In recent years, different approaches for motion planning of autonomous vehicles have been proposed that can handle complex situations. This competition provides the opportunity for researchers to compare their approaches with others based on a large database of traffic scenarios.
In this competition, participants are provided with an extensive database of traffic scenarios with planning problems for each of which a feasible trajectory needs to be planned. Apart from the number of solved scenarios, submissions are assessed with regard to efficiency, safety, and compliance with traffic rules.
The competition is divided into two categories:
The traffic scenarios are provided in the CommonRoad format and solution trajectories are uploaded to commonroad.in.tum.de. Our scenario database contains highways, urban environments, and features several types of traffic participants. Tools for parsing scenarios and useful software for motion planning are openly available within the CommonRoad framework.
After the public release of the benchmark problems, solutions can be submitted online before the submission deadline. The best participants will have the opportunity to present their approaches at the final workshop as part of the IEEE International Conference on Intelligent Transportation Systems (ITSC) 2021. Only registered teams will be able to participate in the final workshop.
The task of the jury is to come up with a consensus for working groups, define the problem instances, and ensure a fair competition. Anybody participating in the competition can become a jury member. We accept applications from academia and industry; a short email before March 31 to commonroad@lists.lrz.de suffices.
Each working group has a leader who is responsible for organizational aspects (presiding discussions, ensuring the workflow, etc.). The definition of the problem instances and the evaluation criteria are carried out by the working group.
Submissions consist of XML solution files submitted through commonroad.in.tum.de, where also details on the provided software, the evaluation, and submission guidelines can be found.
See Rules and Details for more information.
The jury will select the winner of the competition who will receive a prize of 3000 $ provided by our sponsor IVEX. To be eligible for the prize, the participant needs to be registered for the workshop and the motion planner needs to be published as open-source. However, a registration for the workshop is not mandatory in order to participate in the competition, and the registration for the workshop is still possible after the submission deadline.
The final Workshop will take place on Sunday, September 19th, 2021 at 10:00 (EDT) / 14:00 (GMT). The link for the Zoom session has already been sent out by the ITSC organizers to registered participants.
Times in EDT:
10:15 - 10:35: Cui Mingyang (Tsinghua University)
Interactive Planning in CommonRoad - An MCTS & Lattice Planner Method
10:35 - 10:55: Shivesh Khaitan (Carnegie Mellon University):
Motion Planning for Self-driving using Short-term and long-term Prediction and CILQR
10:55 - 11:15: Final Discussion & Plans for Future Competitions
We congratulate the team of Shivesh Khaitan, Qin Lin, and John Dolan from the Carnegie Mellon University as the winner of this competition. They receive the award of $3000 sponsored by IVEX.
01.09.21: The end of the competition is near and there are already several successful submissions. To give all participants more time to improve their planning algorithms, we extend the deadline for two days until Friday, 23:59 GMT.
01.08.21: The submissions for the second phase are now open and details on the submission of docker images have been published. The submission for the first phase is still possible until the end of the competition.
06.07.21: The submission for the first phase is open now.
01.07.21: The scenarios and updated software for the first phase have been published, see Rules and Details.
18.05.21: Sign up to the newsletter to get notified about important updates regarding the competition.
18.05.21: The rules and timeline have been detailed, see Rules and Details.
For a PDF version of this call, see here: download.