AUTOMOTIVEUI '23
AutomotiveUI (or short: AutoUI) is the International ACM SIGCHI Conference on Automotive User Interfaces and Interactive Vehicular Applications. It is the premier forum for UI research in the automotive domain. The conference annually brings together over 200 researchers and practitioners interested in both the technical and the human aspects of in-vehicle user interfaces and applications, to provide a forum for the exchange of technical information concerning research (and practice) and educational activities for motor vehicle user interface development. We have multiple meeting categories in which researchers, practitioners, and other interested parties can take part in our conference and community. We welcome you to engage with us in this exciting field!
What Do You Expect for Your AV? The 2nd Workshop on Behaviors of Autonomous Vehicles in Ambiguous Driving Scenarios
AVs Ambiguous Driving Scenarios
Human drivers are being gradually replaced by highly automated driving systems, and this trend is expected to continue. Alternatives should be available if driving algorithms are incapable of resolving ambiguous driving scenarios. What occurs if an autonomous vehicle follows a vehicle traveling below the posted speed limit? Should the autonomous vehicle cross the leading vehicle or maintain a safe distance? We must have solutions to address such situations. The way an autonomous vehicle responds to a variety of ambiguous driving scenarios is crucial for legal and safety reasons. To improve future road safety and convenience, this workshop aims the enhancement a framework to develop various ambiguous driving scenarios and plausible actions of AV in each of them. The results of this workshop will be an aid to scientists in their strategic policymaking and algorithm design for AVs responses to ambiguous driving scenarios.
SCHEDULE AND ACTIVITIES
The workshop is scheduled in 18th September 09:00 - 13:00
It will consist of three sessions to answer three RQs,
• RQ1: What type factors and entities are included in ambiguous driving scenario framework?
• RQ2: What user perceptions are significant for determining AVs plausible actions in different ambiguous driving scenarios?
• RQ3: What are the possible actions of AVs in different types of ambiguous driving scenarios?
respectively. The organizers will send participants a short questionnaire in advance to survey their interests and to finalize the discussion agenda. Table 1 presents the tentative schedule and activities
It will consist of three sessions to answer three RQs,
• RQ1: What type factors and entities are included in ambiguous driving scenario framework?
• RQ2: What user perceptions are significant for determining AVs plausible actions in different ambiguous driving scenarios?
• RQ3: What are the possible actions of AVs in different types of ambiguous driving scenarios?
respectively. The organizers will send participants a short questionnaire in advance to survey their interests and to finalize the discussion agenda. Table 1 presents the tentative schedule and activities
Workshop Introduction
The workshop will begin with a brief introduction of the organizers and topics. The goal, RQs, and activities will be conveyed to the participants. As all activities will be conducted as a group, the participants will be split into four or five groups (depending on the number of participants). Before beginning the primary sessions, all participants will be asked to introduce themselves. |
Session 1
We will first briefly present relevant works to facilitate understanding the topics/research questions and highlight the importance of the research area. Then, we will provide images and descriptions of various ambiguous driving scenarios involving AV and the draft framework developed by the research group. In each group, participants will discuss plausible actions for the AV in each ambiguous driving scenario, based on the list of determinants and their respective areas of expertise. They can also add relevant entities to the framework developed. They will be required to write their responses on the provided paper so that the results can be readily distributed. |
Session 2
Using the Brain Storming Wall, we will look at AV alternatives and considerations in ambiguous driving scenarios in Session 2. The elicited scenarios (i.e., scenario drawings, descriptions and updated framework) will be affixed to the wall, and the participants will be asked to freely contribute their thoughts in two categories (AV's alternatives and considerations). |
Session 3
In the final session, we will execute a brief survey to collect preferences regarding the behavior and policy of autonomous vehicles (AVs) in ambiguous driving scenarios. We will discuss the survey results and the significant evaluation factors they consider, and then discuss the future directions of research. |
Attendance
We anticipate there will be approximately twenty participants there, excluding the people who organized the event. In addition to being the regular attendees of the AutomotiveUI workshop, we will also try to get in touch with scholars and practitioners who may have an interest in the subject matter of the workshop's discussion. In addition, the number of participants is subject to change based on the structure of the workshop. Because the scholars are interested in the subject matter of the workshop, we anticipate having a larger number of attendees. |
EXPECTED OUTCOME
The main outcome will be a roadmap to answer the three RQs:
• A framework to develop AVs ambiguous driving scenarios • Relevant user priorities and perceptions for determining AVs plausible actions in different ambiguous driving scenarios • Plausible actions of AV in different ambiguous driving scenarios and considerations for the preferences and evaluations on driving style and policy. These outcomes will facilitate future discussions on: • introducing the state-of-art research of driving styles to researchers and practitioners, • identifying research gaps and challenges for the implementation of AVs, • exploring the design of AV driving styles and facilitating future collaborations. We will post the results on the workshop website to share them with participants. The results will include a summary of discussions and pictures of interactive activities, including brainstorming walls and scenario drawings. We will seek to publish academic findings of the results in related venues, e.g., the future ACM AutomotiveUI conference. Further, given that this workshop is a starting point with the expert group, a follow-up study might be valuable. We plan to conduct an experimental or survey study involving normal users. |
ACKNOWLEDGMENTS
This work was supported under the framework of international joint research program managed by the National Research Foundation of Korea (NRF-2022K1A3A1A18080244)
This work was supported under the framework of international joint research program managed by the National Research Foundation of Korea (NRF-2022K1A3A1A18080244)