Human-Vehicle-Environment Cooperation in Automated Driving
The Next Stage of a Classic TopicWorkshop@AutomotiveUI 2021, September 9/10
MOTIVATIONS AND GOALS
It appears that autonomous systems are replacing human in the driving task. However, autonomous driving abilities do not mean that vehicles should not interact with their drivers/passengers anymore. There are still many scenarios where either the automated system cannot handle the driving very well, or human wants to spontaneously influence the behavior of the system to meet their preferences. Thus, beyond the hype of autonomous driving, a large space opens for human-vehicle cooperation at a different level of automated driving. It is expected that the workshop will consolidate existing knowledge of human-vehicle-environment cooperation and provide insight for future works.
In this workshop, three topics will be discussed:
Why does the human driver want/need to intervene in the cooperative system?
The necessity of human driver’s involvement remains controversial, especially among automated driving engineers. Therefore, it is necessary to discuss the fundamental motivation to conduct cooperative driving, thus, researchers will have a base for further investigation in this direction.
Which scenarios are appropriate for human-vehicle cooperation?
The cooperation need not happen in all locations or scenarios as most of the time a qualified automated system should drive by itself. It is necessary to find out the typical use scenarios in which the system either the system or the human intent to cooperate.
By which interaction means can human driver and vehicle accomplish the cooperation?
Normally, an automated vehicle moves fast on the road, which leaves a limited time window for human driver/passenger and the system to exchange their intent, negotiate to meet an agreement of what action will be taken next. The novel interaction technologies, such as gaze-, speech- or brain-interaction will be discussed associated with the use cases.
Related Concepts and Studies
Augmented Driving Concept by Honda at CES 2020
Honda’s Augmented Driving Concept features a seamless transition from autonomous to semi-autonomous driving operation. To respond quickly to the user’s curiosity, the autonomous driving system is constantly on standby, ready to intervene and control the vehicle when needed. The driving system changes between automatic and manual mode with a switch, and features more than eight modes between fully autonomous and semi-autonomous operation. Various sensors in the vehicle continuously read the user’s intention to smoothly shift between these modes, creating an instinctive driving experience.
Designing for Prediction-Level Collaboration Between a Human Driver and an Automated Driving System
A prototype was implemented in a driving simulator driven by a functional AD system that has been partially validated on the public road. We designed and implemented a gaze-button input for intuitive vehicle referencing and a graphical user interface (GUI) for enhancing the explainability of the AD system. Three typical driving scenarios in which an AD could take advantage of the human driver's anticipation to drive more comfortable and personalized were created for subsequent evaluation.
AutoMate
The success of future automated vehicles will depend on how well they interact, communicate and cooperate with humans both inside and outside the vehicle. Within the EU project „AutoMate“ the question of how to turn vehicle automation into a team player for the human driver was addressed. The object of design is not the automation system but the overall driver-automation team. Driver and automation need to cooperatively pursue the driving task by dynamically sharing and distributing tasks depending on driver’s and automation’s current capabilities and situational demands. The developed driver-vehicle interface allows the human driver to interact with the vehicle in situations which the automation cannot handle in an efficient manner alone. In our driving scenario the vehicles sensor cannot grab enough information to assure a safe overtaking manoeuvre and asks the driver to reassure a hazard free road situation. After driver’s reassurance the automation performs the overtaking manoeuvre.
Project Vorreiter
One of the central ideas of the Vorreiter (Trailblazer) project is to provide the collected intelligence of the technical system, i.e. the motion assistance or automation, in such a way that it makes possible actions in the form of maneuvers quickly comprehensible in the respective situation and quickly and safely realizable through intuitive steering gestures on the steering wheel or side-stick, if possible. Examples of maneuver gestures are a lane change, e.g., on the highway or in the city, a turning maneuver, stopping and starting in front of a crosswalk, or a parking maneuver.
Such control via steering gestures could already provide a gain in comfort for drivers without disabilities, and also a gain in safety for fast, safety-critical maneuvers. For younger drivers, the suggestion of a movement path through automation could compensate for the still lacking experience, for older drivers a possibly slower reaction time.
Project website: https://vorreiter.iaw.rwth-aachen.de/de/Projektinhalt.html
Organizers
Chao Wang
Chao Wang works as a Senior Scientist at Honda Research Institute Europe. He received a PhD degree from the Eindhoven University of Technology in 2017. His research focuses on HCI of automated driving, human-robot interaction and explainable artificial intelligence.
Marcel Usai
Marcel Usai is a Ph.D. student and research assistant at the department of Human Systems Integration at the RWTH Aachen University. His research focuses on the realization of intuitive human-machine cooperation through design and observation of interaction pattern.
Jingyi Li
Jingyi Li is a Ph.D. student in Media Informatics at the University of Munich. Her research focuses on the safety and comfort of the rear-seat passenger in their use of interactive technologies in the confined space of the car.
Martin Baumann
Martin Baumann is a Professor and the chair for Human Factor at Ulm University. His research interests are Cognitive Modeling, Human-Machine-Interaction and -Cooperation, Comprehension processes and mental representation. He serves as the co-chair at AutoUI2021.
Frank Flemisch
Frank Flemisch is a Professor for Human Systems Integration at the RWTH Aachen University and a branch head for balanced Human Systems Integration at Fraunhofer. He is stunned that after 30 years of research on highly automated and cooperative driving, there is still a lot to discover, that the best time of the cooperation paradigm is yet to come, and that the best automotive user interfaces and concepts are yet to be invented and developed
Greeting From the Organizers
Workshop Schedule (to be finalized)
The workshop is planned as 100 minutes event on:
Day 1, Sep. 9th. 18:15 - 19:50 (Berlin, UTC +2);
17:15 - 18:50 (London); 09:15 - 10:50 (Los Angeles); 12:15 - 13:50 (New York); 21:45 - 22:50 (New Delhi)
Day 2 (repeating session), Sep.10th. 9:30 - 11:10 (Berlin, UTC +2);
15:30 - 17:10 (Beijing); 13:00 - 14:40 (New Delhi); 08:30 - 10:10 (London);
And is structured as follows:
Part 1 (20 minutes)
Introduction and Presentations
The workshop host will welcome the participants and give an introduction to the schedule as well as the motivation for this workshop.
To kick-off the productive phase of the workshop, two expert presenters will give an impulse presentation of seven minutes each on their understanding of what cooperation is and why we need it. The presentations should give a quick overview selected topics of past and current research, then motivate the upcoming workshop phases.
Part 2 (20 minutes)
Discussion about WHY and WHICH
Beginning with this phase, we start dividing all participants into 5 small groups. We first start with a few minutes to break the ice and get to know each other before further discussing the topic of “Why is cooperation needed?” Most important results are documented on the miro board in form of post-its.
From crowded urban area to fast-speed highway, cooperation fits in many scenarios. The groups that formed beforehand will further discuss potential cooperation use cases ranging across all automation levels, traffic types and environmental conditions etc. All discussed use cases are documented on the Miro board in the form of a small visual and/or textual scenario description.
Break (10 minutes)
Part 3 (20 minutes)
Discussion about HOW
Groups are merged and examples of the results of session 2 are discussed. Afterwards, participants re-gather in their former groups and work on the objective to create an interaction design for at least one example scenario and document it on a prepared Miro frame.
Part 4 (30 minutes)
Wrap-up
Each group gives short presentation of their interaction design with focus on the most important aspects of their design based on the documented graphical representation as well as their understanding of why cooperation is needed in the presented scenario.
CALL FOR PARTICIPATION
We sincerely invite any researcher/designer/practitioners both from the automated system domain, such as autonomous-driving researcher, automotive engineers and sensor specialist; as well as human’s perspective, such as human-machine interaction researchers, psychologist and user experience designers.
For participating, please select "Workshop on Human-Vehicle-Environment Cooperation in Automated driving: The Next Stage of a Classic Topic" when register (you can also go back to the registration website and select the workshop at anytime later) of AUTO UI 2021 conference.
Furthermore, participants will be also invited to present (not mandatory) their current projects related to the workshop topics in a 3-min video, which will be played in the workshop. Please send the video (e.g. via googledrive link) to <chao.wang@honda-ri.de>
CONTACT
If you have questions of the workshop, please feel free to contact Chao Wang (chao.wang@honda-ri.de).