human factors & user awareness - connected automated driving
TRANSCRIPT
CARTRE and SCOUT are funded by the European Union Horizon 2020
Work Programme
Human Factors & User AwarenessInteractive Symposium
Vienna, 20 April 2018
Anna Schieben, German Aerospace Center (DLR)
Agenda for today
Interactive Symposium - Vienna, 20 April 2018
• Position paper on Human Factors
• Position paper on Social Acceptance and User Awareness
• Statements, voting and Expert discussion• New HMI for Automated Vehicles
• Driver distraction
• AV’s interaction with other road users
• Social Acceptance and User Awareness
• Testing/Validation procedures for AVs
Our speakers & experts on the podium
ERTRAC/CARTRE Joint Stakeholder Workshop
• Riender Happee, TU Delft
• Olivier Lenz, FIA
• Francois Fischer, ERTICO
• Evangelos Bekiaris, CERTH
• Satoshi Kitazaki, AIST
• Andreas Eustacchio, Attorney at law
• Arjan van Vliet, RDW Netherlands
Voting tool
ERTRAC/CARTRE Joint Stakeholder Workshop
• Polls• Online voting yes/no
• www.sli.do
• Event code #EUCAD2018
• Questions/Statements
• You’re invited to leave your questions, own statements and feedback online
Interactive Symposium - Vienna, 20 April 2018
Position papersRiender Happee, TU Delft
Olivier Lenz, FIA
CARTRE and SCOUT are funded by the European Union Horizon 2020
Work Programme
User and Societal Awareness, Acceptance and BehaviourInteractive Symposium
Vienna, 20 April 2018
Riender Happee, TU Delft
2035Safe, affordable transport systemComplex HF challenges
• Manual driving will be the exception
• AVs will share the road with pedestrians and cyclistswhich may communicate
• We have to design AVs, infrastructure and traffic management for many new AV to human interactions
Human Factors of Automated Driving
• 1-Automated vehicles allowing manual drivingHow can we ensure a safe evolution towards a future where manual driving is the exception?
• 2-All automation levels including driverless vehiclesHow can we ensure a safe interaction with all kinds of other road users?
• 3-Design and verificationCan we sufficiently understand the human interaction with automation to propose systematic approaches for design and verification?
CARTRE and SCOUT are funded by the European Union Horizon 2020
Work Programme
User and Societal Awareness, Acceptance and BehaviourInteractive Symposium
Vienna, 20 April 2018
Olivier Lenz, FIA
Need for User-Centric Further Research
Interactive Symposium - Vienna, 20 April 2018
Society
User
Ethics
Health
Four Horizontal Impact Areas
Interactive Symposium - Vienna, 20 April 2018
Data Management
Mobility
CARTRE and SCOUT are funded by the European Union Horizon 2020
Work Programme
Thank you!www.connectedautomateddriving.eu
Interactive Symposium - Vienna, 20 April 2018
New HMI
Statements yes/no
ERTRAC/CARTRE Joint Stakeholder Workshop
• New smart, personalized HMI concepts, designed for high automation, will increase system robustness, user acceptance and overall comfort.
• We need new, adapted and fully integrated HMI systems that provide advanced comfort during Level 3+ automation.
Interactive Symposium - Vienna, 20 April 2018
Driver DistractionEvangelos Bekiaris, CERTH
Statements yes/no
ERTRAC/CARTRE Joint Stakeholder Workshop
• Driver state monitoring is a pre-requisite for Level 3 (and lower) automated driving to ensure safe and smooth transitions.
• Driver monitoring systems should follow a hybrid approach to reach robust performance and use a combination of driver state, vehicle state, environmental context and personalised driver info.
Evangelos Bekiaris
Mechanical Engineer Ph.D.
CERTH/HIT Director
Driver State Monitoring and its role on advanced safety and comfort of Automated Driving
In general…
To err is human
Lucius Annaeus Seneca
Driver Monitoring: Facts & Challenges
• Must have for L3 and lower AD
• Robustness only through hybrid systems (driver-vehicle-environment state assessment)
• Unobtrusive, “non-contact” monitoring
• Data fusion and decision systems
• Capturing the state is important!
• Sleepiness; Inattention; Distraction; Stress; Excessive Emotion (i.e. anger, rage, anxiety); Rest; Other (i.e. faint)
• Operator monitoring for Level 4+ automated driving (especially for public transport fleets) should be researched.
Holistic approach required for all automated vehicles in all transition cases
ATTENTIVE LONG HAUL TRUCKING
ELECTRIC VEHICLE RANGE ANXIETY
DRIVER STATE-BASED SMOOTH & SAFE AUTOMATION
TRANSITIONS (CAR)
NON-REACTING DRIVER EMERGENCY MANOEUVRE
(CAR)
LONG RANGE ATTENTIVE TOURING
WITH MOTORBIKE
RIDER FAINT PASSENGER PICK UP/DROP OFF AUTOMATION FOR
BUSES
Interactive Symposium - Vienna, 20 April 2018
AV’s interaction with other road users
Satoshi Kitazaki, AIST
AIST Univ. of Tsukuba Keio Univ. DENSO Tokyoto Business Service
SIP-adus Human Factors and HMI Research Consortium
Interaction of AV with other road users
Satoshi Kitazaki, Ph.D.
PI, SIP-adus National Project on Human Factors in Automated Driving
Director, Automotive Human Factors Research Center
National Institute of Advanced Industrial Science and Technology (AIST)
Japan
AIST Univ. of Tsukuba Keio Univ. DENSO Tokyoto Business Service
SIP-adus Human Factors and HMI Research Consortium
23
➢Question;
Do we need additional external HMI for
Automated Vehicles to coordinate actions with
other road users?
➢Answer; Yes.
AIST Univ. of Tsukuba Keio Univ. DENSO Tokyoto Business Service
SIP-adus Human Factors and HMI Research Consortium
24
Social acceptance
SafetyTraffic efficiency
Sense of security
Benefit structure of external communication
AIST Univ. of Tsukuba Keio Univ. DENSO Tokyoto Business Service
SIP-adus Human Factors and HMI Research Consortium
25
Experimental method
⚫ An experiment in a closed field.
⚫ The subjects pressed the button when they made a decision to cross.
⚫ The subjects were also asked about their confidence about their decision.
Automated vehicleSubject pedestrian
Response button
AIST Univ. of Tsukuba Keio Univ. DENSO Tokyoto Business Service
SIP-adus Human Factors and HMI Research Consortium
26
• All the subjects made a decision to cross with deceleration +“I am going to stop”
• However, effects of the messages were generally small when deceleration was
large enough.
14
14
14
14
13
13
13
13
12
12
13
10
14
15
15
13
1
1
3
1
2
0% 20% 40% 60% 80% 100%
HMIなしお先にどうぞとまります
自動走行中
HMIなしお先にどうぞとまります
自動走行中
HMIなしお先にどうぞとまります
自動走行中
HMIなしお先にどうぞとまります
自動走行中
非高齢・免
許あ
り
非高齢・免
許な
し高齢者
学童
学童
高齢者
免許なし
非高齢者
免許なし
免許あり
14名
13名
13名
15名
横断できると判断 横断できないと判断
26
Results; decisions
LicensedN=14
Non- licensedN=13
Non-licensedN=13
Non-licensedN=15
Non elderly
Elderly
Children(8-10 y.o.)
Decel. onlyDecel +”After you”Decel +“I’m going to stop”Decel +”Automed driving”
Decel. onlyDecel +”After you”Decel +“I’m going to stop”Decel +”Automed driving”
Decel. onlyDecel +”After you”Decel +“I’m going to stop”Decel +”Automed driving”
Decel. onlyDecel +”After you”Decel +“I’m going to stop”Decel +”Automed driving”
■Cross ■Do not cross
Traffic efficiency
AIST Univ. of Tsukuba Keio Univ. DENSO Tokyoto Business Service
SIP-adus Human Factors and HMI Research Consortium
3
6
8
6
5
6
7
1
2
7
10
5
8
11
12
4
8
5
5
5
6
6
5
8
5
3
2
2
3
2
4
3
3
1
3
2
1
1
4
5
2
1
3
3
2
3
5
0% 20% 40% 60% 80% 100%
HMIなしお先にどうぞとまります
自動走行中
HMIなしお先にどうぞとまります
自動走行中
HMIなしお先にどうぞとまります
自動走行中
HMIなしお先にどうぞとまります
自動走行中
免許あり
免許なし
免許なし
非高齢者
非高齢者
高齢者
学童
非常に確信した 確信した それ以外
学童
高齢者
免許なし
非高齢者
免許なし
免許あり
2727
Results; confidence
LicensedN=14
Non- licensedN=13
Non-licensedN=13
Non-licensedN=15
Non elderly
Elderly
Children(8-10 y.o.)
Decel. onlyDecel +”After you”Decel +“I’m going to stop”Decel +”Automed driving”
Decel. onlyDecel +”After you”Decel +“I’m going to stop”Decel +”Automed driving”
Decel. onlyDecel +”After you”Decel +“I’m going to stop”Decel +”Automed driving”
Decel. onlyDecel +”After you”Decel +“I’m going to stop”Decel +”Automed driving”
■Very confident ■Confident ■Others
• The messages gave confidence to the subjects except for “Automate driving”.
Sense of security
AIST Univ. of Tsukuba Keio Univ. DENSO Tokyoto Business Service
SIP-adus Human Factors and HMI Research Consortium
➢Question;
Do Automated Vehicles need to fully understand
the intentions of other road users and react
accordingly?
➢Answer; Yes.
AIST Univ. of Tsukuba Keio Univ. DENSO Tokyoto Business Service
SIP-adus Human Factors and HMI Research Consortium
29
Optimistic anticipation (personal opinion)N
um
ber
of
con
flic
ts b
etw
een
AV
an
d o
ther
road
users
No external
communication
One-way external communicationAV ⇒ Other road user
Two-way external
communicationAV ⇔Other road user
Complicated situations
requiring negotiations
Situations requiring multiple
communications
AV
Com 1
Com 2
You must stop!
Statements yes/no
ERTRAC/CARTRE Joint Stakeholder Workshop
• We need additional external HMI for Automated Vehicles to coordinate actions with other road users.
• Automated Vehicles must fully understand the intentions of other road users and react accordingly.
Interactive Symposium - Vienna, 20 April 2018
Social acceptance & user awareness
Andreas Eustacchio, Attorney at law
Statements yes/no
ERTRAC/CARTRE Joint Stakeholder Workshop
• Accident liability should be removed from drivers of conditionally automated cars (SAE Level 3) who show typical and reasonable user behaviour.
• People should have the freedom to change options for the decisions taken by the cars (e.g. driving style).
C AR TR E 2 018 : I N TER AC TI V E SYM P O SI U M O N R ESEA RC H & IN N OVATI O N FO RC O N N EC TED A N D AU TO M ATED D R I V I N G I N EU R O P E
AUTOMATED DRIVING
SOCIAL ACCEPTANCE AND USER AWARENESS
V IE N N A , A P R IL 2 0 , 2 0 1 8
Dr. Andreas Eustacchio, LL.M. (London LSE)
Rechtsanwalt (Attorney-at-Law, Austria)
Hon.Prof.(FH) – Cavaliere
SMART CARS
“Like it or not, the robots are slowly taking over a driver's chores.”
Popular Science Magazine, 1958
34
SMART CARS?
35
▪ connected and intelligent vehicles
▪ combination of mechanical and digital technologies
▪ new ways of communication, information, safety and security
▪ HAV (highly autonomous vehicles) and the humanoid robot on four wheels by way of AI (artificial intelligence)?
SOCIAL BENEFITS – SOCIAL ACCEPTANCE
▪ increase in road safety▪ around 90% of all accidents caused by human error
▪ just a small percentage due to a technical fault related to the vehicle
▪ „crash-optimizing“
▪ power efficiency und e-mobility? “platooning”
▪ „the last driver license holder is already born“ (Mario Herger)
▪ mobility for elderly people, children, physically disabled persons
▪ start-ups, innovation, new business models
36
(UN)WANTED SIDE-EFFECTS?
▪ cost-reduction in the logistics industry
▪ pay as you drive – new insurance models?
▪ hacking and „carnapping“ – security comes first?
▪ big brother is watching you – who will get your data?
▪ what happens with your data when you sell your car?
37
AUTOMATED DRIVING AND THE LAW?
Smart Cars
product liability, product surveillance and product recall
necessity of a driving license?
Infrastructure (5G), car2car,
and test tracks
safety, civil liability, insurance
connected driving and BIG DATA:
data protection and data security
ethical questions: who shall be protected?
criminal liablity
national/international technical standards,
risk assessment
international road traffic agreements
patent law and software license agreementsnational road traffic regulations
are car producers telecom service provider?
38
software-developer as producer?
databox: the cars‘
black box?
Dr. Andreas Eustacchio LL.M. (London LSE)
Rechtsanwalt (Attorney-at-Law, Austria)
Hon.Prof.(FH) – Cavaliere
CARTRE 2018 – Vienna: Interactive Symposium on research &
innovation for connected and automated driving in Europe
Interactive Symposium - Vienna, 20 April 2018
Testing & Validation proceduresArjan van Vliet, RDW
Statements yes/no
ERTRAC/CARTRE Joint Stakeholder Workshop
• Legal admission of SAE Level 3 automation requires human in the loop evaluation of take-over procedures.
• Legal admission of SAE Level 4 automation in vehicles with steer and pedals does not require human in the loop evaluation of take-over procedures.
Symposium on Research & Innovation for
Connected and Automated Driving in Europe
Arjan van Vliet
Vienna, 20 April 2018
RDW’s field of work
CommunicationHD Maps
Vehicles becoming:
• Ever changing
• Ever learning
• Ever communicating
But what about the driver?
Dutch ADAS
practice
• Low penetration rate ADAS <5%
• No clear definitions ADAS
• No clear registration ADAS
• Tax focus on CO2
• Poor instruction new users
Thank you for yourattention!
Interactive Symposium - Vienna, 20 April 2018
Next steps/Future researchneeds
CARTRE and SCOUT are funded by the European Union Horizon 2020
Work Programme
Thank you!www.connectedautomateddriving.eu