This workshop will be held at the 33rd IEEE Intelligent Vehicles Symposium (IV), on June 5th, 2022, in Aachen, Germany in the afternoon.
Location: Eurogress, Conference Room 2, Sunday 13:00h CEST
Scope and Topics
Research on Automated Vehicles enabled safe prototypes, driving in public traffic. However, safety comes at the price of overly conservative inconvenient behavior, e.g. in unprotected left turns or merging scenarios. Reasons can include prediction errors in vehicle interactions or lack of probabilistic considerations in motion planning. While it must be ensured that safe and possibly uncomfortable reactions to worst-case behaviors of others remain feasible, the intended reaction should be comfortable.
Approaches providing sufficient and comfortable actions require sophisticated behavior prediction for other traffic participants. Predictions with maneuver options for other vehicles are needed, which cannot be evaluated against one single “right prediction”. Options may depend on destinations, driving styles and even emotions. Novel methods, not limited to machine learning or common evaluation schemes, are needed.
Similarly, motion planning in highly interactive scenarios lacks one single “ground truth”. In public traffic, a good plan should lead to stereotypical behavior, predictable by other road users. It’s realization requires research on the combination of planning and prediction and its evaluation and benchmarking.
The organizers support research with the published INTERACTION dataset https://interaction-dataset.com/ of trajectory data from highly interactive traffic scenarios on high-definition maps but all contributions within the scope of the workshop are welcome, with or without use of the dataset.
|Time PDT||Time CEST||Speaker||Affiliation||Talk Title|
|04:10:00||13:10:00||Joseph Gesnouin||MINES ParisTech/VEDECOM||Assessing Cross-dataset Generalization of Pedestrian Crossing Predictors|
|04:30:00||13:30:00||Hannes Weinreuter||Karlsruher Institut für Technologie: KIT||Cooperation behavior and complexity at inner city intersections|
|04:50:00||13:50:00||Christoph Burger||Karlsruher Institut für Technologie: KIT||Interaction aware trajectory optimization using multi-agent models|
|05:10:00||14:10:00||Julian Schmidt||Mercedes-Benz AG/University of Ulm||Vehicle Trajectory Prediction with Crystal Graph Convolutional Neural Networks and Multi-Head Self-Attention|
|05:30:00||14:30:00||Linda Miller||University of Ulm||How should automated vehicles behave? On social driving behavior and drivers’ interpretation of vehicle motion|
|06:10:00||15:10:00||Max Mertens||University of Ulm||Probabilistic Prediction and Maneuver Planning for Connected and Legacy Vehicles at Unsignalized Intersections|
|06:30:00||15:30:00||Faris Janjos||Bosch||SAN: Scene Anchor Networks for Joint Action-Space Prediction|
|06:50:00||15:50:00||Steffen Busch||University of Hannover||LUMPI: The Leibniz University Multi-Perspective Intersection Dataset|
|07:10:00||16:10:00||Topan Sever||McGill University/NVIDIA||Interaction-Dynamics-Aware Perception Zones for Obstacle Detection Safety Evaluation|
|07:30:00||16:30:00||Karen Leung||NVIDIA/University of Washington||Towards Data-Driven Safety Synthesis for Safe AV Interactions|
- Sascha Hornauer, MINES ParisTech. Email: email@example.com
- Maximilian Naumann, Bosch Center for Artificial Intelligence (BCAI)
- Eike Rehder, Mercedes-Benz AG
- Jiachen Li, Stanford University
- Wei Zhan, University of California at Berkeley
- Martin Lauer, Karlsruhe Institute of Technology (KIT)
- Masayoshi Tomizuka, University of California at Berkeley
- Arnaud de La Fortelle, MINES ParisTech
- Christoph Stiller, Karlsruhe Institute of Technology (KIT)
Please get in touch with the organizers in case you have any further questions.
At IV2021, the organizers hosted a previous edition of this workshop: https://kit-mrt.github.io/iv2021-workshop/.