End-to-End Autonomous Driving:

Emerging Tasks and Challenges

CVPR 2023 Workshop

June 18, 2023, Vancouver, Canada





  • Contact us at workshop-e2e-ad@googlegroups.com




    • Overview

    Welcome to the CVPR 2023 Workshop on End-to-End Autonomous Driving!

    This workshop aims to bring together leading researchers and practitioners to discuss upcoming paradigms for autonomous vehicles. Central to the program is a series of invited talks and four new challenges in the self-driving domain. Each challenge combines new perspectives of multiple components in perception and planning compared to conventional pipelines. Winners of the challenges will present their results and insights as part of the workshop. We invite researchers around the world to build new algorithms to tackle these challenging, real-world autonomous driving tasks!

    We have a sibling workshop on ICLR 2023. The submission deadline is Mar 1st Anywhere on Earth. Check it out!

    Contact us at workshop-e2e-ad@googlegroups.com.

    Sign up for updates via the link.

    Date

    Challenge

    →15 Feb 2023: Challenge Official Announcement
    →15 March 2023: Test Data Release, Server Open
    →Late May 2023: Challenge Submission Deadline
    →4 Jun 2023: Notification of Award
    →15 Jun 2023: Submission of Winners Talk

    Workshop

    →18 Jun 2023, Full day: Workshop Event

    Challenges [Website]

  • We are proud to announce four brand-new challenges this year, in collaboration with our partners - Vision-Centric Autonomous Driving Workshop and the nuPlan team from Motional.

  • The challenge is not solely about winning top place with tricks and ensembles. Each track is set with cash prizes to award both top entries as well as innovative submissions. The award committee will select the most innovative submissions to present their work live.


    • OpenLane Topology Challenge

      World's First Topology Reasoning Benchmark for Scene Structure.

    • Online HD Map Construction Challenge

      Dynamically construct local HD maps using onboard sensors.

    • 3D Occupancy Prediction Challenge

      Occupancy as New Representation for General Object Perception.

    • nuPlan Planning Challenge

      World's First Large-scale Planning Benchmark.



    General Rules at a Glance

  • Each track runs in parallel and would not affect the others. That being said, one can claim multiple awards.
  • We organizers unanimously agree that research should be equal to everyone in the world and apart from other factors. As such, all companies/individuals, regardless of region/belief/etc. can participate in the challenges.
  • Workshop organizers and their teams can participate in a particular track, as long as they are concealed from test data and not a member of the Award Committee.
  • Any endeavors found to hack test data/annotation will lead to disqualification from the competition.

    • For more details, please refer to the full rules here.

    Speakers

    Schedule

    Time Speaker Theme
    9:00, Jun 18 Hongyang Li, Shanghai AI Lab Welcome and opening remarks
    9:10, Jun 18 Speaker 1 Awesome Talk
    9:45, Jun 18 Speaker 2 Awesome Talk
    10:20, Jun 18 Coffee Break Coffee Break
    10:30, Jun 18 Organizers and Winners Track 1: OpenLane Topology Challenge
    11:30, Jun 18 Speaker 3 Awesome Talk
    12:05, Jun 18 Lunch Break Lunch Break
    13:30, Jun 18 Speaker 4 Awesome Talk
    14:05, Jun 18 Organizers and Winners Track 4: nuPlan Planning Challenge
    15:05, Jun 18 Coffee Break Coffee Break
    15:15, Jun 18 Speaker 5 Awesome Talk
    15:50, Jun 18 Speaker 6 Awesome Talk
    16:25, Jun 18 Speaker 7 Awesome Talk
    17:00-17:35, Jun 18 TBD Moderator Panel discussion
    Note: Presentation and winners of Track 2: Online HD Map Construction Challenge and Track 3: 3D Occupancy Prediction Challenge would be scheduled at Vision-Centric Autonomous Driving Workshop.

    Organizers

    Workshop Organizers



  • Hongyang Li

    Hongyang Li

    Shanghai AI Lab
  • Kashyap Chitta

    Kashyap Chitta

    Univeristy of Tübingen



  • Holger Caesar

    TU Delft
  • Shenlong Wang

    Shenlong Wang

    UIUC
  • Ziwei Liu

    Ziwei Liu

    NTU


  • Tai Wang

    CUHK

    • Challenge Organizers


    Track 1: OpenLane Topolopy Challenge




  • Huijie Wang

    Shanghai AI Lab



  • Yang Li

    Shanghai AI Lab

  • Feng Wen

    Feng Wen

    Huawei
  • Yang Li

    Chunjing Xu

    Huawei


  • Hang Xu

    Huawei

    • Track 2: Online HD Map Construction Challenge





  • Yicheng Liu

    Tsinghua University



  • Tianyuan Yuan

    Tsinghua University

    • Track 3: 3D Occupancy Prediction Challenge





  • Hang Zhao

    Tsinghua University



  • Xiaoyu Tian

    Tsinghua University



  • Chonghao Sima

    Shanghai AI Lab



  • Yue Wang

    NVIDIA



  • Lewei Lu

    SenseTime

    • Track 4: nuPlan Planning Challenge





  • Pat
    Karnchanachari

    Motional
  • Feng Wen

    Dimitris
    Geromichalos

    Motional


  • Holger
    Caesar

    TU Delft

    • Steering Committee



  • Yu Qiao

    Shanghai AI Lab



  • Dahua Lin

    Shanghai AI Lab



  • Xiaogang Wang

    SenseTime



  • Wanli Ouyang

    Shanghai AI Lab
  • Jifeng Dai

    Jifeng Dai

    Tsinghua / Shanghai AI Lab


  • Junchi Yan

    Shanghai Jiao Tong University


  • Andreas Geiger

    University of Tübingen
    • OpenDriveLab