Overview – Waymo

Challenge 1

Motion Prediction

Given agents' tracks for the past 1 second on a corresponding map, predict the positions of up to 8 agents for 8 seconds into the future.

Challenge 2

Interaction Prediction

Given agents' tracks for the past 1 second on a corresponding map, predict the joint future positions of 2 interacting agents for 8 seconds into the future.

Challenge 3

Real-time 3D Detection

Given three lidar range images and the associated camera images, produce a set of 3D upright boxes for the objects in the scene.

Challenge 4

Real-time 2D Detection

Given a set of camera images, produce a set of 2D boxes for the objects in the scene, with a latency requirement.

Winners

Challenge 1: Motion Prediction	Method	Technical Report	Authors	Country	Affiliation
First Place	Tsinghua MARS - DenseTNT	Report	Junru Gu, Qiao Sun, Hang Zhao	China	Tsinghua University
Second Place	ReCoAt	Report	Xiaoyu Mo, Zhiyu Huang, Chen Lyu	Singapore	Nanyang Technological University
Third Place	SimpleCNNOnRaster	Report	Stepan Konev, Artsiom Sanakoyeu, Kirill Brodt	Russia, Germany	Skolkovo Institute of Science and Technology, Heidelberg University, Novosibirsk State University

Challenge 2: Interaction Prediction	Method	Technical Report	Authors	Country	Affiliation
First Place	HeatIRm4	Report	Xiaoyu Mo, Zhiyu Huang, Chen Lyu	Singapore	Nanyang Technological University

Challenge 3: Real Time 3D Detection	Method	Technical Report	Authors	Country	Affiliation
First Place	AFDetV2	Report	Yihan Hu, Zhuangzhuang Ding, Runzhou Ge, Wenxin Shao, Li Huang, Kun Li, Qiang Liu	USA	Horizon Robotics
Second Place	CenterPoint++	Report	Tianwei Yin, Xingyi Zhou, Philipp Krähenbühl	USA	University of Texas, Austin
Third Place	X_Autonomous3D	Report	Lu Tian, Jinzhang Peng, Rongzhang Zheng, Han Liu	China	Xilinx

Challenge 4: Real Time 2D Detection	Method	Technical Report	Authors	Country	Affiliation
First Place	LeapMotor_Det	Report	Fenfen Wang, Qiankun Xie, Lindong Li, Yaonong Wang, Hongtao Zhou, Han Xu, Yu Wang, Xiubo Ye	China	Leapmotor Technology
Second Place	DIDI MapVision	Report	Yueming Zhang, Xiaolin Song, Bing Bai, Tengfei Xing, Chao Liu, Xin Gao, Zhihui Wang, Yawei Wen, Haojin Liao, Guoshan Zhang, Pengfei Xu	China	Tianjin University, DIDI
Third Place	dereyly_self _ensemble	Report	Nikolay Sergievskiy	Russia	XIX.ai

Past Challenges - 2020

The past challenges below do not have any awards, but the leaderboards are available for submissions.

General Challenge Information

Each team member must agree to the Official Challenge Rules here.
The submission formats are defined on each challenge’s page.
The metrics for each challenge are defined on each challenge’s page.
You may enter by completing the steps identified in the Official Challenge Rules during the Submission Period.
Submissions may not be created using any data other than the Waymo Open Dataset. However, you may pretrain your models on one of the following datasets if you wish: ImageNet, Coco, or Kitti.
Each entrant (up to 10 team members) can submit for evaluation against the Validation Set as many times as desired.
To appear on the leaderboard, entrants must submit their results for evaluation against the Test Set.
Each entrant can submit for evaluation against the Test Set up to 3 times within 30 days. Submission will fail if that number is exceeded. If you are caught using multiple Google accounts for the purposes of circumventing this limit, your submission will be removed and you will be disqualified from the challenges.
Uploading a new submission with the same Google Account and method name as a previous submission will overwrite it.
Try to avoid using the same method name as other methods on the leaderboard.
Do not submit multiple times for the same method with different object types. Combine them into one submission.
Awards will be given to the top three entries on the leaderboard of each of the challenges after submissions of technical reports, verification of eligibility, and other documentation are complete.
Waymo reserves the right to remove any entry at its discretion pursuant to the Official Challenge Rules.

Challenges

Timeline

Awards for Each Challenge