used to compare InSync and TOD signal
timing performances cover both current (regular) traffic conditions and a set of irregular traffic
conditions when sudden events alter traffic flows. The use of irregular conditions is critical to
evaluate the adaptive benefit of the InSync system. Examples of the sudden events considered in
this study are: 1. A freeway incident where a significant portion of traffic from the freeway is
diverted to the evaluated corridor, 2. Inclement weather conditions in Florida (e.g. flooded streets
after a heavy rain) when travel speeds and saturation flow rates are reduced, 3. Frequent rail
preemption calls for a freight railway system that intersects the evaluated corridor and 4. A
2
sudden surge of traffic (on all roads; e.g. similar to evacuation conditions), which oversaturates a
number of turning movements at most of the intersections on the corridor.
To properly cover a variety of approaches that traffic signal professionals would use to retime
signals on SR-421(where traffic demand levels [and v/c ratios] vary significantly at various
intersections) several TOD scenarios were tested. An optimized Single Section (SS) approach
was used to provide good coordination and fewer stops by putting all of the intersections under
the same cycle length. On the other hand, dividing the entire SR 421 segment into optimized
Multiple Sections (MS) with different cycle lengths was an approach selected to accommodate
the local intersections’ traffic demands and thus reduce delays. The third tested TOD scenario
represented signal timings which currently run in the field.
The regular field traffic operations are evaluated under the three TOD signal timing scenarios
and InSync for all three time periods (AM, MD and PM peaks). The evaluations are performed
on the intersection, corridor and network-wide levels, as well as on main-street vs. side-street
performance. Also, the traffic signal regimes are evaluated for their environmental and safety
performance.
Irregular traffic operations are evaluated under the three TOD signal timing scenarios and InSync
only for PM peak conditions (when traffic demand is highest). The evaluations are again
performed on the various levels (intersection, corridor, etc.) as in the previous case. To gauge
performance under “shock” traffic conditions, the various regimes are evaluated using the same
configurations and timing plans as with the regular traffic conditions. As “shock” traffic
conditions authors refer here to those conditions when there is a sudden and large increase in
traffic demand (on one or more traffic routes in the network) which acts as a ‘shock’ on the
‘system’ and it lasts for a given period of time. More specifically, two scenarios considered in
this study that can be qualified as ‘shock’ scenarios are: ‘freeway incident’ scenario (described
under section 2.5.2) and ‘oversaturation’ scenario (section 2.5.5).
1.1. Overview of InSync
InSync is an adaptive traffic signal system developed by Rhythm Engineering (Lenexa, Kansas)
whose basic version uses video detection and image processing to detect vehicles and adjust
3
signal timings in real time to address changes in traffic demand. InSync has two major aspects of
operations: 1. it automatically adjusts local signal timings; and 2. it coordinates signals along
roadway arterials according to real time traffic demand. To learn more about InSync adaptive
logics, readers are encouraged to read detailed descriptions provided elsewhere (3, 4).
1.1.1 Traffic Progression
InSync uses time tunnels/green waves to ensure successive turning of green lights at consecutive
traffic signals. Time tunnels are commanded by a facilitator intersection. This usually means that
the speed lines from upstream intersections for the main two directions of travel intersect at the
facilitator intersection; however, they can also intersect between intersections. The facilitator
intersection decides a time at which it will serve a green band for the coordinated tunnel phases
and communicates that time with the adjacent intersections.
1.1.2 Period Length Evaluation
The time between tunnels is called a period (similar to cycle length in actuated-coordinated
operations). The local adaptive system at each intersection continually analyzes its queue lengths
and the percentage of occupancy for each phase to facilitate dynamic periods. InSync determines
its dynamic period lengths in three ways. First, each intersection reports whether the current
period length was enough to clear all its queues. Second, each intersection frequently reports
whether the current period length was adequate to clear each queue for all the periods during the
most recent rolling 15-minute timeframe. Third, InSync also considers the anticipated period
needed for the next 15 minutes based on that same 15-minute window by time-of-day and day-
of-week over the last four weeks. In this way InSync is performing a limited amount of
predictive modeling of period lengths.
1.1.3 Local Adaptive Logic
Beyond the constraints communicated by the facilitator as “tunnel messages,” the signals operate
in “intelligent fully-actuated” mode. If a period is 90 s in duration and a green light is guaranteed
for the main directions at each intersection for 10 s, then 80 s are available for the local optimizer
to schedule states (phase pairs) at each intersection according to its intelligent scheduling. The
local optimizer embodies the dominant logic and algorithm of the adaptive capacities of the
4
system. The local optimizer also has the ability to start the tunnel early and to extend or truncate
the tunnel’s duration based on local demand.
1.1.4 Digital Signal Control Concepts - Finite Number of Signal States
InSync needs to decide: 1. which phase sequence will serve the intersection with the greatest
delay reduction, 2. when to initiate a state (phase pair) and 3. the duration of that state. InSync
utilizes traffic controllers in free mode by disabling all volume density functions and by
managing placement of the detector calls to the controller. This design permits the controller to
react quickly and change the traffic signal in response to the target calls served by InSync.
Another important advantage of this “state-changing” architecture is transition-less operations,
which eliminate traffic flow disruption caused by the transition between static cyclical signal
timings.
1.1.5 Scheduling of States
Considering that InSync does not follow any constant sequence of signal phases, it needs to
decide which signal phases will run after the current green state. There are three main factors
InSync's adaptive logic considers when scheduling phases:
1) If it is close to the initiation of a new tunnel, it will schedule a main-street
sequence of states. After the tunnel(s) are serviced, the local optimizer will
schedule states where demand deems necessary.
2) If a tunnel has recently ended, it will schedule a cross street sequence as its
priority. The amount of time for the cross street is based on a balance between the
actual amount of clearance time needed and anticipated time needed.
3) States can be rescheduled if extra time is available within the period. Queues that
have been waiting the longest receive priority. Wait times being equal, the phase
with the largest queue is scheduled.
5
1.1.6 Review of Previous InSync Evaluations
Despite its relatively recent appearance in the ATCS market (2008), InSync has been evaluated
many times. For brevity reasons, the authors cannot dwell in the details of these evaluations, but
we provide for each of the key studies a range of improvement percentages (in travel times,
number of stops and delays) after InSync deployment. The most notable InSync evaluations are
those performed on corridors in the following studies: Lee’s Summit, MO by Midwest Research
Institute (5) with 17-87% improvements; San Ramon, CA by DKS Associates (6) with 0-85%
improvements; Upper Merion, PA by Pennoni Associates Inc. (7) with 10-53% improvements;
Salinas, CA by TJKM Transportation Consultants (8) with 37-91% improvements; Mt. Pleasant,
SC by HDR Engineering, Inc. (9) with 17-29% improvements; and Hillsboro, OR by Kittleson
and Associates, Inc. (10) with 4-24% improvements. In addition, the only previous InSync
evaluation in microsimulation (2) reported 5-24% improvements in travel times, stops and
delays.
2. METHODOLOGY
2.1. Project Description
The study area represents a four-mile corridor of SR-421 (Dunlawton Ave) in Port Orange, FL,
from Williamson Blvd to US-1. This corridor was selected for InSync evaluation because it
represents a common (sub)urban corridor that is relatively difficult to retime due to variable
traffic demand levels at various intersections on the corridor. It also faces significant seasonal
and annual variations in traffic flow and it serves as an evacuation route during catastrophic
events (e.g. hurricanes). The study area (shown in Figure 2.1) is a divided 4-6 lane facility with
the following signalized intersections: Williamson Blvd, Interchange with I-95 (NB and SB
ramp), Taylor Rd, Yorktowne Blvd, County Rd 483 (Clyde Morris Blvd), Victoria Garden Blvd,
Village Trail, Nova Rd, Spruce Creek Rd and US-1. With the exclusion of Nova Rd and Spruce
Creek Rd (which are free-running intersections) and US-1 (which is a part of a different
coordinated corridor), the intersections operate in actuated-coordinated mode, with coordination
along SR-421. Protected left turns are present at almost every intersection, while only a few of
them facilitate protected-permitted operations. Speed limits range from 35 to 50 mph. Similarly,
the Annual Average Daily Traffic (AADT) varies between 25,000 and 35,000 vehicles per day,
depending on corridor section (Figure 2.2). SR-421 is crossed by freight railroad tracks (usually
not operational during peak periods) and it has a school zone active during morning and early
afternoon hours.
7
Figure 2.1: Test-Bed Network: Dunlawton Avenue, Port Orange, FL
10,000
15,000
20,000
25,000
30,000
35,000
40,000
Vehicles/day
Williamson to Clyde Morris Blvd. Clyde Morris Blvd. to SR 5A/Nova Rd.
SR5A/Nova Rd. to Spruce Creek Rd. Spruce Creek Rd. to US 1
Figure 2.2: Historic AADT Data for SR-421
8
Traffic operations vary for different time periods during the day, as well as for different
segments along the corridor. Corridor travel times using the floating car technique were collected
in spring of 2012 for all three periods (AM, MD and PM). These data were used to calculate the
average travel speeds and LOS presented in Table 2.1.
Table 2.1: Corridor Speeds and LOS
No.
From
To
Length
(mi)
AM Peak MD Peak PM Peak Speed (mph)
LOS Speed (mph)
LOS Speed (mph)
LOS
EA
ST
BO
UN
D
1 Williamson I-95 SB 0.134 32.2 B 26.4 C 29.7 B2 I-95 SB I-95 NB 0.090 32.5 B 31.7 B 38.3 A3 I-95 NB Taylor 0.139 37.5 A 37.0 A 40.0 A4 Taylor Yorktowne 0.307 33.5 B 29.6 B 42.5 A5 Yorktowne Clyde Morris 0.337 26.8 C 37.1 A 39.5 A6 Clyde Morris Victoria 0.404 34.1 B 29.8 B 26.4 C7 Victoria/City Swallowtail 0.410 26.8 C 23.5 C 18.4 D8 Swallowtail Nova 0.506 20.5 D 19.4 D 14.0 E9 Nova Spruce 0.910 41.3 A 34.1 B 43.0 A
10 Spruce US 1 0.713 19.4 D 23.9 C 17.9 D
WE
ST
BO
UN
D
11 US 1 Spruce 0.716 30.3 B 32.4 B 36.3 A
12 Spruce Nova 0.906 33.3 B 28.6 B 27.9 C
13 Nova Swallowtail 0.506 25.4 C 29.6 B 29.5 B
14 Swallowtail Victoria 0.410 27.7 C 27.4 C 21.2 D
15 Victoria/City Clyde Morris 0.406 19.2 D 25.1 C 28.1 B
16 Clyde Morris Yorktowne 0.344 24.9 C 37.2 A 39.3 A
17 Yorktowne Taylor 0.307 21.2 D 36.5 A 32.9 B
18 Taylor I-95 NB 0.128 30.1 B 26.7 C 29.8 B
19 I-95 NB I-95 SB 0.090 16.5 E 30.9 B 28.8 B
20 I-95 SB Williamson 0.134 16.9 E 7.3 F 7.6 F
Total 21 Williamson US 1 3.951 27.6 C 27.3 C 24.6 C
22 US 1 Williamson 3.947 25.9 C 27.1 C 27.0 C
2.2. VISSIM Model Development
VISSIM microsimulation software was selected as a tool to perform experiments because it is, so
far, the only microsimulation model interfaced with InSync software. Three VISSIM models for
AM, MD and PM peak periods were developed based on the existing geometry, traffic
operations and traffic control. The inputs used in model development were as follows:
Geometry. The VISSIM network was developed based on the aerial images and street
view obtained through Google Maps. A set of thirteen aerial images of the corridor were
scaled and imported as the background in VISSIM. A highly detailed model was coded
9
using the background as a base. The geometry was checked through street view to make
sure that the model objectively represented actual conditions. A total of twelve signalized
intersections and a number of major unsignalized intersections were modeled, including
pedestrian paths where applicable. It should be noted here that additional intersection at
Summer Trees Rd and SR-421 was added to the model. This was necessary to ensure that
incoming EB traffic flows at Williamson Blvd and SR-421 are modeled properly.
However, considering that no field traffic data for this intersection were available (and
thus its performance could not be calibrated and validated) this intersection was not
analyzed and included in the results from this report together with other intersections.
Traffic. Data on traffic movement counts for signalized intersections along the corridor
were obtained from Volusia County/FDOT District 5 (D5). These data were used to
define initial traffic in the network, especially turning movement percentages at
signalized intersections for routing decisions. The available data on pedestrian traffic
were included in the model. The signalized intersections were defined as nodes in
VISSIM, and these nodes were coded to collect traffic movements from the simulation.
An Excel spreadsheet was designed to use VISSIM outputs and to make a comparison to
traffic counts from the field. This spreadsheet was later used for traffic volume balancing
and model calibration.
Traffic signals. Signal timing data for the corridor were also obtained from Volusia
County/FDOT D5. Traffic signals in VISSIM were developed using the built-in Ring
Barrier Controller (RBC); all the settings match the actual signal timings from the field in
every detail: signal phase allocation, phase timing data, phase sequence and coordination
settings. Street view was again used for a precise definition of signal heads, left turn
treatment and stop bar/pedestrian crossings position. Detector size, location and function
were modeled according to the NEMA standards.
Road signs. Street view was again used to locate stop and speed limit signs along the
corridor and side-streets, and the signs were coded in the model. Special attention was
given to the speed limits; the corresponding speed decisions were created in simulation.
The speed reduction zone in the area of the Port Orange Elementary school was also
10
considered for eventual impacts. In addition to speed limits, actual recorded travel times
and speeds were used to update model parameters.
Other model elements. Priority rules and turning speed reductions were also modeled in
order to closely represent real conditions. The railroad located west of Orange Ave was
incorporated into the model and prepared to simulate eventual rail traffic.
2.3. VISSIM Models Calibration and Validation
The SR 421 VISSIM models were calibrated and validated based on driving behavior captured
through multiple traffic metrics collected in the field. The calibration process was performed
manually by adjusting elements of the model, including speed distributions, desired speed
decisions and saturation flows at intersection approaches. Routing decisions and traffic inputs
were adjusted to reflect manually collected traffic counts. The model was validated through a
comparison of modeled and field travel times measured along intersection segments. Travel
times were collected both by GPS devices (floating car method) and by video recordings of
travel runs. Video recordings were especially instrumental in observing queues and general
congestion levels at each intersection; these observations were later used to validate the virtual
traffic conditions in microsimulation models. Figure 2.3 presents the calibration and validation
results, which both indicate close matches between field data and VISSIM outputs (with
coefficients of determination [R2] ranging from 0.90 to 0.96).
11
a) AM Peak
b) Mid-day
c) PM Peak
R² = 0.910
0
20
40
60
80
100
120
140
160
0 20 40 60 80 100 120 140 160
Mod
eled
Tra
vel T
imes
(s)
Field Travel Times (s)
R² = 0.906
0
20
40
60
80
100
120
140
0 20 40 60 80 100 120 140
Mod
eled
Tra
vel T
imes
(s)
Field Travel Times (s)
R² = 0.955
0
20
40
60
80
100
120
140
0 20 40 60 80 100 120 140
Mod
eled
Tra
vel T
imes
(s)
Field Travel Times (s)
R² = 0.925
0
500
1000
1500
2000
2500
3000
3500
Mod
eled
Tur
ning
Cou
nts
(veh
/2 h
rs)
Field Turning Counts (veh/2 hrs)
R² = 0.967
0
500
1000
1500
2000
2500
3000
3500
Mod
eled
Tur
ning
Cou
nts
(veh
/2 h
rs)
Field Turning Counts (veh/2 hrs)
R² = 0.962
0
500
1000
1500
2000
2500
3000
3500
Mod
eled
Tur
ning
Cou
nts
(veh
/2 h
rs)
Field Turning Counts (veh/2 hrs)
CALIBRATION VALIDATION
Figure 2.3: VISSIM Models Calibration and Validation
12
2.4. Modeling of InSync ATCS in VISSIM Models
The VISSIM-InSync integration framework can be classified as Software-in-the-Loop-
Simulation, as the software that controls traffic in VISSIM is the same as software that controls
traffic at InSync deployments. Figure 2.4 shows a few screenshots of the VISSIM-InSync
interface framework.
a) Interface software: Left – General Tab; Right – An Intersection Tab
13
If a picture is used, use this layout. Picsare much preferred over bullets or other text.
b) Traffic Operations Windows: Left – InSync; Right – VISSIM
c) InSync Detectors in VISSIM
Figure 2.4: InSync-VISSIM Interface
To interface InSync software with VISSIM microsimulation, it was necessary to perform four
steps. First, detectors in VISSIM were set up to resemble field video detection systems, which
are the most common detection types for InSync. Once the simulation starts, vehicles were
detected by a series of detectors emulating field-like video detection zones from field InSync
14
deployments. Unlike in default VISSIM/RBC (Ring-Barrier Controller) cases, these detectors
were numbered according to nomenclature that ensured no calls were passed to relevant RBC
phases. For example, on an intersection approach served by the controller’s phase # 8 (and
hosting detector # 8), detectors numbered 801, 802 …808 were placed. Activation of these
detectors was noted by InSync but not by the RBC controller. Once InSync decided that phase #
8 needed to be served, it activated detector # 8 and held the call (while suppressing calls for all
other phases) until its logic decided it was time to cease that call and give a green to another
phase.
Second, controllers in VISSIM did not constrain ‘hold’ and ‘force off’ inputs from InSync. Local
controllers, which were hosted in VISSIM’s RBC files, were set up to run ‘free’ or
uncoordinated operations. However, additional steps were taken to ensure ‘free operations’ did
not constrain InSync’s ability to fully control green allocations and durations. The following
steps, as applied in any InSync field deployment, were executed when setting up RBC controllers
in VISSIM:
1) Set controller into “free/uncoordinated” mode
2) Enable Detector Diagnostic Failure Mode
3) Set all “Minimum Green” times to 5 s, or similar
4) Leave “Maximum Green” times unchanged
5) Set “Passage Gap” or “Observed Gap” to 1 s
6) For protected/permitted left turns, omit the left turn call when the opposing through
movement is green
7) Enable “Soft Recall” on the mainline phases
8) Disable “Yellow Lock” and “Red Lock” detector locking
9) Set all “Detector Delays” to 0 s
10) Disable all recalls: Max, Min, Hard, Vehicle, Phase, etc.
11) Remove “Extensions”
12) Disable “Anti-Backup” or “Left Turn Trap”
13) Enable “Max Recall Inhibit”
15
Third, a program that interfaces VISSIM and InSync was set up (configurations are written in
xml files) to connect individual InSync processors, each of which resided on a separate virtual
machine running with InSync software that connected with VISSIM. Each InSync processor had
an IP address and communicated with other processors through TCPIP protocols.
Fourth, once InSync’s main configurations were developed for each peak period, InSync
operations were fine-tuned to customize InSync performance for a particular peak period. Both
development of major InSync configurations and fine-tuning were performed by Rhythm
Engineering staff, whose involvement guaranteed optimal, practical InSync setup that would
perform as well as any InSync field deployment.
2.5. Scenario Development
The InSync evaluation was performed by comparing the system to field signal settings, and those
optimized by Synchro for single and multiple sections, under various traffic conditions. First, the
signal timing scenarios were defined for all three periods (AM, MD and PM) as follows:
VISSIM model with existing signal timings from the field
VISSIM model with Synchro optimized signal timings for single sections (SS Optimized)
VISSIM model with Synchro optimized signal timings for multiple sections (MS
Optimized)
VISSIM model with signal timings optimized in real time by InSync
Existing TOD signal timings were provided by the traffic engineering division of Volusia
County from field controllers. Based on the quality of their performance, it seems these signal
timings were periodically/recently adjusted to accommodate changes in traffic demand.
Considering traffic demand levels (and v/c ratios) vary significantly at various intersections on
SR-421, there was a dilemma in determining which approach should be taken when optimizing
signal timings. Putting all of the intersections under the same cycle length (Single Section [SS])
would provide good coordination and fewer stops. On the other hand, dividing the entire SR 421
segment into Multiple Sections (MS) with different cycle lengths would accommodate the local
16
intersections’ traffic demands and thus reduce delays. To cover most of the possible benefits
from TOD plans, both approaches were applied. In total, three TOD plans were compared with
InSync. Table 2.2 graphically shows differences in cycle lengths for the various traffic signal
control regimes.
Table 2.2: Corridor Field Cycle Lengths Generated by Various TOD Signal Timing Plans
Intersection Field SS Optimized MS Optimized Field SS Optimized MS Optimized Field SS Optimized MS Optimized
Williamson 115 110 68 120 124 76 125 144 100
I-95 SB 115 110 68 120 124 76 125 144 100
I-95 NB 115 110 68 120 124 76 125 144 100
Taylor rd 115 110 68 120 124 76 125 144 100
Yorktowne 115 110 68 120 124 76 125 144 80
Clyde Morris 115 110 116 120 124 130 125 144 160
Victoria 115 110 116 120 124 130 125 144 160
Village 115 110 116 120 124 130 125 144 160
Nova A 110 116 A 124 130 A 144 160
Spruce A 110 90 A 124 74 A 144 88
US 1 140 110 90 140 124 74 145 144 88
MD PeakAM Peak PM Peak
A–Intersections which run uncoordinated vehicle-actuated operations
These three signal timing scenarios were evaluated for different operational scenarios
characterizing different traffic operations along the corridor. The operational scenarios were
defined as follows:
Existing field traffic operations
Traffic operations for an incident on the I-95 freeway, at the end of the corridor
Traffic operations under inclement weather conditions (heavy rain)
Traffic operations under frequent rail preemption calls
Traffic operations under oversaturated conditions
Ten VISSIM simulation runs with different random seeds were implemented for each scenario
under the three signal timing scenarios and the five operational scenarios. Each simulation was 2
hours and 15 minutes long, with a 15-minute build-up time and two hours of evaluations.
17
2.5.1 Existing Field Traffic Operations
This operational scenario represented the existing field conditions in the AM, MD and PM peak
periods, and was also the base VISSIM model. The corresponding VISSIM models were created,
calibrated and validated based on the field inputs for traffic volumes and travel times/speeds. The
traffic volumes (intersection counts) were collected and provided by Volusia County/FDOT D5.
The majority of traffic counts were collected in February 2010, with the exception of Spruce
Creek Rd (August 2008) and US-1 (August 2011), and were provided on 15-minute time bases.
Corridor travel times were recorded in the spring of 2012 by Volusia County/FDOT D5 for the
three time periods and were also used in model creation and validation.
2.5.2 Traffic Operations for an I-95 Freeway Incident Situation
In this operational scenario, an incident on I-95 freeway between SR-421 and SR-400 was
incorporated into the model. The incident was modeled by closing one lane for one hour, and the
one-lane-equivalent traffic demand (approximately 1000 vph) was diverted to SR 421. The
alternative routes to I-95 were defined through two major corridors (US-1 and Nova Rd), which
both carried an additional 300 vph, and two minor corridors (Clyde Morris Blvd and Williamson
Blvd), which carried an additional 200 vph. The major corridors could be used to take the
travelers to their final destinations, while the minor corridors could be used as bypasses to I-95.
The incident and the expected traffic distributions are shown in Figure 2.5. This operational
scenario was evaluated for the PM peak only, since this was the period with highest traffic
volumes.
18
Figure 2.5: Freeway Incident and Diversion Routes
2.5.3 Traffic Operations under Inclement Weather Conditions
This operational scenario represented inclement weather conditions such as the heavy rains that
may occur in this part of Florida. It was modeled for the PM peak only, which was the worst case
scenario because of the heaviest demand. Based on the field experiences for this type of traffic
operations, travel speeds were reduced by an average of 15%, where all the existing speed
distributions were reduced by 10% to 20%.
The saturation flow rates were reduced by 20% and all saturation flow rates in the network were
set to 1500 vphpl. Table 2.3 summarizes findings reported in (11), which were used to model
traffic conditions under inclement weather.
19
Table 2.3: Traffic Performance under Inclement Weather Reported by Literature
Areas Facilities
Reductions Increases
Average Speed
Free-Flow Speed
AverageVolume
Sat. Flow Rate
Travel Time Delay
Start-Up
Delay
AKRural network with 5 arterials, 24 intersections
16%11% -15%
VTIntersection, 1 uphill approach
2% -21%
MNArterial with 5 intersections
40%15% -30%
11% 50%
UT 2 intersections10% -30%
6% -20%
50%5% -23%
DCUrban network with 15 arterials
12% -48%
4 urban networks 6% 11%
2.5.4 Traffic Operations under Frequent Rail Preemption Calls
This operational scenario represents inclement traffic conditions under frequent rail preemption
calls. The freight train tracks were located on the east side of the network, approximately 0.3
miles west of US-1. In existing conditions, about 18 trains crossed SR-421 per day, and there
were no trains during the PM peak. In this operational scenario, four trains were modeled during
the 2-hour PM peak period, and each train consisted of 20-100 cars (modeled 20, 30, 70 and
100). The speed of the trains was modeled in the range between 42 and 48 mph, based on the
average field speed of 45 mph. Trains were distributed evenly every 30 minutes, and full train
preemption was provided across SR-421. An average time of 4 seconds for the gates to close and
open was implemented in the model.
2.5.5 Traffic Operations under Oversaturated Conditions
This operational scenario evaluated the responsiveness of different signal timings under the
conditions of increased demand that lead to oversaturation. In this scenario, a 20% increase in
traffic demand was introduced at all traffic generators (VISSIM inputs). Other simulation
parameters, including the intersection turning proportions, remained the same as in the current
(regular) scenario.
20
3. RESULTS AND DISCUSSION
3.1. Existing Field Traffic Operations
The existing field traffic operations were assessed for the four signal timing scenarios and all
three time periods (AM, MD and PM peaks). The evaluations were performed on different
levels: intersection, corridor and network-wide, as well as on main-street vs. side-street,
environmental and safety performance.
3.1.1 Intersection Performance
The most detailed performance measures were assessed on the intersection level. The values for
average delays and average number of stops per vehicle were aggregated on the intersection level
and presented in Tables 3.1 and 3.2 respectively for each scenario and time period. The average
delay-based intersection Level of Service (LOS) is also presented in Table 3.1.
InSync was better than the TOD plans at the intersection level. Depending on the time period,
InSync reduced delays (when compared to field timings) at 80%-90% of the intersections in the
network, with delay reductions between 8% and 50%. Overall, InSync was better than any of the
TOD plans in about 50% of the cases (17 out of 33 comparisons) and it was never the worst
scenario. Taylor Road, Clyde Morris Boulevard, Nova Road and the I-95 ramps were among
those intersections where InSync outperformed other optimizations significantly, which was also
shown by their LOS values.
The same trend was observed when the average number of stops (Table 3.2) per vehicle was
analyzed. The reduction in number of stops was, however, more significant for each optimization
than the reduction in delays, as shown by the T-test. InSync yielded the best results for this
measure of effectiveness (MOE) at the majority of intersections. Significant improvements over
other optimization methods were observed at the same intersections as for the delay results.
21
Table 3.1: Average Intersection Delay and Level of Service
AM Peak Average Intersection Delay per Vehicle (s) (LOS) Intersection Field SS MS InSync Williamson 30.9 (C) 25.5 (C) 21.61,2 (C) 26.93 (C) I-95 SB 20.8 (C) 14.81 (B) 11.31 (B) 16.5 (B) I-95 NB 8.8 (A) 8.0 (A) 8.5 (A) 7.7 (A) Taylor 10.7 (B) 14.4 (B) 10.8 (B) 8.3 (A) Yorktowne 13.7 (B) 11.2 (B) 9.2 (A) 10.0 (A) CMB 26.6 (C) 31.0 (C) 32.42 (C) 27.1 (C) Victoria 17.8 (B) 18.9 (B) 18.0 (B) 15.71 (B) Village 34.8 (C) 22.1 (C) 19.0 (B) 17.4 (B) Nova 33.4 (C) 31.0 (C) 30.6 (C) 28.4 (C) Spruce 19.3 (B) 21.0 (C) 23.21 (C) 16.32 (B) US 1 36.6 (D) 28.2 (C) 24.51,2 (C) 31.03 (C)
Midday Average Intersection Delay per Vehicle (s) (LOS) Intersection Field SS MS InSync Williamson 26.5 (C) 22.9 (C) 17.71,2 (B) 23.6 (C) I-95 SB 18.6 (B) 16.0 (B) 12.6 (B) 11.01,2 (B) I-95 NB 3.7 (A) 3.6 (A) 7.2 (A) 2.6 (A) Taylor 13.6 (B) 12.9 (B) 13.8 (B) 11.0 (B) Yorktowne 15.9 (B) 21.1 (C) 16.0 (B) 16.4 (B) CMB 31.0 (C) 34.6 (C) 37.12 (D) 27.23 (C) Victoria 25.1 (C) 23.4 (C) 22.7 (C) 21.8 (C) Village 22.4 (C) 17.2 (B) 17.72 (B) 17.21,3 (B) Nova 44.6 (D) 38.4 (D) 38.1 (D) 30.41 (C) Spruce 20.1 (C) 20.5 (C) 17.01,2 (B) 18.53 (B) US 1 36.5 (D) 31.8 (C) 21.51,2 (C) 30.11,3 (C)
PM Peak Average Intersection Delay per Vehicle (s) (LOS) Intersection Field SS MS InSync Williamson 43.8 (D) 36.8 (D) 30.12 (C) 32.8 (C) I-95 SB 25.9 (C) 27.8 (C) 23.1 (C) 26.7 (C) I-95 NB 8.7 (A) 11.3 (B) 10.1 (B) 5.7 (A) Taylor 17.3 (B) 16.3 (B) 14.1 (B) 13.7 (B) Yorktowne 13.0 (B) 19.21 (B) 15.4 (B) 13.22 (B) CMB 37.7 (D) 43.8 (D) 45.21 (D) 32.51,2,3 (C) Victoria 25.4 (C) 29.41 (C) 31.01,2 (C) 25.5 (C) Village 21.5 (C) 16.1 (B) 15.12 (B) 15.21,2,3 (B) Nova 59.4 (E) 69.2 (E) 73.8 (E) 41.61,3 (D) Spruce 32.9 (C) 31.1 (C) 25.22 (C) 26.3 (C) US 1 38.7 (D) 34.7 (C) 24.91,2 (C) 32.41,2 (C)
1 - value statistically different from corresponding Field value 2 - value statistically different from corresponding SS Optimized value 3 - value statistically different from corresponding MS Optimized value
22
Table 3.2: Average Number of Stops per Vehicle
AM Peak Average Number of Stops per Vehicle Intersection Field SS MS InSync Williamson 0.65 0.561 0.661,2 0.561,2 I-95 SB 0.51 0.341 0.341 0.391 I-95 NB 0.34 0.24 0.342 0.282 Taylor 0.24 0.38 0.36 0.28 Yorktowne 0.47 0.381 0.401,2 0.441,2 CMB 0.64 0.721 0.741,2 0.651,2 Victoria 0.45 0.481 0.461,2 0.421,2 Village 0.66 0.501 0.411,2 0.431,2 Nova 0.71 0.711 0.641,2 0.601,2 Spruce 0.60 0.611 0.651,2 0.501,2 US 1 0.70 0.681 0.681,2 0.731,2,3
Midday Average Number of Stops per Vehicle Intersection Field SS MS InSync Williamson 0.59 0.461 0.501,2 0.551,2 I-95 SB 0.42 0.361 0.441,2 0.341,2 I-95 NB 0.10 0.09 0.19 0.06 Taylor 0.36 0.32 0.50 0.29 Yorktowne 0.45 0.621 0.571,2 0.531,2 CMB 0.69 0.761 0.761,2 0.681,2 Victoria 0.56 0.511 0.491,2 0.531,2 Village 0.56 0.401 0.411,2 0.451,2 Nova 0.78 0.761 0.741,2 0.631,2 Spruce 0.61 0.511 0.581,2 0.521,2 US 1 0.73 0.741 0.721,2 0.721,2
PM Peak Average Number of Stops per Vehicle Intersection Field SS MS InSync Williamson 0.77 0.631 0.631,2 0.651,2 I-95 SB 0.49 0.471 0.551,2 0.581,2 I-95 NB 0.25 0.38 0.31 0.19 Taylor 0.46 0.40 0.42 0.37 Yorktowne 0.24 0.371 0.381,2 0.291,2 CMB 0.80 0.861 0.841,2 0.821,2 Victoria 0.58 0.571 0.591,2 0.581,2 Village 0.53 0.361 0.321,2 0.421,2 Nova 0.88 1.091 1.051,2 0.721,2 Spruce 0.74 0.711 0.691,2 0.641,2 US 1 0.72 0.691 0.691,2 0.731,2
1 - value statistically different from corresponding Field value 2 - value statistically different from corresponding SS Optimized value 3 - value statistically different from corresponding MS Optimized value
23
3.1.2 Corridor Travel Times
The success of signal coordination and traffic progression along the main corridor is best
assessed through a corridor travel time analysis. The values of travel times between each pair of
signalized intersections averaged from ten simulation runs for every scenario are presented in
Table 3.3.
SS and MS TOD plans yielded similar benefits to corridor travel times when compared to travel
times from the field (under existing TOD plans). However, SS and MS plans resulted in longer
travel times in the EB direction during the PM peak, and the WB direction during the AM peak.
Savings in total travel times for SS and MS plans, when compared to the field, varied between
1% and 16%, depending on the time period and direction. The highest benefits were observed in
the midday period. InSync, on the other hand, yielded better travel times than those from any
other TOD plan during each period and in both directions. With InSync, total corridor travel
times from the field were reduced 11%-17%, depending on the direction and time period, and
they were lower than those achieved with SS or MS optimizations in all time periods.
3.1.3 Main-Street vs. Side-Street Performance
Facilitating main movements along a corridor can have certain impacts on side-street traffic.
There is a notion among signal timing professionals that InSync provides good progression for
main-street traffic at the expense of increasing delay for side-street traffic. For this reason, the
authors specifically wanted to investigate main-street vs. side-street delays, which were based on
intersection delays. Figure 3.1 shows a comparison of main-street and side-street delays for each
simulation scenario and time period, aggregated for all eleven intersections (Summer Trees Rd &
Sr-421 not included).
The analysis of intersection delays clearly shows that InSync outperformed all of the other
scenarios for main-street delay for all three peak periods. Field signal timings were worst for
most of the cases, while SS and MS optimized timings yield similar results for the main-street
traffic. Delays in these scenarios were reduced 8%-14% in AM and MD periods, while they were
slightly increased in PM periods when compared to the field. InSync resulted in the best
24
performance for the main-street traffic, reducing these delays by about 30% during each period
when compared to the field results.
Although different optimization methods had different effects on side-street traffic, each of them
still yielded better results than those observed in the field. InSync results in general were good.
They were never either best or worst. For MD and PM peaks, InSync performed second best
(after MS optimized signal timings), while for the PM peak, InSync was the second worst. When
compared to field signal timings, InSync consistently resulted in a 6% reduction in side-street
delays for each time period. When compared to Synchro (MS) optimized signal timings, InSync
consistently imposed more delay on side-streets, in the range of 5% to 25%, depending on the
time of day.
Table 3.3: Corridor (SR 421) Travel Times
Travel Times EB (s) AM Peak Midday PM Peak From To Field SS MS InSync Field SS MS InSync Field SS MS InSync
Williamson I-95 SB 37 17 14 14 27 17 17 13 24 24 20 23 I-95 SB I-95 NB 8 9 9 9 8 8 9 8 9 8 12 9 I-95 NB Taylor 14 24 19 16 20 22 21 16 22 20 18 17 Taylor Yorktowne 32 27 27 29 32 49 32 36 26 27 32 31 Yorktowne CMB 40 52 53 39 37 46 55 37 42 62 59 39 CMB Victoria 43 43 43 42 48 43 44 43 51 50 55 41 Victoria/City Swallowtail 54 56 53 51 58 51 52 50 59 51 46 47 Swallowtail Nova 78 63 58 56 92 61 65 62 125 174 180 81 Nova Spruce 82 89 86 73 97 90 97 92 92 102 96 71 Spruce US 1 129 93 95 98 114 87 84 80 128 94 94 103
Total EB 508 472 454 4361 523 442 454 4361,2 557 593 603 4761
Travel Times WB (s) AM Peak Midday PM Peak From To Field SS MS InSync Field SS MS InSync Field SS MS InSync
US 1 Spruce 87 75 99 75 78 73 67 66 84 69 77 72 Spruce Nova 96 107 95 74 121 131 119 90 127 141 131 110 Nova Swallowtail 63 54 55 55 63 52 51 51 65 53 51 51 Swallowtail Victoria 48 58 51 52 48 52 51 51 49 52 52 58 Victoria/City CMB 58 64 65 53 57 58 57 51 49 50 52 46 CMB Yorktowne 42 44 45 43 36 39 45 40 34 51 43 34 Yorktowne Taylor 37 40 42 37 34 32 42 32 34 29 32 28 Taylor I-95 NB 17 33 34 19 17 16 23 14 22 19 26 16 I-95 NB I-95 SB 9 10 11 13 11 15 21 15 14 13 21 13 I-95 SB Williamson 50 17 27 20 41 20 21 20 42 29 25 20
Total WB 501 521 530 4483 500 485 495 4433 530 506 492 4681,3
1 - value statistically different from corresponding Field value 2 - value statistically different from corresponding SS Optimized value 3 - value statistically different from corresponding MS Optimized value
25
0
10
20
30
40
50
60
Main street Side street
Agg
rega
te D
elay
per
Veh
icle
(s)
Field SS Optimized MS Optimized InSync Optimized
1.141.00 0.94
0.87
4.31
4.01
3.57
4.70
0
10
20
30
40
50
60
Main street Side street
Agg
rega
te D
elay
per
Veh
icle
(s)
Field SS Optimized MS Optimized InSync Optimized
1.681.73 1.63
1.28
4.84 4.89
4.454.95
0
10
20
30
40
50
60
Main street Side street
Agg
rega
te D
elay
per
Veh
icle
(s)
Field SS Optimized MS Optimized InSync Optimized
1.161.15 1.05
0.84
4.524.62
3.99
4.86
Figure 3.1: Main-Street vs. Side-Street Average Vehicle Delays
26
3.1.4 Network Performance
The effects of each analyzed system were also assessed on a network-wide level. The total
network delays were selected as the best representative network performance measure. A
comparison of total network delays for the four scenarios and each time period is given in Figure
3.2.
InSync yielded a total network delay reduction of about 17% for each time period, when
compared to the field signal timings. It performed better than any other optimization method,
with the best performance observed during the most congested PM peak period. At the same
time, SS optimized and MS optimized signal timings reduced delay, when compared to the field,
by 5-8% and 2-14%, respectively. The exception was SS optimized signal timings, which
increased delay by 4% in the PM peak.
3.1.5 Environmental Performance Measures
To assess the performance of various traffic signal control regimes in terms of environmental
outputs, the authors used the Comprehensive Modal Emission Model (CMEM) to obtain
estimates of fuel consumption and other emission outputs (12). Connections between traffic
microsimulation tools, such as VISSIM, and instantaneous emission models, such as CMEM,
have been described elsewhere (13). A program built in C++ connected VISSIM with the light-
duty vehicle (LDV) (98%) and heavy-duty diesel vehicle (HDD) (2%) models in CMEM. For
each vehicle, VISSIM provided simulation time, a vehicle identifier, a vehicle type (light-duty
vehicle or truck), speed and acceleration/deceleration on a second-by-second basis. The C++
interface program imported the VISSIM output file to CMEM, which depending on the vehicle
type, utilized either the LDV or HDD model and computed fuel consumption and vehicular
emissions (Carbon monoxide, CO; Hydrocarbons, HC; Nitrogen oxides, NOx; and Carbon
dioxide, CO2). The C++ program summarized values for each vehicle to obtain totals for the
entire network.
Table 3.4 shows the results of aggregated fuel consumption and emission estimates for various
traffic signal control regimes for the AM peak. The last column represents commonly used fuel
efficiency measures (miles/gallon). While most of the results are statistically different, it is
27
obvious the differences in fuel consumption are not large. InSync saved around 4% of fuel when
compared to field signal timings and 2% when compared to any of the optimized plans. It should
be noted here that increased fuel consumption is not only an outcome of degraded traffic
conditions (stops, congestion, etc.) but sometimes it is also associated with improvement in
traffic conditions (e.g. higher vehicular speed). Considering that InSync yielded higher average
speed than the other traffic signal control regimes, further research is needed to isolate the
impacts of two InSync-generated factors on fuel consumption: improved traffic progression
(reduces fuel consumption) and increased traffic speed (may increase fuel consumption).
28
0
100
200
300
400
500
Tota
l Net
wor
k D
elay
(h)
Field SS Optimized MS Optimized InSync Optimized
0
100
200
300
400
500
600
700
Tota
l Net
wor
k D
elay
(h)
Field SS Optimized MS Optimized InSync Optimized
0
100
200
300
400
500
600
700
800
900
Tota
l Net
wor
k D
elay
(h)
Field SS Optimized MS Optimized InSync Optimized
Figure 3.2: Total Network Delays
29
Table 3.4: Fuel Consumption and Emission Estimates
Scenario Statistic CO2 (Kg)
CO (Kg)
HC (Kg)
NOx (Kg)
Fuel (Kg)
Distance (Miles)
Fuel Eff. (Mi/gal)
Field Mean 13333.5 661.2 12.7 28.9 4534.3 26356.0 29.1
Std. Dev. 98.2 3.6 0.1 0.6 30.7 107.8 0.1Optimal SS
Mean 13137.4 645.3 12.5 28.4 4464.8 26492.3 29.7Std. Dev. 121.1 4.5 0.1 0.7 37.7 116.1 0.2
Optimal MS
Mean 13077.7 652.6 12.6 28.6 4449.8 26465.2 29.7*Std. Dev. 72.9 3.1 0.0 0.5 22.8 99.9 0.1
InSync Mean 12848.8 654.6 12.5 27.5 4379.6 26562.4 30.3
Std. Dev. 662.2 25.4 0.4 2.4 200.2 1152.7 0.4
* - value is not statistically different from corresponding SS Optimal value
3.1.6 Safety Performance Measures
The Surrogate Safety Assessment Model (SSAM) (14) was used to assess safety aspects of
InSync’s and TOD plans’ performances. For each scenario and each peak period, VISSIM
created trajectory data files later processed by SSAM to identify the frequency, severity and type
of conflicts (15). Table 3.5 shows estimates for three types of conflicts reported by SSAM, as
well as for the total number of conflicts (various conflicts reported by SSAM are defined in 14).
InSync reduced the number of rear-end and lane-changing conflicts, which were lowest among
all of the signal timing scenarios. However, crossing conflicts, which are often considered to be
more dangerous than the other conflict types (16), were the highest for the InSync scenario. It
seems that InSync increased crossing conflicts due to its flexible phasing design and the fact that
it gave more green time to major streets than other conventional control scenarios. Further
investigation is needed to document and justify differences in the numbers of various conflicts
reported by SSAM.
30
Table 3.5: Number of Estimated Safety Conflicts Between Vehicles
AM Peak Statistic/Conflict Total Crossing Rear-End Lane-Changing
Field Mean 2074.1 268.3 1516.8 289.0 Std. Deviation 44.1 16.2 30.6 23.2
Optimal SS Mean 1978.8 275.0 1412.6 291.2 Std. Deviation 53.7 16.5 45.0 12.1
Optimal MS Mean 2065.2 278.1 1498.1 289.0 Std. Deviation 34.5 19.6 32.5 20.3
InSync Mean 1951.0 302.0 1376.2 272.8 Std. Deviation 48.5 21.4 43.0 16.5
Midday Statistic/Conflict Total Crossing Rear-End Lane-Changing
Field Mean 2570.3 435.7 1819.8 314.8 Std. Deviation 76.1 28.3 61.6 18.7
Optimal SS Mean 2346.1 410.3 1636.8 299.0 Std. Deviation 38.2 22.1 35.7 14.2
Optimal MS Mean 2759.7 431.3 2005.4 323.0 Std. Deviation 33.4 20.4 34.2 28.5
InSync Mean 2354.4 485.7 1574.0 294.7 Std. Deviation 42.8 19.2 39.2 12.9
PM Peak Statistic/Conflict Total Crossing Rear-End Lane-Changing
Field Mean 2808.3 428.1 2015.7 364.5 Std. Deviation 72.8 21.3 79.8 14.2
Optimal SS Mean 2830.8 390.1 2084.5 356.2 Std. Deviation 74.2 20.5 63.9 20.3
Optimal MS Mean 2982.2 418.2 2179.2 384.8 Std. Deviation 77.7 9.8 74.8 28.6
InSync Mean 2750.2 473.1 1928.3 348.8 Std. Deviation 55.9 18.3 52.4 19.0
31
3.2. Traffic Operations for an I-95 Freeway Incident Situation
This operational scenario is evaluated for the PM peak only for the four traffic signal scenarios.
The evaluations are performed on the intersection, corridor and network level, as well as for
main-street vs. side-street performance.
3.2.1 Intersection Performance
Figures 3.3 - 3.5 show the comparison of average intersection delays, number of stops and
queues respectively, for the eleven analyzed intersections along the SR-421 corridor. For
intersection delays, InSync yielded better results than other signal timing plans for the majority
of intersections, except Williamson, I-95 SB and Spruce Creek Rd, where at least one other plan
showed better results. Averaged for all intersections, InSync yielded 11% lower delays than field
signal timings, 19% lower delays than SS optimized and 13% lower delays than MS optimized
timings. InSync showed the biggest benefits in delay reduction at the I-95NB interchange, as
well as the Clyde Morris Blvd and US-1 intersections. Similar results were observed for the
average number of stops per vehicle. On average, the number of stops observed for InSync was
8% lower than field, 13% lower than SS optimized and 18% lower than MS optimized timings.
InSync yielded the best results for the average queue results, with 9% shorter queues than field,
24% lower than SS optimized and 20% lower than MS optimized timings. The comparison of
intersection LOS results is given in Table 3.6. InSync never showed lower LOS than other plans,
while for the I-95 NB and Nova Rd intersections, the InSync LOS was higher than all other
plans.
32
0
20
40
60
80
100
120
Tot
al 2
hr
Inte
rsec
tion
Del
ay (
s)
Intersection
Field SS Optimized MS Optimized InSync Optimized
Figure 3.3: Comparison of Intersection Delays: Freeway Incident Scenario
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
Nu
mb
er o
f S
top
s p
er V
ehic
le
Intersection
Field SS Optimized MS Optimized InSync Optimized
Figure 3.4: Comparison of Number of Stops: Freeway Incident Scenario
33
0
50
100
150
200
250
300
350
400
450
Ave
rage
Qu
eue
(ft)
Intersection
Field SS Optimized MS Optimized InSync Optimized
Figure 3.5: Comparison of Average Queues: Freeway Incident Scenario
Table 3.6: LOS Comparison: Freeway Incident Scenario
Intersection Field SS Optimized MS Optimized InSync Optimized
Williamson D D C CI-95 SB C C C CI-95 NB D C D BTaylor rd B B B BYorktowne B B B BClyde Morris D E D DVictoria C C C CVillage C C C CNova E F F DSpruce C C C CUS 1 D D D D
LOS
34
3.2.2 Corridor Travel Times
InSync yielded the lowest travel times along the corridor in both directions under the freeway
incident scenario. Figures 3.6 and 3.7 show segment travel times in the EB and WB directions
respectively, while Figure 3.8 shows a comparison of travel times for the entire corridor. In the
EB direction, the total corridor travel times in the InSync scenario were 25%, 34% and 41%
lower than the field, SS optimized and MS optimized timings. The travel times in InSync were
also lower in the WB direction, although the savings were not as emphasized as in the EB
direction. For the entire corridor, the reported InSync travel times were 14%, 9% and 6% lower
than the field, SS optimized and MS optimized timings. The average LOS along the corridor in
the InSync scenario was B in both directions, which was significantly better than any other signal
timing plan, as shown in Table 3.7.
3.2.3 Main-Street vs. Side-Street Performance
InSync showed advantages over other signal timing plans along the main corridor and for the
main-street movements. However, for a complete assessment of intersection operations, it is
important to compare the performance of side-street movements as well. Figure 3.9 shows a
comparison of recorded delays for through and left movements on the main and side-streets
separately, while Figure 3.10 shows aggregated delays for main and side-streets. The results
show InSync increased delays for side-street through movements 29% to 38% when compared to
other plans; while for left turns, InSync decreased delays 8% to 22%. On average, the percentage
change in side-street delays for InSync varied between -2% (a decrease when compared to field
timings), to 11% (an increase when compared to MS optimized timings).
35
0
50
100
150
200
250
300
Tra
vel T
ime
(s)
Segments
Field SS Optimized MS Optimized InSync Optimized
Figure 3.6: Comparison of EB Segment Travel Times: Freeway Incident Scenario
0
50
100
150
Tra
vel T
ime
(s)
Segments
Field SS Optimized MS Optimized InSync Optimized
Figure 3.7: Comparison of WB Segment Travel Times: Freeway Incident Scenario
36
0
100
200
300
400
500
600
700
800
900
Tra
vel T
ime
(s)
Direction
Field SS Optimized MS Optimized InSync Optimized
Figure 3.8: Comparison of Corridor Travel Times: Freeway Incident Scenario
Table 3.7: Travel Speed and LOS Comparison: Freeway Incident Scenario
Field SS Optimized MS Optimized InSync Optimized
No. From To Speed (mph) LOS Speed (mph) LOS Speed (mph) LOS Speed (mph) LOS
1 Williamson I-95 SB 20.2 D 18.9 D 23.8 C 20.1 D2 I-95 SB I-95 NB 37.0 A 38.0 A 24.9 C 34.6 B3 I-95 NB Taylor 21.3 D 23.7 C 25.4 C 26.7 C4 Taylor Yorktowne 41.0 A 40.2 A 33.9 B 33.2 B5 Yorktowne Clyde Morris 27.3 C 17.0 E 18.8 D 27.7 C6 Clyde Morris Victoria 28.2 B 26.8 C 25.1 C 30.7 B7 Victoria/City Swallowtail 22.0 D 16.5 E 17.2 D 24.0 C8 Swallowtail Nova 9.2 F 6.5 F 6.4 F 19.6 D9 Nova Spruce 36.0 A 31.8 B 30.2 B 40.5 A
10 Spruce US 1 18.2 D 25.7 C 15.6 E 27.3 C
11 US 1 Spruce 30.7 B 37.9 A 33.2 B 31.6 B12 Spruce Nova 24.9 C 22.9 C 24.6 C 31.7 B13 Nova Swallowtail 28.7 B 35.2 A 36.6 A 32.3 B14 Swallowtail Victoria 30.5 B 28.0 B 28.4 B 27.7 C15 Victoria/City Clyde Morris 29.2 B 29.4 B 28.4 B 30.5 B16 Clyde Morris Yorktowne 36.7 A 24.1 C 29.1 B 33.5 B17 Yorktowne Taylor 32.9 B 37.6 A 34.0 B 37.3 A18 Taylor I-95 NB 20.3 D 22.4 C 16.1 E 24.4 C19 I-95 NB I-95 SB 23.1 C 24.5 C 15.8 E 25.8 C
20 I-95 SB Williamson 11.1 F 15.5 E 19.5 D 24.5 C
21 Williamson US 1 21.5 D 18.8 D 16.9 E 28.5 B
22 US 1 Williamson 26.6 C 28.0 C 28.9 B 30.8 B
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0
Agg
rega
te D
elay
per
Veh
icle
(s)
Street / Movement
Field SS Optimized MS Optimized InSync Optimized
Figure 3.9: Main-Street vs. Side-Street Movement Delays: Freeway Incident Scenario
0.0
10.0
20.0
30.0
40.0
50.0
60.0
Agg
rega
te D
elay
per
Veh
icle
(s)
Street
Field SS Optimized MS Optimized InSync Optimized
Figure 3.10: Main-Street vs. Side-Street: Aggregate Delays: Freeway Incident Scenario
38
3.2.4 Network Performance
On the network-wide level, InSync showed better results than other timing plans in all aspects.
Figures 3.11 - 3.13 show the comparison of total network delays, stops and travel time. In total
network delays, InSync reported 24%, 30% and 32% decreases compared to field, SS optimized
and MS optimized signal timings. Similar results were reported for the total network number of
stops. The total network travel time was 12% to 17% lower for the InSync scenario compared to
other timing plans.
The performance of InSync under the freeway incident scenario showed better results and
adaptability than other plans when compared to the existing field performance. The comparison
of results on the intersection, corridor and network levels is shown in Tables 3.8 - 3.10. The
summary of results for InSync compared to other timing plans for the incident operational
scenario is given in Figure 3.14. InSync outperformed other plans on all levels, except for side-
street performance, which was better for the MS optimized signal timings.
0
200
400
600
800
1000
1200
1400
Tot
al N
etw
ork
Del
ay (
h)
Signal Timings
Field SS Optimized MS Optimized InSync Optimized
Figure 3.11: Comparison of Total Network Delay: Freeway Incident Scenario
39
Figure 3.12: Comparison of Total Network Number of Stops: Freeway Incident Scenario
Figure 3.13: Comparison of Total Network Travel Times: Freeway Incident Scenario
40
Table 3.8: Regular vs. Incident Intersection Performance: Freeway Incident Scenario
Field SS MS InSync InSync vs Field InSync vs SS InSync vs MS
Regular 26 28 27 19 -28.9 -32.8 -31.0Incident 32 38 38 23 -28.8 -40.1 -40.9
Regular 51 51 46 48 -5.9 -5.8 5.1Incident 56 54 50 55 -1.7 1.0 10.6
Aggregate intersection delays (s) % Change
Main street
Side street
Table 3.9: Regular vs. Incident Corridor Performance: Freeway Incident Scenario
TT (s) Speed (mph) LOS TT (s) Speed (mph) LOS TT (s) Speed (mph) LOS TT (s) Speed (mph) LOS
Regular 556.8 25.5 C 593 24 C 603.1 23.6 C 476.2 29.9 BIncident 661.8 21.5 D 755.2 18.8 D 839.3 16.9 E 498.3 28.5 B
Regular 535.1 26.6 C 508.1 28 C 492.1 28.9 B 461.4 30.8 BIncident 530.3 26.8 C 506.4 28.1 B 491.6 28.9 B 467.6 30.4 B
EB
WB
Field SS Optimized MS Optimized InSync Optimized
InSync vs Field InSync vs SS InSync vs MS
Regular -14.5 -19.7 -21.0Incident -24.7 -34.0 -40.6
Regular -13.8 -9.2 -6.2Incident -11.8 -7.7 -4.9
EB
WB
% TT Change
Table 3.10: Regular vs. Incident Network Performance: Freeway Incident Scenario
Field SS Optimized MS Optimized InSync Optimized InSync vs Field InSync vs SS InSync vs MS
Regular 771.3 804.4 756.9 633.4 -17.9 -21.3 -16.3Incident 1031.2 1127.9 1160.9 784.2 -24.0 -30.5 -32.4
Regular 54738 55652 55294 50462 -7.8 -9.3 -8.7Incident 76966 77249 84786 57106 -25.8 -26.1 -32.6
Regular 1678.1 1710.3 1661.8 1539.0 -8.3 -10.0 -7.4Incident 1999.2 2094.1 2120.9 1756.2 -12.2 -16.1 -17.2
Total Delay Time (h)
% change
Total Number of Stops
Total Travel Time (h)
Figure 3.14: Incident Scenario Comparison on Different Levels: Freeway Incident Scenario
41
3.3. Traffic Operations under Inclement Weather Conditions
The evaluation of this operational scenario was performed for the PM peak only for the four
traffic signal scenarios. As with the freeway incident scenario, the evaluations were performed
on the intersection, corridor and network level, as well as for main-street vs. side-street
performance.
3.3.1 Intersection Performance
Figures 3.15 - 3.17 show the comparison of average intersection delays, number of stops and
queues. For intersection delays, InSync yielded better results than other signal timing plans for
the majority of intersections, except I-95 SB, where it resulted in the highest delays, and Spruce
Creek Rd and US-1, where the SS and MS optimized timings showed better results. Averaged
for all intersections, InSync yielded 17% lower delays than field signal timings, 12% lower
delays than SS optimized and 4% lower delays than MS optimized timings. InSync showed the
biggest benefits in delay reduction at the I-95NB interchange and the Nova Rd intersection.
More conservative savings with InSync implementation were observed for the average number
of stops per vehicle. On average, the average number of stops observed for InSync was 6% lower
than field, 1% lower than SS optimized and 4% lower than MS optimized timings.
The reported average queues in InSync were 19% shorter than the queues in the field scenario,
6% shorter than SS optimized queues and 2% longer than MS optimized timings queues. The
comparison of intersection LOS results is given in Table 3.11. InSync never showed lower LOS
than other plans, while for the Nova Rd intersection, the InSync LOS was higher than all other
plans.
42
0
10
20
30
40
50
60
70
80
90
100
Tot
al 2
hr
Inte
rsec
tion
Del
ay (
s)
Intersection
Field SS Optimized MS Optimized InSync Optimized
Figure 3.15: Comparison of Intersection Delays: Inclement Weather
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
Nu
mbe
r of
Sto
ps
per
Veh
icle
Intersection
Field SS Optimized MS Optimized InSync Optimized
Figure 3.16: Comparison of Number of Stops: Inclement Weather
43
0
50
100
150
200
250
300
350
400
Ave
rage
Qu
eue
(ft)
Intersection
Field SS Optimized MS Optimized InSync Optimized
Figure 3.17: Comparison of Average Queues: Inclement Weather
Table 3.11: LOS Comparison: Inclement Weather
Intersection Field SS Optimized MS Optimized InSync Optimized
Williamson E D C DI-95 SB C C C CI-95 NB A B B ATaylor rd B B B BYorktowne B C B BClyde Morris D D D DVictoria C C D CVillage C B B BNova E F F DSpruce E C D DUS 1 D C C D
LOS
44
3.3.2 Corridor Travel Times
InSync yielded the lowest travel times along the corridor in both directions under the inclement
weather scenario. Figures 3.18 and 3.19 show segment travel times in the EB and WB directions
respectively, while Figure 3.20 shows a comparison of travel times for the entire corridor. In the
EB direction, the total corridor travel times in the InSync scenario were 20%, 28% and 32%
lower than the field, SS optimized and MS optimized timings. The travel times in InSync were
also lower in the WB direction, although the savings are not as emphasized as in the EB
direction. For the entire corridor, the reported InSync travel times were 11%, 5% and 7% lower
than the field, SS optimized and MS optimized timings respectively. The average LOS along the
corridor is shown in Table 3.12. In the InSync scenario, the reported LOS was C in both
directions, which was better than any other signal timing plan in the EB direction, and the same
as other signal timing plans in the WB direction.
3.3.3 Main-Street vs. Side-Street Performance
Figure 3.21 shows a comparison of recorded delays for through and left movements on the main
and side-streets separately, while Figure 3.22 shows aggregated delays for the main and side-
streets. The results show InSync increased delays for side-street through movements 20% to 50%
when compared to other plans, while for left turns, the change in delays varied from -25% to 4%.
On average, the percentage change in side-street delays for InSync varied between -6% (a
decrease when compared to field timings) and 24% (an increase when compared to MS
optimized timings).
45
0
50
100
150
200
250
300
350
Tra
vel T
ime
(s)
Segments
Field SS Optimized MS Optimized InSync Optimized
Figure 3.18: Comparison of EB Segment Travel Times: Inclement Weather
0
50
100
150
Tra
vel T
ime
(s)
Segments
Field SS Optimized MS Optimized InSync Optimized
Figure 3.19: Comparison of WB Segment Travel Times: Inclement Weather
46
0
100
200
300
400
500
600
700
800
900
Tra
vel T
ime
(s)
Direction
Field SS Optimized MS Optimized InSync Optimized
Figure 3.20: Comparison of Corridor Travel Times: Inclement Weather
Table 3.12: Travel Speed and LOS Comparison: Inclement Weather
Field SS Optimized MS Optimized InSync Optimized
No. From To Speed (mph) LOS Speed (mph) LOS Speed (mph) LOS Speed (mph) LOS
1 Williamson I-95 SB 19.0 D 18.3 D 21.3 D 19.6 D2 I-95 SB I-95 NB 35.0 A 35.2 A 24.1 C 32.5 B3 I-95 NB Taylor 21.3 D 22.5 C 24.0 C 26.8 C4 Taylor Yorktowne 38.5 A 35.5 A 31.7 B 33.0 B5 Yorktowne Clyde Morris 27.6 C 17.7 D 18.2 D 27.0 C6 Clyde Morris Victoria 28.4 B 29.7 B 26.8 C 33.0 B7 Victoria/City Swallowtail 20.4 D 25.6 C 25.4 C 25.8 C8 Swallowtail Nova 9.2 F 6.1 F 6.2 F 19.5 D9 Nova Spruce 32.2 B 30.2 B 28.6 B 37.0 A
10 Spruce US 1 17.3 D 26.2 C 21.3 D 21.3 D
11 US 1 Spruce 19.2 D 33.0 B 24.6 C 31.5 B12 Spruce Nova 24.2 C 23.1 C 22.8 C 26.1 C13 Nova Swallowtail 26.9 C 33.9 B 34.5 B 32.5 B14 Swallowtail Victoria 27.4 C 25.9 C 25.6 C 23.6 C15 Victoria/City Clyde Morris 26.5 C 23.9 C 24.4 C 29.1 B16 Clyde Morris Yorktowne 30.8 B 20.7 D 25.4 C 29.4 B17 Yorktowne Taylor 32.8 B 33.7 B 32.3 B 36.4 A18 Taylor I-95 NB 19.8 D 22.2 C 16.3 E 26.8 C19 I-95 NB I-95 SB 23.2 C 21.8 D 15.0 E 23.0 C
20 I-95 SB Williamson 9.7 F 15.0 E 15.8 E 22.6 C
21 Williamson US 1 20.9 D 18.6 D 17.7 D 25.9 C
22 US 1 Williamson 24.3 C 25.9 C 25.4 C 27.2 C
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
Agg
rega
te D
elay
per
Veh
icle
(s)
Street / Movement
Field SS Optimized MS Optimized InSync Optimized
Figure 3.21: Main-Street vs. Side-Street Movement Delays: Inclement Weather
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
Agg
rega
te D
elay
per
Veh
icle
(s)
Street
Field SS Optimized MS Optimized InSync Optimized
Figure 3.22: Main-Street vs. Side-Street: Aggregate Delays: Inclement Weather
48
3.3.4 Network Performance
On the network-wide level, InSync showed better results than other timing plans in all aspects.
Figures 3.23 - 3.25 show the comparison of total network delays, stops and travel time. In total
network delays, InSync reported 23%, 18% and 15% decreases compared to field, SS optimized
and MS optimized signal timings. The total network number of stops reported by InSync was
12%, 7% and 9% lower than for field, SS optimized and MS optimized signal timings. The total
network travel time was 7% to 11% lower for the InSync scenario compared to other timing
plans.
The comparison of results between the existing and inclement weather scenarios on the
intersection, corridor and network levels is shown in Tables 3.13 - 3.15. The summary of results
for InSync compared to other timing plans for the inclement weather operational scenario is
given in Figure 3.26. Similarly to the freeway incident scenario, InSync outperformed other
plans on all levels, except for side-street performance, which was better for the MS optimized
signal timings.
0
200
400
600
800
1000
1200
Tot
al N
etw
ork
Del
ay (
h)
Signal Timings
Field SS Optimized MS Optimized InSync Optimized
Figure 3.23: Comparison of Total Network Delay: Inclement Weather
49
Figure 3.24: Comparison of Total Network Number of Stops: Inclement Weather
Figure 3.25: Comparison of Total Network Travel Times: Inclement Weather
50
Table 3.13: Regular vs. Incident Intersection Performance: Inclement Weather
Field SS MS InSync InSync vs Field InSync vs SS InSync vs MS
Regular 26 28 27 19 -28.9 -32.8 -31.0Inclement Weather 32 32 33 21 -33.8 -35.3 -36.1
Regular 51 51 46 48 -5.9 -5.8 5.1Inclement Weather 67 56 51 63 -6.7 12.5 23.9
Aggregate intersection delays (s) % Change
Main street
Side street
Table 3.14: Regular vs. Incident Corridor Performance: Inclement Weather
TT (s) Speed (mph) LOS TT (s) Speed (mph) LOS TT (s) Speed (mph) LOS TT (s) Speed (mph) LOS
Regular 556.8 25.5 C 593.0 24.0 C 603.1 23.6 C 476.2 29.9 BInclement Weather 681.8 20.9 D 763.5 18.6 D 802.2 17.7 D 548.5 25.9 C
Regular 535.1 26.6 C 508.1 28.0 C 492.1 28.9 B 461.4 30.8 BInclement Weather 585.0 24.3 C 548.0 25.9 C 559.3 25.4 C 521.6 27.2 C
EB
WB
Field SS Optimized MS Optimized InSync Optimized
InSync vs Field InSync vs SS InSync vs MS
Regular -14.5 -19.7 -21.0Inclement Weather -19.5 -28.2 -31.6
Regular -13.8 -9.2 -6.2Inclement Weather -10.8 -4.8 -6.7
% TT Change
EB
WB
Table 3.15: Regular vs. Incident Network Performance: Inclement Weather
Field SS Optimized MS Optimized InSync Optimized InSync vs Field InSync vs SS InSync vs MS
Regular 771.3 804.4 756.9 633.4 -17.9 -21.3 -16.3Inclement Weather 1007.7 941.4 911.7 771.0 -23.5 -18.1 -15.4
Regular 54738 55652 55294 50462 -7.8 -9.3 -8.7Inclement Weather 64432 61488 62387 56929 -11.6 -7.4 -8.7
Regular 1678.1 1710.3 1661.8 1539.0 -8.3 -10.0 -7.4Inclement Weather 2007.9 1949.7 1919.2 1779.4 -11.4 -8.7 -7.3
Total Delay Time (h)
% change
Total Number of Stops
Total Travel Time (h)
Figure 3.26: Incident Scenario Comparison on Different Levels: Inclement Weather
51
3.4. Traffic Operations under Frequent Rail Preemption Calls
The evaluation of this operational scenario was performed for the PM peak only for the four
traffic signal scenarios. As with the previous two scenarios, the evaluations were performed on
the intersection, corridor and network level, as well as for main-street vs. side-street
performance.
3.4.1 Intersection Performance
Figures 3.27 - 3.29 show the comparison of average intersection delays, number of stops and
queues. The results under this scenario were similar to the results obtained for the existing
conditions. For intersection delays, InSync showed better results than other signal timing plans
for the majority of intersections. Averaged for all intersections, InSync yielded 16% lower delays
than field signal timings, 19% lower delays than SS optimized and 8% lower delays than MS
optimized timings. The average number of stops per vehicle observed for InSync was about 7%
lower than all other signal timing plans.
The reported average queues in InSync were 22%, 23% and 11% shorter than the queues in the
field, SS optimized and MS optimized timing scenarios. The comparison of intersection LOS
results is given in Table 3.16. InSync never showed lower LOS than other plans, while for the
Clyde Morris Blvd and Nova Rd intersections, the InSync LOS was higher than all other plans.
52
0
10
20
30
40
50
60
70
80
Tot
al 2
hr
Inte
rsec
tion
Del
ay (
s)
Intersection
Field SS Optimized MS Optimized InSync Optimized
Figure 3.27: Comparison of Intersection Delays: Rail Preemption
0.00
0.20
0.40
0.60
0.80
1.00
1.20
Nu
mb
er o
f S
tops
per
Veh
icle
Intersection
Field SS Optimized MS Optimized InSync Optimized
Figure 3.28: Comparison of Number of Stops: Rail Preemption
53
0
50
100
150
200
250
Ave
rage
Qu
eue
(ft)
Intersection
Field SS Optimized MS Optimized InSync Optimized
Figure 3.29: Comparison of Average Queues: Rail Preemption
Table 3.16: LOS Comparison: Rail Preemption
Intersection Field SS Optimized MS Optimized InSync Optimized
Williamson D D C CI-95 SB C C C CI-95 NB A B B ATaylor rd B B B BYorktowne B B B BClyde Morris D D D CVictoria C C C CVillage C B B BNova E E E DSpruce C C C CUS 1 D C C CR Xing A A A A
LOS
3.4.2 Corridor Travel Times
The corridor travel times under the rail preemption scenario also showed similarities to the
existing conditions. InSync yielded the lowest travel times along the corridor in both directions
54
under the rail preemption scenario. Figures 3.30 and 3.31 show segment travel times in the EB
and WB directions respectively, while Figure 3.32 shows a comparison of travel times for the
entire corridor. In the EB direction, the total corridor travel times in the InSync scenario were
18%, 22% and 24% lower than the field, SS optimized and MS optimized timings. In the WB
direction, the reported InSync travel times were 15%, 12% and 9% lower than the field, SS
optimized and MS optimized timings. The average LOS along the corridor is shown in Table
3.17. In the InSync scenario, the reported LOS was B in both directions, which was better than
other signal timing plans.
3.4.3 Main-Street vs. Side-Street Performance
Figure 3.33 shows a comparison of recorded delays for through and left movements on the main
and side-streets separately, while Figure 3.34 shows aggregated delays for main and side-streets.
The results show InSync increased delays for side-street through movements 16% to 24% when
compared to other plans, while for left turns, the delays decreased between 6% and 21%. On
average, the percentage change in side-street delays for InSync varied between -4% (a decrease
when compared to field timings) and 7% (an increase when compared to MS optimized timings).
0
50
100
150
200
Tra
vel T
ime
(s)
Segments
Field SS Optimized MS Optimized InSync Optimized
Figure 3.30: Comparison of EB Segment Travel Times: Rail Preemption
55
0
50
100
150
Tra
vel T
ime
(s)
Segments
Field SS Optimized MS Optimized InSync Optimized
Figure 3.31: Comparison of WB Segment Travel Times: Rail Preemption
0
100
200
300
400
500
600
700
Tra
vel T
ime
(s)
Direction
Field SS Optimized MS Optimized InSync Optimized
Figure 3.32: Comparison of Corridor Travel Times: Rail Preemption
56
Table 3.17: Travel Speed and LOS Comparison: Rail Preemption
Field SS Optimized MS Optimized InSync Optimized
No. From To Speed (mph) LOS Speed (mph) LOS Speed (mph) LOS Speed (mph) LOS
1 Williamson I-95 SB 20.0 D 20.2 D 23.3 C 20.5 D2 I-95 SB I-95 NB 37.9 A 38.3 A 26.6 C 34.6 B3 I-95 NB Taylor 22.8 C 25.4 C 27.0 C 29.4 B4 Taylor Yorktowne 41.7 A 41.0 A 34.9 B 35.4 A5 Yorktowne Clyde Morris 28.7 B 19.5 D 20.5 D 30.8 B6 Clyde Morris Victoria 28.7 B 29.0 B 26.2 C 34.1 B7 Victoria/City Swallowtail 25.1 C 29.0 B 31.9 B 31.7 B8 Swallowtail Nova 14.6 E 10.8 F 10.2 F 25.4 C9 Nova Spruce 36.2 A 31.9 B 34.1 B 45.8 A
10 Spruce US 1 18.9 D 25.6 C 25.6 C 25.0 C
11 US 1 Spruce 29.5 B 35.1 A 31.6 B 33.4 B12 Spruce Nova 25.9 C 23.0 C 24.7 C 32.2 B13 Nova Swallowtail 28.1 B 34.8 B 36.0 A 36.9 A14 Swallowtail Victoria 30.5 B 28.0 B 28.3 B 27.2 C15 Victoria/City Clyde Morris 29.9 B 29.1 B 28.3 B 32.1 B16 Clyde Morris Yorktowne 36.4 A 24.2 C 29.6 B 36.0 A17 Yorktowne Taylor 32.9 B 37.9 A 34.0 B 39.2 A18 Taylor I-95 NB 21.4 D 24.9 C 17.3 D 28.8 B19 I-95 NB I-95 SB 22.4 C 24.5 C 15.6 E 25.0 C
20 I-95 SB Williamson 11.7 F 16.2 E 19.4 D 24.9 C
21 Williamson US 1 25.4 C 24.0 C 23.4 C 31.0 B
22 US 1 Williamson 26.7 C 27.7 C 28.7 B 31.4 B
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
Agg
rega
te D
elay
per
Veh
icle
(s)
Street / Movement
Field SS Optimized MS Optimized InSync Optimized
Figure 3.33: Main-Street vs. Side-Street Movement Delays: Rail Preemption
0.0
10.0
20.0
30.0
40.0
50.0
60.0
Agg
rega
te D
elay
per
Veh
icle
(s)
Street
Field SS Optimized MS Optimized InSync Optimized
Figure 3.34: Main-Street vs. Side-Street: Aggregate Delays: Rail Preemption
58
3.4.4 Network Performance
On the network-wide level, InSync showed better results than other timing plans in all aspects.
Figures 3.35 - 3.37 show the comparison of total network delays, stops and travel time. In total
network delays, InSync reported 18%, 21% and 17% decreases compared to field, SS optimized
and MS optimized signal timings. The total network number of stops reported by InSync was 9%
lower than for field and 11% lower than the SS and MS optimized signal timings. The total
network travel time was 8% to 10% lower for the InSync scenario compared to other timing
plans.
The comparison of results between the existing and rail preemption scenarios on the intersection,
corridor and network levels is shown in Tables 3.18 - 3.20. The performance between the two
operational scenarios was similar. The summary of results for InSync compared to other timing
plans for the rail preemption operational scenario is given in Figure 3.38. Similar to the other
scenarios, InSync outperformed other plans on all levels, except for side-street performance,
which was better for the MS optimized signal timings.
0
100
200
300
400
500
600
700
800
900
Tot
al N
etw
ork
Del
ay (
h)
Signal Timings
Field SS Optimized MS Optimized InSync Optimized
Figure 3.35: Comparison of Total Network Delay: Rail Preemption
59
Figure 3.36: Comparison of Total Network Number of Stops: Rail Preemption
Figure 3.37: Comparison of Total Network Travel Times: Rail Preemption
60
Table 3.18: Regular vs. Incident Intersection Performance: Rail Preemption
Field SS MS InSync InSync vs Field InSync vs SS InSync vs MS
Regular 26 28 27 19 -28.9 -32.8 -31.0Rail Preemption 26 28 27 18 -30.7 -34.4 -33.1
Regular 51 51 46 48 -5.9 -5.8 5.1Rail Preemption 51 51 46 49 -4.1 -3.9 7.3
Aggregate intersection delays (s) % Change
Main street
Side street
Table 3.19: Regular vs. Incident Corridor Performance: Rail Preemption
TT (s) Speed (mph) LOS TT (s) Speed (mph) LOS TT (s) Speed (mph) LOS TT (s) Speed (mph) LOS
Regular 556.8 25.5 C 593 24 C 603.1 23.6 C 476.2 29.9 BRail Preemption 560.6 25.4 C 591.6 24.0 C 606.6 23.4 C 459.2 31.0 B
Regular 535.1 26.6 C 508.1 28 C 492.1 28.9 B 461.4 30.8 BRail Preemption 532.1 26.7 C 513.6 27.7 C 494.5 28.7 B 452.5 31.4 B
EB
WB
Field SS Optimized MS Optimized InSync Optimized
InSync vs Field InSync vs SS InSync vs MS
Regular -14.5 -19.7 -21.0Rail Preemption -18.1 -22.4 -24.3
Regular -13.8 -9.2 -6.2Rail Preemption -15.0 -11.9 -8.5
% TT Change
EB
WB
Table 3.20: Regular vs. Incident Network Performance: Rail Preemption
Field SS Optimized MS Optimized InSync Optimized InSync vs Field InSync vs SS InSync vs MS
Regular 771.3 804.4 756.9 633.4 -17.9 -21.3 -16.3Rail Preemption 776.4 807.3 762.0 635.1 -18.2 -21.3 -16.6
Regular 54738 55652 55294 50462 -7.8 -9.3 -8.7Rail Preemption 54942 55727 55581 49751 -9.4 -10.7 -10.5
Regular 1678.1 1710.3 1661.8 1539.0 -8.3 -10.0 -7.4Rail Preemption 1683.1 1713.7 1667.3 1540.6 -8.5 -10.1 -7.6
Total Delay Time (h)
% change
Total Number of Stops
Total Travel Time (h)
Figure 3.38: Incident Scenario Comparison on Different Levels: Rail Preemption
61
3.5. Traffic Operations under Oversaturated Conditions
The evaluation of this operational scenario was performed for the PM peak only for the four
traffic signal scenarios. As with the previous scenarios, the evaluations were performed on the
intersection, corridor and network levels, as well as for main-street vs. side-street performance.
3.5.1 Intersection Performance
Figures 3.39 - 3.41 show the comparison of average intersection delays, number of stops and
queues, respectively. For intersection delays, InSync showed better results than other signal
timing plans for the majority of intersections. Averaged for all intersections, InSync yielded 17%
lower delays than field signal timings, 18% lower delays than SS optimized and 8% lower delays
than MS optimized timings. The average number of stops per vehicle observed for InSync was
about 8% lower than for field signal timings and 11% lower than SS and MS optimized signal
timing plans.
The reported average queues in InSync were 19%, 22% and 4% shorter than the queues in the
field, SS optimized and MS optimized timings scenarios. The comparison of intersection LOS
results is given in Table 3.21. InSync never showed lower LOS than other plans, while for the I-
95 NB interchange, Taylor Rd, Village Rd and Nova Rd intersections, the InSync LOS was
higher than all other plans.
62
0
20
40
60
80
100
120
Tot
al 2
hr
Inte
rsec
tion
Del
ay (
s)
Intersection
Field SS Optimized MS Optimized InSync Optimized
Figure 3.39: Comparison of Intersection Delays: Oversaturation
0.00
0.20
0.40
0.60
0.80
1.00
1.20
1.40
1.60
1.80
2.00
Nu
mb
er o
f S
top
s p
er V
ehic
le
Intersection
Field SS Optimized MS Optimized InSync Optimized
Figure 3.40: Comparison of Number of Stops: Oversaturation
63
0
50
100
150
200
250
300
350
400
450
500A
vera
ge Q
ueu
e (f
t)
Intersection
Field SS Optimized MS Optimized InSync Optimized
Figure 3.41: Comparison of Average Queues: Oversaturation
Table 3.21: LOS Comparison: Oversaturation
Intersection Field SS Optimized MS Optimized InSync Optimized
Williamson E D C DI-95 SB D D C DI-95 NB B B B ATaylor rd C C C BYorktowne C D C CClyde Morris E E E EVictoria D E E DVillage D F E BNova F F F ESpruce F D C DUS 1 D D C D
LOS
64
3.5.2 Corridor Travel Times
The corridor travel times in InSync under the oversaturated conditions yielded the lowest travel
times along the corridor in both directions. Figures 3.42 and 3.43 show segment travel times in
the EB and WB directions respectively, while Figure 3.44 shows a comparison of travel times for
the entire corridor. In the EB direction, the total corridor travel times in the InSync scenario were
18%, 22% and 24% lower than the field, SS optimized and MS optimized timings. In the WB
direction, the reported InSync travel times were 15%, 12% and 9% lower than the field, SS
optimized and MS optimized timings. The average LOS along the corridor is shown in Table
3.22. In the InSync scenario, the reported LOS was C in both directions, which was significantly
better than other signal timing plans in the EB direction, and equal to other signal timing plans in
the WB direction.
3.5.3 Main-Street vs. Side-Street Performance
Figure 3.45 shows a comparison of recorded delays for through and left movements on the main
and side-streets separately, while Figure 3.46 shows aggregated delays for the main and side-
streets. The results show InSync increased delays for side-street through movements 20% to 55%
when compared to other plans, while for left turns, the delays varied between -19% (a decrease
compared to the field timings) and 7% (an increase compared to the MS optimized timings). On
average, the percentage change in side-street delays for InSync varied between -4% (a decrease
when compared to field timings) and 26% (an increase when compared to MS optimized
timings).
65
0
50
100
150
200
250
300
350
400
450
500
Tra
vel T
ime
(s)
Segments
Field SS Optimized MS Optimized InSync Optimized
Figure 3.42: Comparison of EB Segment Travel Times: Oversaturation
0
50
100
150
Tra
vel T
ime
(s)
Segments
Field SS Optimized MS Optimized InSync Optimized
Figure 3.43: Comparison of WB Segment Travel Times: Oversaturation
66
0
200
400
600
800
1000
1200
Tra
vel T
ime
(s)
Direction
Field SS Optimized MS Optimized InSync Optimized
Figure 3.44: Comparison of Corridor Travel Times: Oversaturation
Table 3.22: Travel Speed and LOS Comparison: Oversaturation
Field SS Optimized MS Optimized InSync Optimized
No. From To Speed (mph) LOS Speed (mph) LOS Speed (mph) LOS Speed (mph) LOS
1 Williamson I-95 SB 18.0 D 16.6 E 20.3 D 18.7 D2 I-95 SB I-95 NB 37.4 A 37.8 A 23.5 C 32.7 B3 I-95 NB Taylor 21.7 D 20.2 D 22.4 C 26.1 C4 Taylor Yorktowne 40.8 A 38.1 A 32.5 B 35.2 A5 Yorktowne Clyde Morris 28.0 B 17.7 D 18.3 D 29.5 B6 Clyde Morris Victoria 27.8 C 15.9 E 14.2 E 33.7 B7 Victoria/City Swallowtail 17.2 D 6.8 F 7.1 F 25.7 C8 Swallowtail Nova 5.1 F 3.8 F 3.9 F 15.4 E9 Nova Spruce 35.8 A 30.8 B 32.8 B 44.9 A
10 Spruce US 1 18.1 D 27.3 C 25.6 C 25.5 C
11 US 1 Spruce 19.8 D 36.0 A 30.6 B 35.6 A12 Spruce Nova 24.2 C 22.7 C 22.9 C 28.7 B13 Nova Swallowtail 27.4 C 33.5 B 35.2 A 34.8 B14 Swallowtail Victoria 26.8 C 24.1 C 24.5 C 24.0 C15 Victoria/City Clyde Morris 18.4 D 13.4 E 13.9 E 18.0 D16 Clyde Morris Yorktowne 18.1 D 14.5 E 16.5 E 16.2 E17 Yorktowne Taylor 31.2 B 33.4 B 31.0 B 34.3 B18 Taylor I-95 NB 17.1 D 19.4 D 13.1 E 21.3 D19 I-95 NB I-95 SB 19.4 D 20.4 D 13.1 E 19.4 D
20 I-95 SB Williamson 9.7 F 14.2 E 16.6 E 23.0 C
21 Williamson US 1 16.9 E 12.6 F 12.5 F 26.8 C
22 US 1 Williamson 22.5 C 22.8 C 23.1 C 25.5 C
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
90.0
100.0
Agg
rega
te D
elay
per
Veh
icle
(s)
Street / Movement
Field SS Optimized MS Optimized InSync Optimized
Figure 3.45: Main-Street vs. Side-Street Movement Delays: Oversaturation
0.0
10.0
20.0
30.0
40.0
50.0
60.0
70.0
80.0
Agg
rega
te D
elay
per
Veh
icle
(s)
Street
Field SS Optimized MS Optimized InSync Optimized
Figure 3.46: Main-Street vs. Side-Street: Aggregate Delays: Oversaturation
68
3.5.4 Network Performance
On the network-wide level, InSync showed better results than other timing plans in all aspects.
Figures 3.47 - 3.49 show the comparison of total network delays, stops and travel time. In total
network delays, InSync reported 26%, 30% and 28% decreases compared to field, SS optimized
and MS optimized signal timings. The total network number of stops reported by InSync was
15%, 20% and 19% lower than for field, SS and MS optimized signal timings. The total network
travel time was 15% to 18% lower for the InSync scenario compared to other timing plans.
The comparison of results between the current (regular) and oversaturation scenarios on the
intersection, corridor and network levels is shown in Tables 3.23 - 3.25. The summary of results
for InSync compared to other timing plans for the oversaturation operational scenario is given in
Figure 3.50. Similar to the other scenarios, InSync outperformed other plans on all levels, except
for side-street performance, which was better for the MS optimized signal timings.
0
200
400
600
800
1000
1200
1400
1600
1800
2000
Tot
al N
etw
ork
Del
ay (
h)
Signal Timings
Field SS Optimized MS Optimized InSync Optimized
Figure 3.47: Comparison of Total Network Delay: Oversaturation
69
Figure 3.48: Comparison of Total Network Number of Stops: Oversaturation
Figure 3.49: Comparison of Total Network Travel Times: Oversaturation
70
Table 3.23: Regular vs. Incident Intersection Performance: Oversaturation
Field SS MS InSync InSync vs Field InSync vs SS InSync vs MS
Regular 26 28 27 19 -28.9 -32.8 -31.0Oversaturation 44 53 52 28 -35.4 -46.6 -45.7
Regular 51 51 46 48 -5.9 -5.8 5.1Oversaturation 78 65 59 75 -3.5 15.0 26.3
Aggregate intersection delays (s) % Change
Main street
Side street
Table 3.24: Regular vs. Incident Corridor Performance: Oversaturation
TT (s) Speed (mph) LOS TT (s) Speed (mph) LOS TT (s) Speed (mph) LOS TT (s) Speed (mph) LOS
Regular 556.8 25.5 C 593.0 24.0 C 603.1 23.6 C 476.2 29.9 BOversaturation 842.9 16.9 E 1128.2 12.6 F 1137.2 12.5 F 529.9 26.8 C
Regular 535.1 26.6 C 508.1 28.0 C 492.1 28.9 B 461.4 30.8 BOversaturation 631.5 22.5 C 624.2 22.8 C 615.7 23.1 C 556.3 25.5 C
EB
WB
Field SS Optimized MS Optimized InSync Optimized
InSync vs Field InSync vs SS InSync vs MS
Regular -14.5 -19.7 -21.0Oversaturation -37.1 -53.0 -53.4
Regular -13.8 -9.2 -6.2Oversaturation -11.9 -10.9 -9.6
% TT Change
EB
WB
Table 3.25: Regular vs. Incident Network Performance: Oversaturation
Field SS Optimized MS Optimized InSync Optimized InSync vs Field InSync vs SS InSync vs MS
Regular 771.3 804.4 756.9 633.4 -17.9 -21.3 -16.3Oversaturation 1660.2 1773.8 1706.1 1236.5 -25.5 -30.3 -27.5
Regular 54738 55652 55294 50462 -7.8 -9.3 -8.7Oversaturation 101413 108797 106976 86520 -14.7 -20.5 -19.1
Regular 1678.1 1710.3 1661.8 1539.0 -8.3 -10.0 -7.4Oversaturation 2753.8 2872.5 2806.5 2350.3 -14.7 -18.2 -16.3
Total Delay Time (h)
% change
Total Number of Stops
Total Travel Time (h)
Figure 3.50: Incident Scenario Comparison on Different Levels: Oversaturation
71
4. CONCLUSIONS
The following generalized conclusions can be reached based on the results of this study:
Under the existing conditions, InSync outperformed TOD signal timings in terms of
traffic efficiency. For this corridor and its traffic demand, InSync was undoubtedly better
than existing signal timings and two optimal signal timing plans in terms of overall
network performances (delay, stops, average speed, etc.), corridor travel times,
intersection delays and stops, and main-street delays. The only aspect where InSync was
not clearly better than any other TOD signal timings was side-street delay, where InSync
was neither best nor worst.
Under the existing conditions, InSync outperformed TOD signal timings in terms of fuel
efficiency and most of the emissions outputs. Benefits from InSync were not large,
although they were statistically significant: InSync saved between 2-4% when compared
to other timing plans. It is speculated that relatively low savings in fuel consumption
were due to the nonlinear relationship between fuel consumption and traffic efficiency
metrics. Thus, InSync (and other signal timing regimes) sometimes get penalized in
higher fuel consumption for providing better LOS and higher travel speeds.
Under the existing conditions, InSync outperformed TOD signal timings in terms of total
number of vehicular conflicts – a surrogate safety measure reported by SSAM. InSync
was better in terms of total, rear-end and lane-changing conflicts. InSync yielded a higher
number of crossing conflicts than the TOD scenarios. Further research is needed to
investigate the relationship between InSync and TOD operations and numbers of various
conflicts reported by SSAM.
The performance of the four signal timing plans was evaluated under different
operational scenarios: freeway incident, inclement weather, frequent rail preemption and
oversaturation. Under all operational scenarios, InSync outperformed other signal timing
plans on the intersection, corridor and network-wide levels. Traffic operations on side-
72
streets were the only aspect where InSync was outperformed by MS (and in some cases
SS) optimized signal timings.
InSync savings reported in this study were similar to those reported in the study performed by
HDR (2). However, savings reported by most of the field evaluation studies (5-10) were much
higher than the findings from microsimulation studies. This difference opens room for
speculations – maybe most of the TOD signal timings assessed in before and after field
evaluation studies were outdated; or, perhaps simulation models do not provide enough traffic
variations to yield higher benefits. Other speculations may also apply. Further research is
necessary to investigate the differences in benefits reported in field and simulation studies.
5. REFERENCES
1. Rhythm Engineering – Projects. Website: http://rhythmtraffic.com. Accessed July, 2012.
2. HDR Inc. An unpublished study on InSync performance in VISSIM. Website: http://fixcongestion.com/validation.htm. Accessed July, 2012.
3. Chandra et al. (2012). External Adaptive Control Systems and Methods. US Patent No. 8103436 B1. Website: http://www.google.com/patents/US8103436. Accessed July, 2012.
4. Chandra et al. (2012). Adaptive Control Systems and Methods. US Patent No. 8050854 B1. Website: http://www.google.com/patents/US8050854. Accessed July, 2012.
5. Hutton, J. M., Bokenkroger, C. D., and Meyer, M. M. (2010). “Evaluation of an Adaptive Traffic Signal System: Route 291 in Lee’s Summit”, Report prepared by Midwest Research Institute and Missouri Department of Transportation.
6. DKS Associates (2010). “Evaluation of an Adaptive Traffic Signal System – Before and After Study: Crow Canyon Road and Bollinger Canyon Road in San Ramon, California.” Report prepared for the City of San Ramon.
7. Pennoni Associates Inc. (2010). “Traffic Signal System Comparison Route 202 and Gulph Road/Mall Boulevard Upper Merion Township, Montgomery County”. Report prepared for Pennsylvania Department of Transportation.
8. TJKM Transportation Consultants. (2011). “Evaluation of Main-Street Adaptive Traffic Signal System.” Report prepared for the City of Salinas.
9. HDR Engineering, Inc. (2011). “US Route 17 Adaptive Traffic Control System: Phase I Before and After Travel Time and Delay Study.” Report prepared for Town of Mt. Pleasant, SC.
10. Hathaway, E., Urbanik, T., and Tsoi, S. (Kittleson and Associates, Inc.). (2012). “Cornell Road InSync System Evaluation.” Project #: 11075 Memorandum. Prepared for Oregon Department of Transportation.
11. Goodwin , L. C., and Pisano, P. A. Weather Responsive Traffic Signal Control, ITE Journal, June 2004.
12. Scora, G., Barth M. (2006). Comprehensive Modal Emission Model (CMEM) version 3.01 User’s Guide, University of California, Riverside, California.
13. Stevanovic, A., Stevanovic, J., Zhang, K., and Batterman, S. (2009), “Optimizing Traffic Control to Reduce Fuel Consumption and Vehicular Emissions: An Integrated Approach of VISSIM, CMEM, and VISGASOT”, Transportation Research Record: Journal of the Transportation Research Board, 2128, 105–113.
14. Gettman, D. and Head, L. "Surrogate Safety Measures From Traffic Simulation Models", Report No. FHWA-RD-03-050, FHWA, Washington, DC, 2003.
15. Stevanovic, A.Z., Stevanovic, J., and Jolovic, D. (2012). “Retiming Traffic Signals to Minimize Surrogate Safety Measures on Signalized Road Networks.” Presented at the 91st TRB Annual Meeting, Transportation Research Board, Washington D.C.
74
16. Gettman, D.,Pu, L., Sayed, T., and Shelby, S.G. "Surrogate Safety Assessment Model and Validation: Final Report", Report No. FHWA-HRT-08-051, Federal Highway Administration, Washington, DC, 2008.