Upload
others
View
0
Download
0
Embed Size (px)
Citation preview
Smart Working Environments for All Ages
This project has received funding from the European Union’s Horizon 2020 research and
innovation programme under grant agreement N. 826232
The content of this deliverable does not reflect the official opinion of the European
Union. Responsibility for the information and views expressed therein lies entirely with the
author(s)
D2.5 Field Test Strategy
D2.5 Field Test Strategy
2
WP2 – User Centric Design
D2.5 – Field Test Strategy
Project number: 826232
Project acronym: WorkingAge
Project title: Smart Working Environments for All Ages
Author(s): J. Czerniak, S. Comai, V. Rick, L. Sbattella
Partners contributed: POLIMI, RWTH
Delivery date: M8 (September 2019)
Document date: 17/10/2019
Version: 1
Revision: 1
Deliverable type: Report
Remarks:
Status: PU (Public)
PP Restricted to other programme participants (including
the Commission Services)
Restricted to a group specified by the consortium (including
the Commission Services) (please specify the group)
Confidential, only for members of the consortium (including
the Commission Services
D2.5 Field Test Strategy
3
Document Revision Log
Acronyms and Terminology
The following table reports the acronyms used in this deliverable.
Term Definition
AR Augmented Reality
ECG Electrocardiogram
EEG Electroencephalogram
EMG Electromyography
EOG Electrooculogram
GA Grupo Antolín
GSR Galvanic Skin Response
HCI Human Computer Interaction
ST Skin Temperature
VERSION REVISION DATE DESCRIPTION AUTHOR
0 0 12/9/19 Just a starting point to foster
discussion.
S. Comai, L. Sbattella
0 1 23/9/19 Description of use cases L. Sbattella
0 2 27/9/19 Info on lab tests J. Czerniak, V. Rick
0 3 1/10/19 Info on tests L. Sbattella
0 4 3/10/19 Info on field tests L. Sbattella
0 5 3/10/19 Review M. van Gasteren
0 6 3/10/19 Review G. Pelosi
0 7 7/10/19 System integration in Lab tests L. Sbattella
0 8 8/10/19 Sensor about “pose” for the
Driving use case
L. Sbattella
0 9 15/10/19 Details about planning and
size of the pilot. Pilot for
“home environments”. Fixes
to tables on sensor
estimations. Added the
“smartband” sensor.
Review by BS and TPZ.
Added figures from DoA for
supporting test schedule.
Time schedule fixed to meet
Tab 5 of DoA part B.
Added the Skin Temperature
and gesture sensors.
L. Sbattella,
B. Valentin, V.Ronca,
G. Borghini
1 0 17/10/19 TPZ sensor added to tab 5 L. Sbattella
1 1 18/10/19 Final review before submission M. van Gasteren
D2.5 Field Test Strategy
4
Executive Summary
This deliverable is one of the outputs of the project task T2.5 DEFINITION OF PILOT
TESTS, whose main objective is to address the specifications of the pilot
applications to be deployed.
D2.5 defines the how & when of the lab & company pilots: time schedule,
workplaces, description of sample workers.
Notice that the protocol for system evaluation is given in D2.6. D2.5 provides
guidelines for preparing and conducting the tests.
This document is the first release of D2.5, the second one being scheduled for
M16. Thus, the document provides as we currently envision the pilot test
strategy. Notice, however, that the second release of D2.5 could bring many
changes due to refinements or more deep changes.
The output of the work depicted in D2.5 and D2.6 will be the basis for WP3, WP4,
and WP9; in particular WP9 is in charge of providing document D9.1 Pilot
Operational Manual specifying the field test protocol.
This document is structured as follows:
Section 1 introduces the goals of the test strategy and its relationships
with D2.6;
Section 2 describes the use cases;
Section 3 specifies what we are going to measure;
Section 4 focuses on the description of Lab Tests;
Section 5 introduces the Field Tests;
Section 6 further specifies short-term Field Tests;
Section 7 further specifies long-term Field Tests.
D2.5 Field Test Strategy
5
Table of Contents Executive Summary........................................................................................................ 4
1 Introduction ............................................................................................................. 7
1.1 Goals of the test strategy .............................................................................. 7
1.2 Main interrelationships with other deliverables .......................................... 7
2 Use cases.................................................................................................................. 9
2.1 Office ............................................................................................................... 9
2.2 Driving .............................................................................................................. 9
2.3 Manufacturing ................................................................................................ 9
2.3.1 Workplace ................................................................................................... 9
2.3.2 Supporting infrastructure ......................................................................... 10
2.3.3 Possible issues affecting sensor deployment ........................................ 10
2.3.4 Privacy concerns ...................................................................................... 10
2.3.5 Health and stress problems ..................................................................... 10
2.3.6 Human Machine Interaction .................................................................. 10
2.3.7 Notes and issues ....................................................................................... 10
3 Measurements and sensors ................................................................................. 11
3.1 Sensors ........................................................................................................... 11
3.2 Questionnaires .............................................................................................. 12
3.3 Measurement-use case matrix ................................................................... 13
3.4 Measurement-data owner matrix.............................................................. 13
4 Lab Tests ................................................................................................................. 15
4.1 Goals .............................................................................................................. 15
4.2 Recruitment ................................................................................................... 15
4.3 Informed consent forms .............................................................................. 15
4.4 Research questions ...................................................................................... 16
4.5 Study Design ................................................................................................. 16
4.6 Time schedule ............................................................................................... 17
5 Field Tests: general info ........................................................................................ 19
5.1 Informed consent forms .............................................................................. 19
5.2 Test setting at work ....................................................................................... 19
5.2.1 Common to all use cases ....................................................................... 20
5.2.2 Use case: Office ....................................................................................... 20
5.2.3 Use case: Driving ...................................................................................... 20
5.2.4 Use case: Manufacturing ........................................................................ 21
D2.5 Field Test Strategy
6
5.2.5 Augmented Reality at the Manufacturing use case .......................... 21
5.3 Test setting at home ..................................................................................... 21
5.4 Privacy, anonymisation, and obfuscation ................................................ 22
5.4.1 Privacy ....................................................................................................... 22
5.4.2 Anonymisation, and obfuscation .......................................................... 22
5.5 Logistics and management........................................................................ 22
5.6 Time schedule ............................................................................................... 24
6 Field Tests: short-term study ................................................................................. 27
6.1 Goals .............................................................................................................. 27
6.2 Recruitment ................................................................................................... 27
6.3 Time schedule ............................................................................................... 27
6.4 Protocols ........................................................................................................ 28
7 Field Tests: Long-term study ................................................................................. 29
7.1 Goals .............................................................................................................. 29
7.1.1 Use case: Office ....................................................................................... 29
7.1.2 Use case: Driving ...................................................................................... 29
7.1.3 Use case: Manufacturing ........................................................................ 29
7.2 Testing approach ......................................................................................... 29
7.3 Recruitment ................................................................................................... 31
7.4 Time schedule ............................................................................................... 31
7.5 Protocol ......................................................................................................... 34
D2.5 Field Test Strategy
7
1 Introduction
1.1 Goals of the test strategy
The goal of the test strategy is providing a guideline for conducting the testing
activity about all the modules of the WA Tool.
In particular, the testing strategy is divided into three parts:
Lab Tests
Short-term Field Tests
Long-term Field Test
For each of them, this deliverable provides an initial set of guidelines, and
specifies the differences about the three case studies where the WA Tool will be
deployed.
Such case studies are:
Office
Driving
Manufacturing
1.2 Main interrelationships with other
deliverables
The present deliverable is strictly bound to D2.6 “Intervention Protocol”. It
receives inputs and/or provides guidelines and requirements to:
WP3, in particular for the definition of measures to be adopted at the
different pilot sites;
WP4 where Lab Tests will be performed;
WP7, dealing with ethical, security and privacy issues;
WP9, dealing with the final Pilot Operational Manual.
D2.5 version1 will be refined at M16 with the inputs from WP3 and WP4.
It will provide basic input to WP9 for the definition of the “Pilot Operational
Manual” (D9.1) that will describe in detail the pilot application, the documents
to be used during the pilot tests (e.g., informal consent, questionnaires, etc.),
user recruitment rules at each pilot site, performance metrics and their
assessment procedures.
Figure 1 depicts the interrelationships between D2.5 and other WPs.
Finally, D2.2 “Analysis of Available and Suitable Sensors” provides an in-deep
analysis of the technologies adopted by sensors we are going to deploy, while
D2.3 “Data Management Plan” provides details on the pseudo-anonymisation
and encryption.
D2.5 Field Test Strategy
8
Figure 1 - Interrelationships between D2.5 and other deliverables/work packages
D2.5 Field Test Strategy
9
2 Use cases The WA project will be tested in three reference use cases:
Office
Driving
Manufacturing
Such use cases provide a good overview of the typical workplaces, and thus
we believe they will permit to effectively testing the WA Tool.
2.1 Office
The Office use case will be at Mutua Universal in Barcelona. The Piraeus Bank
offices mentioned in the DoA turned out to hardly have personnel over 50 years
old.
The description of the office workplaces at Mutua Universal will be included in
the M16 version of this deliverable.
2.2 Driving
The company IPLUSF, that had signed a Letter of Interest to provide the Driving
use cases, unfortunately withdrew from the project.
The consortium entered in contact with an Athens-based ambulance
company, First Aid Ambulance, who showed great interest in the project.
Unfortunately, as occurred with the bank offices, they do not have sufficient
drivers aged over 50 to be able to host the field tests.
The consortium actively searched for alternatives and we are currently very
close to an agreement with a large company. Again, more information in the
M16 version of this document.
Since general workflows at both office and driving workplaces are quite well
defined, delays caused by these company switches are reduced.
2.3 Manufacturing
The factory, belonging to the GA RyA, which agreed to be the WA project pilot
site for the “Manufacturing” use case, is in Valladolid (Spain). The factory
manufactures car interior components such as dashboards, doors, etc.
2.3.1 Workplace
The factory is organised into several “islands” where specific jobs are performed
by workers with the aid of specific tools1. Many different jobs are performed,
such as assembling a dashboard, for example, which require several steps.
1 As this is a public deliverable, no images of the specific workplace are included due
to confidentiality requirements from GA RyA.
D2.5 Field Test Strategy
10
After observing the whole process we selected some feasible “island” that
allowed, more easily than others, to place sensors without interfering too much
with the worker’s activity:
Leather inspection G-Class
Stitching G-Class
Assembly G-Class
Assembly XFB
Welding XFB
In such settings, workers do not move a lot (which is good for sensors based on
cameras) and the environment should not harm the electronics (it’s not too
humid or hot). Unfortunately, the whole factory is way too noisy for any kind of
microphone (even noise-cancelling ones) to be employed to the porpoise of
the WA Tool, so voice recording is not feasible.
Temperature and humidity could make the environment not conformable for
workers, especially for some of the selected “islands”, while light seems good
(intensity and position).
2.3.2 Supporting infrastructure
Electric power supply is available.
A room could be used to deploy WA servers.
2.3.3 Possible issues affecting sensor deployment
In some areas of the factory, heat and humidity could harm sensors; thus, we
selected the “islands” to avoid such risks.
2.3.4 Privacy concerns
Since audio recordings are not to be collected, privacy concerns are only
related to the public disclosure of images containing sensitive information for
the factory.
2.3.5 Health and stress problems
The major health and stress problems at PDS are due to the posture, and to the
stress of keeping up to the pace required by the job procedure.
2.3.6 Human Machine Interaction
The job could be dangerous (some machines require high attention to avoid
injuries), so an interactive HCI could feasible but requires special care.
In other words, the WA Tool should be able to provide suggestions and
recommendations anytime the worker needs it (for example, when she/he is
getting overloaded, or overstressed); thus, the WA Tool should be able to
interrupt the worker also during the working activities, but considering her/his
status (for example, when the worker is taking a break).
The HCI could be based on a smartphone where a simple touch-based UI will
be present. Interaction could be both WA tool-initiated (sort of “interrupt”) and
worker-initiated (a “request”).
2.3.7 Notes and issues
None.
D2.5 Field Test Strategy
11
3 Measurements and sensors The WA system is going to deploy several sensors and gather their
measurements. The following sections provide an overview of what is going to
be collected by the WA system.
Note that we have two measurement typologies:
Utilising electronic sensors
By means of questionnaires
These two different methodologies to gather data about workers permit to
compare the behaviour of the WA tool, assessing its perceived effect on
workers’ life.
3.1 Sensors
The following body sensors will be analysed in Lab Tests and, if proving useful,
deployed in Field Tests; this list was decided after the technical survey provided
in D2.2.
ELECTROCARDIOGRAM (ECG)
ELECTROENCEPHALOGRAM (EEG)
ELECTROMYOGRAPHY (EMG)
GALVANIC SKIN RESPONSE (GSR) / SKIN TEMPERATURE (ST)2
FACIAL EXPRESSION
VOICE ANALYSIS: BLUETOOTH HEADSET WITH NOISE-CANCELLING MICROPHONE
ELECTROOCULOGRAM (EOG)
EYE MOVEMENT AND PUPIL DIAMETER
BODY POSTURE
GESTURE RECOGNITION
Additionally, the following environmental sensors will be tested:
NOISE
THERMO-HYGROMETRIC
LIGHT
USER LOCATION
Correlations between strain and environmental conditions will be analysed
according to the best practices described in the current public academic
literature.
As already derived in D2.1, the strain types mentioned in Table 1 can be
investigated by means of the physiological parameters monitored by the
sensors.
2 GSR and ST are alternative; to be decided during Lab Tests.
D2.5 Field Test Strategy
12
Table 1. Measurement of mental, emotional and physical strain with the Sensors
used in the WorkingAge project3
X:direct indicator
O:indirect indicator Ele
ctr
oc
ard
iog
ram
(EC
G)
Ele
ctr
om
yo
gra
ph
y (
EM
G)
Ga
lva
nic
skin
re
spo
nse
(G
SR
)(*)
Ele
ctr
oe
nc
ep
ha
log
ram
(EEG
)
Ele
ctr
oo
cu
log
ram
(EO
G)
Fa
cia
l e
xpre
ssio
n
Vo
ice
An
aly
sis
Ge
stu
re r
ec
og
nitio
n
Pu
pil
dia
me
ter
(PD
)
Bo
dy P
ost
ure
Eye
-Mo
ve
me
nt
psychological strain
o
mental o o x o o
sensory
o o o
discriminatory
o o o
combinatory
o x o
emotional o x o o o o o
physical strain
o
muscular o x o
dynamic x x o
static o x o o
cardiovascular x o o
skeletal o o o
(*) ST will be tested as an alternative sensor, for collecting comparable
indicators.
3.2 Questionnaires
We also plan to collect information on workers by administering specific
questionnaires to collect different type of data such as:
Demographic data, e.g. age, gender
Health status
Cognitive and emotional situation
Home time (nutrition, exercise, sleep, etc.)
Such questionnaires could be paper based or electronic. In particular, during
Lab Tests questionnaires will be probably paper based, while during Field Tests,
an electronic format will be adopted (the device used as system HCI will be
used to administer and manage questionnaires).
3 Adapted and extended following Kirchner (1986).
D2.5 Field Test Strategy
13
3.3 Measurement-use case matrix
Because each use cases has its specific constraints WA is not going to deploy
exactly the same set of sensors for the three use cases. Table 2 shows what WA
is going to deploy, for each use case, at the workplace and at home.
Note that the questionnaire (described in Section 3.2) is considered here as a
measurement since they will also gather data related to each WA user.
Table 2 - Measurements and use cases (W=workplace, H=home)
Use cases
Office Driving Manufacturing
Me
asu
rem
en
ts
ECG W W W EEG W W w EMG (*) - - - GSR / ST (**) W W W Facial expression W W W Microphone W W - EOG W W W Eye movement, Pupil diameter W W W Body posture W W W Gesture recognition W W W Noise W W W Thermo-hygrometric W W W Light W W W User location W W W Smartband H H H Questionnaires W/H W/H W/H
(*) EMG sensor will be only deployed in Lab Tests, to derive theoretical
considerations/concepts for the future integration of the EMG system
into the WA Tool
(**) GSR and ST are alternative; Lab Tests will help in selecting the best one,
in terms of comfort, signal stability, etc.
3.4 Measurement-data owner matrix
Each measurement is under the responsibility of one or more WA partners. Table
3 shows the partner(s) responsible for each measured data.
D2.5 Field Test Strategy
14
Table 3 - Measurements and data owners
WA partners
POLIMI AUD ITCL BS INTRAS RWTH UCAM TPZ EXUS
Me
asu
rem
en
ts
ECG X EEG X EMG X GSR / ST X Facial
expression X
Microphone X X EOG X Eye
movement,
Pupil
diameter
X
Body
posture X
Gesture
recognition X
Noise X Thermo-
hygrometric X
Light X User
location X
Smartband X Question-
naires X
D2.5 Field Test Strategy
15
4 Lab Tests Lab Tests will assess system acceptability, usability, and validity. Each partner
designs its own tests and manages its own documents, hardware, software, and
data.
4.1 Goals
Lab Tests represent the proof-of-concept of the WA sensor components. At this
stage, the sensor components for the final WA system will be selected, with
respect to the study findings.
The main goal is to assess acceptability, usability, and validity. Each partner
should indicate here its own specific:
- goals
- statistical methods
- evaluation methodologies
- Safety assessment
At the end of Lab Tests, acceptable, usable, and valid sensors will be
considered for Field Tests.
An integration study, testing the whole WA system will be conducted at the end
of Lab tests. During the integration study, we aim at finding:
- Possible interferences with other partners’ sensors
- Acceptability, usability, and validity of the whole WA system
4.2 Recruitment
The WA project aims at a well-defined category of workers; the main
requirements are:
Possibly, age 50+
Possibly, gender-balanced groups
No disabilities
The Lab Tests, however, are a part of the development cycle, and the goal is to
test the reliability of the subsystems that will compose the WA Tool. Therefore,
the requirements mentioned above are not strictly enforced. In other words,
each WA partner is free to select the more convenient set of users for testing its
equipment.
4.3 Informed consent forms
It is under the responsibility of each WA partner to collect the informed consent
forms from the users involved in Lab Tests. The information consent will involve
the permission to:
Measure
Elaborate
Store
the data collected by the WA partner.
D2.5 Field Test Strategy
16
The informed consent form will specify how long data will be retained by the
WA partner, and how users can request to delete their own data.
Other WA partners won’t be allowed to access the data, unless the informed
consent form explicitly mentions that possibility, and the user explicitly allows it.
The consent form in paper format will be stored in the correspondent country in
which they are generated.
4.4 Research questions
The research questions to be answered for Lab Tests refer to sensors and the
appropriateness of physiological indicators as measure of mental, emotional,
and physical strain itself.
About sensors, the objective is to prove that each of them is valid, feasible, and
usable to measure its foreseen type of strain in the suggested way.
The general research questions for each individual sensor should be:
1. Validate the measurement parameter/sensor (Does the sensor measure
what we want to measure?)
2. Evaluate differences between younger and older participants.
Then, Lab Tests will aim at confirming the appropriateness of the chosen
physiological indicators to gauge the mental, emotional and physical strain of
workers.
In order to determine the specific research questions, a literature survey for the
above-mentioned sensors is necessary. This has to consider especially the
following points:
1. Nature of the stimulus:
a. Cognitive or informational task influence data collected by eye
tracking, GSR, heartbeat measurement devices (sensors)
b. emotional stimuli influence data collected by face and voice
recognition devices;
c. physical stimuli influence data collected by EMG as well as the
body posture analyses;
2. Age-related differences of the measured parameters.
4.5 Study Design
The actual planning of the study depends on the kind of strain considered. As
Figure 2 shows, uniform questionnaires are used in all settings to collect
demographic and other study dependant data, e.g. control variables. To
validate the sensors, the stimulus is changed in randomized cycles if it is
required so by study design.
Closing questionnaires ask for parameters related to the measurement and the
hypothetical system of the WA tool, such as acceptance and usability. The
following list details the study workflow.
1. Questionnaires concerning participants are employed:
D2.5 Field Test Strategy
17
To collect data about age, gender and health restrictions of the
participants.
2. Measurement:
alternating strain
approximately about 3 to 5 different strain levels (two levels, for
mental stress)
several cycles
3. Test Persons:
Female and male test persons equally distributed
Two age groups: younger than and older than 50 years,
dependant on the requirements
The number of participants is related to the study design. For the
pre-studies only a small number of participants is enough to
recognize tendencies.
4. Additional subjective strain measurements, e.g.:
NASA-TLX: To evaluate the relationship of measured and
subjective strain
Rating Scale of Mental Effort (RSME)
5. Statistical Analyses, for instance:
Within-subject-design
Mean value difference and two-way repeated measure of
variance analysis tests for:
i. Significance of mean value difference between no strain
and strain condition, and between different strain levels
ii. Significance of mean value difference between 50+ and
younger test subjects (as stated above, the requirement
about subject recruitment are not strictly enforced, as the
goal is to find the more convenient set of users for testing
the equipment),
Correlations between the evaluated physiological and subjective
parameters
6. Closing questionnaires
Figure 2 - Study design example
4.6 Time schedule
The time schedule for the Lab Tests can be derived from the Gantt chart in the
DoA. A recommendation for testing and data analysing periods is addressed in
the following sections. The final time schedule will be discussed in consultation
with the work package leader of WP2 (RWTH).
D2.5 Field Test Strategy
18
According to the DoA the Lab Tests will be prepared between Month 11 and
Month 16 (Dec 19 – May 20). During this period the study design will be
developed and pre-tests for the preparation of the pilot tests will be completed
including the analysis of the results. During this phase, we will involve a total of
30 subjects.
The conduction of further Lab Tests will take place between Month 20 and
Month 25. Results of these tests will be used as basis for the pilots. During this
phase, we will involve a total of 60 subjects (30 being the same involved in
phase I).
The following time schedule can be estimated:
M11 (Dec 2019): Planning of the Lab Tests & developing study design
M16 (May 2020): End of Lab Tests phase I
M20 (Sept 2020): Planning of further Lab Tests phase II
M25 (Feb 2021): End of Lab Tests phase II
D2.5 Field Test Strategy
19
5 Field Tests: general info Field Tests aim at assessing the system reliability and effectiveness. There are
two Field Test typologies: short-time Field Tests and long-time Field Tests. In the
following, the characteristics that these typologies share will be introduced.
Note that the protocol for system evaluation is given in D2.6. D2.5 provides
guidelines for preparing and conducting the tests.
5.1 Informed consent forms
Each WA partner will adopt a standardised set of informed consent forms:
Involving users
o Document addressing a specific kind of data collected by a
specific WA partner. Such documents will involve the permission
to:
Measure
Elaborate
Store
data collected by the WA partner. Each user involved into Field
Tests must agree with these documents.
o Document addressing elaboration of data by means of a
centralised Agent, in change of providing users with personalised
advice. Each user involved in Field Tests must agree with these
documents.
Involving the company
o Document addressing a specific kind of data collected by a
specific WA partner. Such documents will involve the permission
to:
Measure
Elaborate
Store
data collected by the WA partner. Each company involved into
Field Tests must agree with these documents.
o Document addressing elaboration of data by means of a
centralised Agent, in change of providing users with personalised
advices. Each company involved into Field Tests must agree with
these documents.
The consent forms in paper format will be stored in the correspondent country
in which they are generated (D2.3).
5.2 Test setting at work
Each sensor needs to meet a set of requirements, due to the very nature of the
sensor and its technological limitations. Therefore, the test settings should be
selected in a way that permits to all sensors to operate correctly.
Notice that sensors listed in Table 4 will be actually deployed after the study in
Lab Tests shows they are acceptable, usable, and valid.
D2.5 Field Test Strategy
20
5.2.1 Common to all use cases
In general, a static setting is preferable, as cameras and eye trackers cannot
“follow” users while they move. Moreover, a “reasonably quiet” environment is
needed for the microphones to collect usable vocal samples4. Finally, the
environment should not harm the sensors (no excessive humidity, dust, etc.)
The company hosting the use-case should provide a room for deploying some
WA servers (see Section 5.3). The company should also provide places where
the WA “smart Hot Spots”, provided by GC, will be deployed.
5.2.2 Use case: Office
Sensors will be deployed on the worker’s body and on the desk. The sensors we
plan to deploy are:
ECG, EEG, GSR, EOG, microphone, user location: on the worker’s body
Facial expression: a camera standing in front of the worker, on the
monitor or on the desk
Eye movement, pupil diameter: a device put on the desk, in front of the
worker (a device mounted on glasses is under evaluation, too)
Body posture: a camera put sideways, e.g. on a wall, observing the
whole worker’s body
Gesture recognition: a device put in front of the worker
Noise, thermo-hygrometric, illumination: sensors put in the office, not far
from the worker
Questionnaires: provided by means of the system HCI
The environments should not pose risks about sensor integrity.
5.2.3 Use case: Driving
Sensors will be deployed on the driver’s body and on the dashboard. The
sensors we plan to deploy are:
ECG, EEG, GSR, EOG, microphone: on the worker’s body5
Facial expression: a camera put in front of the worker, on the dashboard
Eye movement, pupil diameter: a device put in front of the worker, on
the dashboard (a device mounted on glasses is under evaluation, too)
User location: a sensor put inside the vehicle
Body posture: a camera put in the cockpit, observing the worker’s body
Gesture recognition: a device put in front of the worker
Noise, thermo-hygrometric, illumination: sensors put in the vehicle
Questionnaires: provided by means of the system HCI
The environments could pose some risks about sensor integrity, as some of them
will be installed under direct sunlight (the dashboard). During the field short-
term study, these aspects will be investigated.
4 We are going to use noise-cancelling microphones, but this technology has limitations. 5 We will check regulations about traffic, in the Country where the Field Test will be
conducted, to ensure sensors can be deployed on a driver’s body. Moreover, during
the Lab Test, we study how to ensure appropriate safety conditions.
D2.5 Field Test Strategy
21
5.2.4 Use case: Manufacturing
Sensors will be deployed on the worker’s body and on the machine operated
by her/him. The sensors we plan to deploy are:
ECG, EEG, GSR, EOG, user location: on the worker’s body
Facial expression: a camera put in front of the worker, on the desk or
machine she/he is operating
Eye movement, pupil diameter: a device put in front of the worker, on
the desk or machine she/he is operating (a device mounted on glasses is
under evaluation, too)
Body posture: a camera put on a wall, observing the whole worker body
Gesture recognition: a device put in front of the worker
Noise, thermo-hygrometric, illumination: sensors put in the workplace, not
far from worker
Questionnaires: provided by means of the system HCI
Sensors we currently do not plan to deploy:
Microphone: the environment is too noisy even for a noise-cancelling
microphone
The environments could pose some risks about sensor integrity, in particular dust,
temperature, and humidity could harm some sensors.
5.2.5 Augmented Reality at the Manufacturing use case
The AR specific application will be a module of the WA Tool. This module will
only be available for the workers in specific workplace(s) selected for it. Due to
the fact that this type of application has to be specifically developed for each
case, the objective of this part of the study is to understand whether this type of
content can contribute to the reduction of the mental overload or the worker.
After the visit to the facilities of GA RyA, the preselected workplace for the AR
experiences is the Kitting XFB: the worker needs to pick up several car door
parts, in a 10-12 meter corridor containing several stacks. She/he has to pick up
the material needed for other workplaces and, in some cases, mount some
parts together, add tags, etc.
In this workplace we envision to provide guidance to the worker exploiting
several devices: Hololens, Magic Leap, etc.
A new visit to the plant and a more detailed work description will be made to
gather all the information needed and start with the development.
5.3 Test setting at home
About data collection at home, we plan to collect information by means of:
Administering specific “home time” questionnaires, as specified in
Section 3.2, using a mobile device
Providing a smartband (e.g. the inexpensive Xiaomi MiBand) to obtain
indication of sleeping habits and exercise measuring sleep and steps.
Data stored into the band will be periodically uploaded into the WA Tool
system
D2.5 Field Test Strategy
22
5.4 Privacy, anonymisation, and obfuscation
Privacy, anonymisation, and obfuscation are very important for the WA project.
In the following sections, we explain how we plan to tackle these matters
(technical detains can be found in D2.3).
5.4.1 Privacy
The WA project must be GDPR compliant. For that reason, privacy was one of
the most important goals, from the very beginning of the project.
To be GDPR compliant, we plan to:
Manage collected data so that only the WA partner that collected
them can access them (encryption)
Pseudo-anonymise users by means of an ID, and all recorded data refer
to that ID. For each use case, a WA member (the WA use case
manager) will maintain an encrypted file containing the mapping
worker’s name/ID; only that person will be able to access the content of
the encrypted file.
Prepare, administer, collect, and store informed consent forms
5.4.2 Anonymisation, and obfuscation
Companies, in general, do not allow us to elaborate sensible information about
their workers, outside their facilities. This issue particularly affects voice recording
and cameras. From those sensors, in fact, it is possible to recognise the person
under measurement. Moreover, voice recordings could contain very private
information (person names, bank accounts, etc.)
To cope with that serious issue, we envision a public-key based encryption
workflow where data collected by sensors are sent to a centralized cloud
storage, managed by BS, where only the WA partners in charge of further
elaborating them can access them.
This way, each WA partner is allowed to access only the data legitimately
intended for him by the data collector. The data pseudo-anonymisation will be
performed on an in-company deployed server. Each data type will be
anonymised and obfuscated in a specific way. A detailed description of such
methodologies will be provided in the second release of the present
Deliverable. Such scheme works by means of a locally deployed server, which is
in charge of applying anonymisation and obfuscation to sensible data (for
example, voice and video recordings). Each data typology will be anonymised
and obfuscated in different ways. A detailed description of such
methodologies will be provided in the second release of the present
Deliverable.
5.5 Logistics and management
This aspect is about how to install and remove hardware and software:
Bringing hardware on the field
o Servers
o Sensors
D2.5 Field Test Strategy
23
o Telecom infrastructure
o Device for HCI (a tablet or a smartphone), if workers can’t use
their own devices
Installing hardware and software, and testing everything
Removing hardware once the test is over
Moreover, during the test:
Helping workers to equip with the sensors
Moving sensors to a workplace to another (for example, from a desk to
another one)
Managing the WA servers
o Deployed at the company facility
o Deployed at WA partners
Managing the WA Tool modules
o Installed at the company facility
o Installed at WA partners
We envision the following roles: a company manager, one or more WA field
managers, several WA home managers, and a WA use case manager.
The WA field managers:
Bring all the WA servers, WA telecom infrastructures, and devices for HCI
to the company facility
Administer the informed consent forms, for workers and company
Install the WA Tool modules on the servers
Bring the sensors, install the environmental sensors,
Instructs the company manager about how to deploy wearable sensors,
and about the WA HCI
Test everything
At the end of the test, remove everything.
The company manager:
Helps workers to equip with the sensors
Instruct workers about how to use the WA HCI
Moves the environmental sensors, when needed
Fixes small hardware/software issues
The WA home managers (one for each WA partner deploying sensors or
telecom infrastructure, or running the Agent):
Monitor software installed in their home server (if any), and fix errors
Monitor the WA telecom infrastructure
The WA use case manager (incorporating the role of local Data Manger):
Monitors that the whole WA Tool works well
Maintains an encrypted file containing the mapping worker’s name/ID,
for that use case
The WA field managers are not required to stay at the company facility during
the long-term Field Tests, whereas during the short-term Field Tests they (the
whole group of WA field managers or part of it) will supervise the test.
Finally, note that the same person could play several roles.
D2.5 Field Test Strategy
24
5.6 Time schedule
The time schedule for Field Tests can be derived from the Gantt chart in the
DoA (see Figure 3) and the methodology for involving companies described in
Figure 4.
Figure 3 - WA general Gantt
Details about the schedule of the testing activities (including Lab Tests) can be
found in Figure 5.
Notice that field studies on the three use cases will be conducted in parallel.
SE
P-2
104
87
208
S
mar
t W
ork
ing e
nv
iro
nm
ents
fo
r al
l A
ges
[826
23
2]
[Wo
rkin
gA
ge]
P
art
B
59
3.
IMP
LE
ME
NT
AT
ION
3.1
W
OR
K P
LA
N —
WO
RK
PA
CK
AG
ES
, D
EL
IVE
RA
BL
ES
The
WA
pro
ject
(3
6M
) as
a
com
ple
x
rese
arch
re
quir
es
care
ful
pla
nnin
g,
man
agem
ent
and
adm
inis
trat
ion
in
its
dev
elopm
ent
and
imple
men
tati
on.
WP
1
wil
l in
clude
all
man
agem
ent
issu
es.
Pri
or
to th
e m
ain re
sear
ch
cycl
e,
the
conso
rtiu
m w
ill
par
tici
pat
e in
WP
2:
This
WP
inte
nds
to s
et
the
bas
es f
or
all
ult
erio
r re
sear
ch i
ncl
udin
g t
he
sele
ctio
n o
f
tools
an
d
par
tici
pan
ts,
opti
miz
ing
the
tim
e fo
r th
e te
sts.
Res
earc
h c
ycl
es (
Expec
ted a
t le
ast
two)
wil
l st
art
wit
h t
he
def
init
ion o
f th
e in
terv
enti
ons
for
the
use
rs (
WP
3)
at w
ork
and i
n d
aily
lif
e pro
cess
. T
his
work
wil
l be
the
bas
e fo
r W
P4
(HC
I pla
tform
), W
P5 (
IOT
infr
astr
uct
ure
and s
ervic
es),
WP
6 (
Dat
a A
nal
ysi
s) w
hic
h o
ccur
in p
aral
lel
(and s
har
ing
pro
duce
d k
now
ledge)
. E
thic
s an
d s
ecuri
ty d
om
ain p
erfo
rmed
duri
ng W
P7
. D
eplo
ym
ent
and I
nte
gra
tion (
WP
8)
in
whic
h t
he
final
pro
toty
pes
and t
he
opti
miz
atio
n o
f th
e re
sear
ch c
ycl
es w
ill
be
adap
ted f
or
the
test
s. T
he
inte
rven
tion
model
s an
d
mea
sure
men
t pro
toty
pes
w
ill
be
then
in
tegra
ted
and
use
d
to
coll
ect
dat
a in
th
e W
P9
(T
est
Per
form
ance
). I
n W
P10 S
tandar
diz
atio
n a
nd b
usi
nes
s dev
elopm
ent,
com
mer
cial
izat
ion a
nd I
PR
man
agem
ent
wil
l
be
work
ed c
onsi
der
ing t
he
futu
re t
he
mar
ket
rel
ease
of
the
studie
d s
olu
tion
. T
his
WP
wil
l al
so s
um
mar
ize
all
the
acti
ons
pro
pose
d f
or
dis
sem
inat
ion i
n t
he
pla
n st
ated
in p
oin
t 2.4
duri
ng t
he
pro
ject
lif
etim
e.
Work
has
bee
n
stru
cture
d i
n t
en w
ork
pac
kag
es:
six c
over
ing R
esea
rch a
nd I
nnov
ati
on w
ork
, T
wo f
or
test
(In
tegra
tion a
nd U
ser
test
s), O
ne
for
exp
loit
atio
n a
nd d
isse
min
atio
n O
ne
for
the
defi
nit
ion o
f al
l sp
eci
fica
tions
and o
ne
man
agem
ent
work
pac
kag
e.
TA
BL
E 3
.1D
: G
AN
TT
CH
AR
T
12
34
56
78
91
01
11
21
31
41
51
61
71
81
92
02
12
22
32
42
52
62
72
82
93
03
13
23
33
43
53
6
WP
1P
RO
JEC
T M
AN
AG
EMEN
T15
1516
1621
32
30
31
29
26
28
28
32
39
32
32
34
31
30
30
29
22
1413
109
910
1010
1011
1111
1110
T1.1
PR
OJE
CT
CO
OR
DIN
ATI
ON
AN
D M
AN
AG
EMEN
T1
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
1
T1.2
QU
ALI
TY A
ND
RIS
K M
AN
AG
EMEN
T 1
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
1
T1.3
SCIE
NTI
FIC
MA
NA
GEM
ENT
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
T1.4
CO
NFI
GU
RA
TIO
N O
F TH
E SU
PP
OR
TIN
G B
OD
IES
11
11
11
11
WP
2U
SER
CEN
TRIC
DES
IGN
4
44
44
42
21
11
15
54
40
00
00
00
00
00
00
00
00
00
0
T2.1
USE
R R
EQU
IREM
ENT
DEF
INIT
ION
11
11
22
22
11
T2.2
HC
I IN
TER
FAC
E A
ND
CO
NC
EPT
DES
IGN
11
11
11
11
11
T2.3
DA
TA M
AN
AG
EMEN
T C
OLL
ECTI
ON
AN
D O
PER
ATI
ON
DEF
INIT
ION
11
11
11
11
11
T2.4
FUN
CTI
ON
AL
SPEC
IFIC
ATI
ON
AN
D A
RC
HIT
ECTU
RE
11
11
11
11
11
11
T2.5
DEF
INIT
ION
OF
PIL
OT
TEST
11
11
11
11
WP
3IN
TER
VEN
TIO
N D
ESIG
N0
00
04
44
44
10
00
44
44
10
00
00
00
00
00
00
00
00
0
T3.1
HEA
LTH
SU
RV
EILL
AN
CE
ASS
ESSM
ENT
AN
D M
ETR
IC D
EFIN
ITIO
N1
11
11
11
11
T3.2
CO
GN
ITIV
E/EM
OTI
ON
AL
ASS
ESSM
ENT
AN
D M
ETR
IC D
EFIN
ITIO
N1
11
11
11
11
T3.3
USE
R B
EHA
VIO
UR
AL
ASS
ESSM
ENT
11
11
11
11
1
T3.4
INTE
RV
ENTI
ON
S D
EFIN
ITIO
N1
11
11
11
11
11
WP
4H
CI
PLA
TFO
RM
00
00
07
77
77
77
77
11
77
77
70
00
00
00
00
00
00
00
T4.1
VIS
UA
LIZA
TIO
N P
LATF
OR
M F
RO
NT
END
IM
PLE
MEN
TATI
ON
11
11
11
11
12
21
11
11
T4.2
GES
TUR
E-B
ASE
D I
NTE
RA
CTI
ON
PLA
TFO
RM
IM
PLE
MEN
TATI
ON
11
11
11
11
11
11
11
T4.3
VO
ICE-
BA
SED
IN
TER
AC
TIO
N P
LATF
OR
M I
MP
LEM
ENTA
TIO
N1
11
11
11
11
11
11
1
T4.4
FAC
IAL
AFF
ECT
AN
ALY
SIS
AN
D E
YE
TRA
CK
ING
PLA
TFO
RM
IM
PLE
MEN
TATI
ON
11
11
11
11
11
11
11
T4.5
NEU
RO
MET
RIC
S1
11
11
11
11
11
11
1
T4.6
SOC
IAL/
WO
RK
ING
/HEA
LTH
NET
WO
RK
IN
TEG
RA
TIO
N F
OR
DIS
TRIB
UTE
D D
EVIC
ES D
EVEL
OP
MEN
T1
11
11
11
11
11
11
1
T4.7
LOC
ATI
ON
FU
NC
TIO
N &
E1
12
SER
VIC
E IM
PLE
MEN
TATI
ON
11
11
11
11
11
11
11
WP
5IO
T0
00
00
00
22
22
22
22
11
12
22
20
00
00
00
00
00
00
0
T5.1
IOT
LOC
AL
DA
TA C
OLL
ECTI
ON
PLA
TFO
RM
DEV
ELO
PM
ENT
11
11
11
11
22
21
11
1
T5.2
IOT
SEN
SOR
S FU
NC
TIO
NA
LITI
ES D
EVEL
OP
MEN
T1
11
11
11
11
11
1
WP
6D
ATA
AN
ALY
SIS
00
00
02
22
22
22
22
11
22
22
22
00
00
00
00
00
00
00
T6.1
DA
TA A
GG
REG
ATI
ON
AN
D F
ILTE
RIN
G1
11
11
11
11
22
11
11
11
T6.2
AG
ENT
BA
CK
END
: M
OD
EL D
EFIN
ITIO
N A
ND
TR
AIN
ING
11
11
11
11
11
11
11
1
T6.3
AG
ENT
BA
CK
END
: R
UN
-TIM
E LE
AR
NIN
G A
ND
REA
SON
ING
11
11
11
11
11
11
11
1
WP
7ET
HIC
S A
ND
SEC
UR
ITY
22
22
22
31
11
11
11
11
11
11
11
11
11
11
11
11
11
11
T7.1
RES
PO
NSI
BLE
RES
EAR
CH
AN
D I
NN
OV
ATI
ON
11
11
11
12
2
T7.2
ETH
ICA
L, L
EGA
L A
ND
SO
CIA
L IM
PLI
CA
TIO
NS
11
11
11
11
11
11
11
11
11
11
11
T7.3
SEC
UR
ITY
AN
D P
RIV
AC
Y M
OD
ELIN
G1
11
11
11
11
11
11
WP
8D
EPLO
YM
ENT
AN
D I
NTE
GR
ATI
ON
00
00
00
00
00
11
13
33
33
33
33
21
00
00
00
00
00
00
T8.1
LOC
AL
ASS
ISTI
VE
SYST
EM I
NTE
GR
ATI
ON
(SE
NSO
R, W
EAR
AB
LES,
MO
BIL
E D
EVIC
E)1
11
11
11
11
T8.2
DIS
TRIB
UTE
D S
YST
EM I
NTE
GR
ATI
ON
(LO
CA
L +
CLO
UD
-BA
SED
SER
VIC
ES)
11
11
11
11
11
T8.3
SYST
EM T
ESTI
NG
AN
D V
ALI
DA
TIO
N1
11
11
11
11
11
11
1
WP
9P
ILO
TS0
00
00
00
00
01
11
13
43
33
32
22
22
11
22
22
22
22
1
T9.1
TRA
ININ
G A
ND
IC
T SU
PP
OR
T1
11
11
T9.2
IN L
AB
TES
TS1
11
11
11
11
11
1
T9.3
IN C
OM
PA
NY
TES
T1
11
11
11
11
11
11
11
11
11
1
T9.4
TEST
CY
CLE
EV
ALU
ATI
ON
11
11
11
11
11
11
11
1
WP
10
DIS
SEM
INA
TIO
N, E
XP
LOIT
ATI
ON
STR
ATE
GY
AN
D V
ALU
E C
HA
IN M
OD
ELLI
NG
33
33
11
11
11
11
11
11
11
11
11
11
11
11
11
13
33
33
T10
.1C
OM
MU
NIC
ATI
ON
AN
D D
ISSE
MIN
ATI
ON
STR
ATE
GY
1
11
11
11
1
T10
.2C
OM
MU
NIC
ATI
ON
AN
D D
ISSE
MIN
ATI
ON
MA
TER
IAL
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
11
T10
.3EX
CH
AN
GE
WIT
H O
THER
PR
OJE
CTS
/CO
UN
TRIE
S1
11
11
11
11
11
T10
.4ST
AN
DA
RD
IZA
TIO
N1
11
12
22
22
22
22
22
22
22
22
22
22
22
22
22
11
11
1
T10
.5EX
PLO
ITA
TIO
N S
TRA
TEG
Y A
ND
VA
LUE
CH
AIN
MO
DEL
LIN
G1
11
11
11
11
T10
.6IP
R M
AN
AG
EMEN
T1
11
11
11
11
Mo
nth
Alfa
Beta
USE
RS
M1
M2
M5
M3
M4
M6
M8
M7
M9
M 11
M10
D
D
D DD
D
D
D
D D
D
D D D D D D D
D
D D D D
D
D
D
D
D
D
DD
D D D
DD
D
D
D2.5 Field Test Strategy
25
Figure 4 - Company implication in WA
SE
P-2
104
87
20
8
Sm
art
Wo
rkin
g e
nvir
on
men
ts f
or
all
Ages
[826232
] [W
ork
ingA
ge]
P
art
B
25
more
eff
ecti
ve
than
corr
ecti
onal
appro
aches
aft
er t
ypic
al d
isea
ses
hav
e al
read
y m
anif
este
d130.
Addit
ional
ly,
this
targ
et p
opu
lati
on i
s le
ss r
esis
tant
to t
he
inse
rtio
n o
f IC
T s
olu
tions
aim
ed t
o i
mpro
ve
thei
r li
fe q
ual
ity,
ensu
ring t
hat
the
dev
eloped
pro
duct
wil
l be
tail
ore
d a
ccord
ingly
wit
h t
hei
r pre
fere
nce
s as
they
age.
It s
hould
be
note
d t
hat
thes
e use
rs a
re P
rim
ary
Use
rs. M
ore
over
, W
A w
ill
involv
e Sec
ondary
Use
rs, nam
ely h
um
an
reso
urc
es p
erso
nnel
of
the
org
aniz
atio
ns,
occ
upat
ional
hea
lth a
nd s
afet
y p
rofe
ssio
nal
s. H
ealt
hca
re s
pec
iali
sts
and
Pla
tform
dev
eloper
s w
ill
be
Ter
tiar
y U
sers
. T
he
role
of
seco
ndar
y a
nd t
erti
ary u
sers
wil
l re
gar
d m
onit
ori
ng a
nd
support
, pro
vid
ing fe
edbac
k as
ex
tern
al use
rs an
d,
more
im
port
antl
y,
ensu
ring th
e co
rrec
tnes
s of
the
on
-lin
e
advis
ing a
nd i
nte
rven
tion s
yst
ems.
Tab
le 4
sum
mar
izes
the
use
r ca
tegori
es a
nd p
arti
cipat
ion i
n t
he
pro
ject
: T
able
4:
Targ
et p
opu
lati
on
U
SE
R
DE
SC
RIP
TIO
N
PA
RT
ICIP
AT
ION
ST
AG
E
WP
M
ET
HO
D
FE
ED
- B
AC
K
PR
IMA
RY
:
Ad
ult
s >
50
Volu
nte
er
Hea
lth
y
Ad
ult
s (5
0+
) li
vin
g
ind
epen
den
tly
Du
rin
g
the
An
alysi
s,
Des
ign
,
Imp
lem
enta
tion
Ph
ase,
Tes
tin
g
WP
2
WP
3
WP
9
WA
, te
sts
Inte
rvie
ws,
wo
rksh
op
s,
Qu
esti
onn
aire
,
Inte
rvie
ws,
focu
s g
rou
ps
SE
CO
ND
AR
Y:
HH
RR
, O
HS
P
rofe
ssio
nal
s ac
tin
g i
n t
he
sup
port
of
use
rs
to
pro
mote
lo
ng
last
ing
work
ing
hea
lth
y
con
dit
ion
s,
Des
ign
an
d T
esti
ng
WP
2
WP
3
WP
9
Inte
rvie
ws,
w
ork
shop
s,
Qu
esti
onn
aire
TE
RT
IAR
Y:
Hea
lth
care
Hea
lth
car
e p
rofe
ssio
nal
s fo
r th
e re
cogn
itio
n
of
dis
ease
s an
d
con
seq
uen
t to
p
rovid
e tr
eatm
ent
and
mon
itori
ng
of
the
use
rs:
Med
ical
D
oct
ors
, N
urs
es,
Pra
ctit
ion
ers,
Nu
trit
ion
ists
, P
har
mac
ist,
Psy
cholo
gis
t, e
tc.
Du
rin
g
des
ign
wit
h
rule
s fo
r
Inte
rop
erab
ilit
y
Fra
mew
ork
at
se
man
tic/
te
chn
ical
le
vel
;
imp
lem
enta
tion
wit
h i
nfo
rmat
ion o
n
use
rs’
resu
lts.
E
val
uat
ion
ex
per
t K
no
wle
dge.
WP
2
WP
3
WP
9
Inte
rvie
ws,
wo
rksh
op
s,
Qu
esti
onn
aire
,
rep
ort
s, a
nal
ysi
s of
dat
a co
llec
ted
du
rin
g t
esti
ng
TE
RT
IAR
Y:
Pla
tform
Dev
elop
ers
Soft
war
e D
evel
op
ers
from
ap
pli
cati
on
s on
Hu
man
R
esou
rces
, O
ccupat
ion
al
hea
lth
care
, H
ealt
h
Car
e,
Big
D
ata,
In
sura
nce
,
Soci
al
Net
wo
rkin
g,
Edu
cati
on
and
En
tert
ain
men
t. S
ecu
rity
of
the
syst
em w
ill
also
be
stu
die
d h
ere
Fin
al
stag
es
of
WA
, E
xp
loit
atio
n
and
Bu
sin
ess
Pla
n
WP
4,
WP
5,
WP
6
WP
7
wo
rksh
op
s,
Inte
rvie
ws,
Qu
esti
onn
aire
,
focu
s g
rou
ps
WA
wil
l ad
op
t an
ite
rati
ve
met
hodolo
gy t
o m
inim
ize
the
over
all
risk
by e
nabli
ng t
he
earl
y i
nte
gra
tion p
repar
atio
n,
the
regula
r av
aila
bil
ity o
f th
e sy
stem
and t
he
adap
tati
on t
o d
iffi
cult
ies
or
new
nee
ds.
For
that
purp
ose
, it
wil
l
imple
men
t m
icro
-cycl
es u
sing t
he
Lea
n S
tart
-Up b
uil
d-m
easu
re-l
earn
piv
oti
ng m
ethodolo
gy (
bas
ed o
n t
he
Agil
e
Met
hodolo
gy).
E
ach
iter
atio
n
involv
es a
tea
m w
ork
ing t
hro
ugh a
full
so
ftw
are
dev
elopm
ent
cycl
e
from
re
quir
emen
t an
alysi
s to
acce
pta
nce
tes
ting w
hen
a r
elea
se
is
dem
onst
rate
d
to
final
use
rs.
Ther
efore
, an
inte
rmed
iate
rel
ease
wil
l be
inte
gra
ted a
nd p
ote
nti
ally
avai
lable
to
en
d-u
sers
ev
ery
two
month
s. I
t w
ill
star
t in
em
bry
onic
form
(e
.g., pap
er pro
toty
pes
) an
d
wil
l be
gra
dual
ly
exte
nded
thro
ughou
t th
e pro
ject
dura
tion
unti
l re
achin
g
the
Fin
al
Rel
ease
stat
us.
Ter
tiary
use
rs’
invo
lvem
ent:
F
unct
ional
re
quir
emen
ts
obta
ined
from
ter
tiar
y u
sers
wil
l be
inges
ted
into
Req
uir
emen
ts B
azaa
r131 1
32 1
33
134,
a to
ol
dev
eloped
by R
WT
H
Aac
hen
Univ
ersi
ty in
the
conte
xt
of
the
RO
LE
pro
ject
. T
he
Req
uir
emen
ts B
azaa
r is
a bro
wse
r bas
ed s
oci
al so
ftw
are
pla
tform
for
Soci
al R
equir
emen
ts E
ngin
eeri
ng (
SR
E)
addre
ssin
g t
he
chal
lenge
of
a fe
edbac
k c
ycl
e bet
wee
n u
sers
and d
evel
oper
s in
a s
oci
al n
etw
ork
ing m
anner
.
Sta
keh
old
ers
from
div
erse
C
om
munit
ies
of
Pra
ctic
e (C
oP
s)
are
bro
ught
toget
her
w
ith
serv
ice
pro
vid
ers
130
Ro
lland,
Y.,
van K
an,
G.
A.,
& V
ell
as,
B.
(2008).
Phy
sical
act
ivit
y a
nd A
lzheim
er'
s d
iseas
e:
fro
m p
reventi
on t
o t
her
apeuti
c p
ers
pecti
ves.
Jo
urn
al
of
the A
mer
ican M
edic
al
Dir
ecto
rs A
ssoci
ati
on,
9(6
), 3
90
-405.
131
htt
p:/
/req
uir
em
ents
-bazaa
r.org
)
132
Kla
mm
a, R
., J
ark
e,
M.,
Hannem
ann,
A.,
& R
enze
l, D
. (2
011
). D
er B
aza
r der
Anfo
rderu
ngen -
Open I
nno
vat
ion i
n e
merg
ente
n C
om
mu
nit
ies.
Info
rmati
k-S
pek
tru
m,
34
(2),
178–
191.
Retr
ieved
fro
m h
ttp
://w
ww
.spri
ng
erl
ink.c
om
/co
nte
nt/
mk28
uq75
g87
4208r/
full
text.
pd
f
133
Law
, E
. L
.-C
., C
hat
terj
ee,
A.,
Ren
zel
, D
., &
Kla
mm
a, R
. (2
012).
T
he S
ocia
l R
eq
uir
em
ents
Engin
eeri
ng (
SR
E)
App
roach t
o D
evelo
pin
g a
Lar
ge-S
cale
Pers
onal
Lea
rnin
g
Envir
onm
ent
Infr
astr
uctu
re.
In A
. R
avensc
roft
, S
. L
indst
aedt,
C.
Delg
ado K
loos,
D.
Hernández-L
eo,
& C
. D
. K
loos
(Eds.
), L
ectu
re N
ote
s in
Co
mp
ute
r S
cie
nce:
Vo
l. 7
563.
21st
Cen
tury
Lea
rnin
g f
or
21st
Centu
ry S
kil
ls.
7th
Euro
pean C
onfe
rence
on T
echno
logy E
nhanced
Lear
nin
g,
EC
-TE
L 2
012,
Saarb
rück
en,
Germ
any,
Septe
mber
18
-21,
201
2,
Pro
cee
din
gs
(pp.
194–
207
). B
erli
n,
Heid
elb
erg
: S
pri
ng
er
Verl
ag;
Sp
ringer
134
R
enzel,
D
., B
ehre
ndt,
M
., K
lam
ma,
R
., &
Ja
rke,
M.
(2013).
R
equir
em
ents
B
azaa
r: S
ocia
l R
eq
uir
em
ents
E
ngin
eeri
ng
fo
r C
om
munit
y-D
riven In
no
vati
on.
In 21st
IE
EE
Inte
rnati
onal
Req
uir
em
ents
Engin
eeri
ng C
onfe
rence.
RE
2013 (
pp.
326–
327
).
Fig
ure
7:
Com
pan
ies
impli
cati
on
in
WA
D2
.5 F
ield
Test S
trate
gy
2
6
Fig
ure
5 –
Sc
he
du
le
D2.5 Field Test Strategy
27
6 Field Tests: short-term study Field tests aim at assessing the WA system reliability, when deployed on the
field.
6.1 Goals
The main goal is to assess the WA reliability; in particular:
Validating the acceptance of the tool by users (user-friendly, non-
intrusive, not problematic for proper work conduction, etc.)
Collect data to validate the data processing algorithms (for instance,
validate that no false alarms are raised by the tool)
Validate reliability of sensors, in particular in potential harmful
environments (for example, under direct sun light)
Validate the architecture of the system (local clouds, global clouds,
databases managements, etc.) when deployed on the field; in
particular: reliability and robustness
6.2 Recruitment
The WA consortium will need to organise a meeting at each company office to
explain the system and the goals of WA and ask for their cooperation.
The main requirements for the recruitment are:
Age 50+
No disabilities
Starting from such requirements, each company will provide a list of possible
candidates; then, the WA consortium will select the persons who will be
involved into the project.
Note that gender aspects will be considered in the data analyses. However,
because it is not straightforward to find a large enough sample of workers, it is
decided not to put gender restrictions in the recruitment requirements. As a
‘nice-to-have’ requirement, in case the luxury of choosing participants exists, a
gender-balanced group will be sought.
6.3 Time schedule
The goal of the short-term Field Tests is to assess that all the WA Tool
components are working as expected. Fields Tests are divided into:
Single Day, aiming at testing the WA components;
Week, aiming at testing the whole WA system.
The general time schedule is:
M15 (Apr 2020): Beginning of recruitment and training
M16 (Mar 2020): Beginning of activities for short-term, Single-Day Field Tests,
phase I
D2.5 Field Test Strategy
28
M20 (Sep 2020): End of activities
M23 (Nov 2020): Beginning of activities for short-term, Single-Day Field Tests,
phase II
M27 (April 2021): End of activities
M20 (Sep 2020): Beginning of activities for short-term, Week Field Tests
M24 (Jan 2021): End of activities
6.4 Protocols
Each component requires a different testing protocol; in particular:
1. Sensor hardware (e.g., the microphone)
2. Server hardware and software, for machines deployed at the company
facility (for example, machine and software needed for voice
obfuscation anonymisation, and encryption)
3. Server hardware and software, for machines deployed at some WA
partner (for example, machine and software needed to infer emotional
state, from an utterance, installed at POLIMI)
4. Network infrastructure installed at the company facility
5. Device for HCI
6. Questionnaires
7. The whole Agent loop: sensors information agent advice
Such protocols will be provided by:
For components 1, 2, and 3, each WA partner will provide a protocol to
be adopted for effectively testing everything.
For component 4, GC will provide a protocol.
For component 5, UCAM will provide a protocol.
For component 6, INTRAS will provide a protocol.
For component 7 (i.e., for testing that the whole WA Tool is working
properly), the WA consortium will agree on a protocol.
Notice that here as “protocol” we mean a document specifying the test
procedures to apply, the expected results, and how to address issues.
Small issues will be addressed by the WA field managers and the company
manager, whereas more complex problems (in particular, issues affecting
machine/software installed at WA partners) will be addressed by WA home
managers.
We aim at involving 30 workers, per use case, during short-term Field Tests. Each
WA partner will provide, as a part of its protocol, how many workers will be
measured, and how many hours of data recordings will be collected.
D2.5 Field Test Strategy
29
7 Field Tests: Long-term study Field tests aim at assessing the system effectiveness, when deployed on the
field.
7.1 Goals
The main goal is to assess system effectiveness, when deployed on the field; in
other words, understanding whether the WA Tool is able to reduce user's strain,
by advices about stress management, body postures, and environmental
control.
All the use cases will permit to highlight issues related to:
Mental stress
Physical strain
Environment conditions
However, each use case, due to its very nature, will be more suitable for one or
more of these issues.
7.1.1 Use case: Office
This use case is focused on body posture and mental stress.
Advices generated by the WA Tool could be in “real time” (i.e., proactively
provided as soon as the system discovers that they are useful) or “offline” (for
example, as a report provided at the end of the shift). The worker could also
ask the WA Tool for any “pending” advice.
7.1.2 Use case: Driving
This use case is focused on mental stress and environment conditions.
Advices generated by the WA Tool will be “offline” (for example, as a report
provided at the end of the shift), to avoid distraction. However, the worker
could ask the WA Tool for any “pending” advice.
7.1.3 Use case: Manufacturing
This use case is focused on body posture, physical strain, mental stress, and
environment conditions.
Advices generated by the WA Tool could be in “real time” (i.e., proactively
provided as soon as the system discovers that they are useful) or “offline” (for
example, as a report provided at the end of the shift). The worker could also
ask the WA Tool for any “pending” advice.
7.2 Testing approach
The long-term Field Test requires measuring 30 workers per use case, along a
period of one year. This means that we should provide enough sensors and
computational power, and ensure that the WA tool stays up and running for
one year.
The following factors must be taken into account:
D2.5 Field Test Strategy
30
Each group deploying sensors has its own constraints in terms of devices
available for the test (mainly due to user invasiveness/acceptance, e.g.
wearing an EEG head cap continuously, or the cost of the device) and
measuring periods (from few hours a day to a full day).
Any failure will be resolved as soon as possible. Additionally, instability of
sensors can be a reason to exclude them from the in-company tests
Thus, to comply with the claims stated in the DoA, we designed the following
strategy, highlighted by Figure 6.
The rationale is monitoring each worker for a few hours per day (depending on
the sensor), for several days (we envision a week). Then, after a while, the same
worker will undergo another testing period under another data collection
period, and so on until we reach the required measuring time.
This way, we ensure that we have enough data about every single worker,
even if we cannot measure all of them at the same time. Moreover, we can
manage to monitor all the workers we planned (30 for each use case) without
deploying too many sensors.
Summing up, we plan to:
Split one year into several so-called data collection periods.
Data collection periods are separated by a technical update iteration in
which the gathered data and feedback will be used to improve WA Tool
performance and stability. Note that whenever possible, of course we
will do any maintenance on the go during the tests.
Each data collection period is split up into several weeks, where a given
number of workers will be monitored.
Weeks are separated by a technical inter-week iteration
Rotation schemes will define per sensor which workplaces will use the
sensor, when and for how long; and how many sensors will be deployed
in parallel. Some examples:
o It is not necessary that an illumination sensor at an office measures
8 hours a day every day as office lighting is expected to be fairly
constant.
o EEG head caps are too uncomfortable to wear 8 hours a day.
o Microphones are not invasive (and inexpensive) and can be used
for the full working shift.
Figure 6 - Testing strategy
D2.5 Field Test Strategy
31
7.3 Recruitment
The WA consortium will need to organise a meeting at each company office to
explain the system and the goals of WA, and ask them for their cooperation.
The main requirements for the recruitment are:
Age 50+
Gender-balanced group
No disabilities
Starting from such requirements, each company will provide a list of possible
candidates; then, the WA consortium will select the persons who will be
involved into the project.
7.4 Time schedule
The general time schedule for the long-term Field Test is:
M23 (Dec 2020): Beginning of recruitment and training
M24 (Jan 2021): Beginning of activities related to long-term Field Tests
M35 (Dec 2021): End of activities
M36 (Jan 2021): Final assessment
Each WA partner provided initial estimations about (see Table 4):
How many workers, at the same time, can be measured. Limitations can
be due to:
o Sensor availability (for example due to cost)
o Software constraints about the number of concurrent data flows
How long the measurement time is. Limitations are mainly due to battery
life, recharging options/schemes necessary to guarantee operation
How long it takes to move (environmental) sensors from working place to
another one
How long it takes to wear (personal) sensors
The suggested working period before maintenance is needed (to the
sensor, to software or to some server)
How long it takes to complete maintenance
From such information, the following schedule details have been estimated:
How many users we can monitor at the same time
How long the “week” is (see Figure 6)
How long the “technical inter-week iteration” is
How long the “technical update iteration” is
How many “data collection periods” we have
Table 4 shows initial estimations about each sensor typology; note that figures
are per use case.
D2.5 Field Test Strategy
32
Table 4 – Information about sensors; initial estimation. Figures are per use case
Sensor information
Home
/
Work
#
workers
Measure
ment
time [h]
Time to
move or
wear
Working
period before
maintenance
[h]
Maintenance
Me
asu
rem
en
ts
ECG
(BS) W 3 (2) 5/6 (1) 5 min 5/6 (1)
recharge
battery
EEG
(BS) W 3 (2) 5/6 (1) 20 min 5/6 (1)
recharge
battery
GSR / ST
(BS) W 3 (2) 5/6 (1)