AI-DO 6: NeurIPS 2021

by
The AI Driving Olympics 2021 are possible thanks to the generous sponsorship of:
Organized by:

Andrea Censi, Jacopo Tani, Emilio Frazzoli (ETH Zurich), Liam Paull (University of Montreal), Holger Caesar (Motional), Matthew R. Walter, Andrea F. Daniele (TTIC), Sahika Genc (Amazon), Sharada Mohanty (AICrowd), Abhinav Valada (University of Freiburg).

Duckietown: learning autonomy

Welcome to the 2021 AI Driving Olympics!

UPDATE! Check the final event schedule here. 

AI-DO Overview

The AI Driving Olympics (AI-DO) is a series of embodied intelligence competitions in the field of autonomous vehicles. The objective of AI-DO is to provide accessible mechanisms for benchmarking progress in autonomy applied to the task of autonomous driving.

Read more about the AI-DO objectives here.

AI-DO 2021

AI-DO 2021 final event will take place at the Neural Information Processing Systems conference (NeurIPS 2021) edition and encompass three leagues: urban driving, advanced perception, and racing.  

The Leagues

Leagues each present several tasks to solve (“challenges”) and run with independent platforms and logistics. Read on to know more!

Urban Driving with Duckietown

In this league, you will submit robotic agents using the Duckietown platform for solving precision autonomous driving tasks in simulation and on real Duckiebots: lane following, lane following with other vehicles, lane following with intersections, and lane following with vehicle and intersections. 

Get started with the Urban league
Advanced Perception with nuScenes

In this league you will use the nuScenes dataset, to tackle three perception challenges related to real-world autonomous vehicles: object detection, panoptic lidar segmentation and panoptic lidar tracking.

Get started with the Advanced perception league
Racing with AWS Deepracer

In this league, you will use the AWS Deepracer platform to train a reinforcement learning agent (i.e., an autonomous car) to drive by interacting with its environment and a simulated track, to complete a lap as fast as possible without going off-track while avoiding crashing into the objects on the track.

Get started with the Racing league

The Final Event

Winners for all leagues will be announced during the competition track of NeurIPS on Friday, Dec. 10th, 2021.

Note: the event can be accessed live by registering at the NeurIPS conference. 

Schedule

The AI Driving Olympics 2021 will have two phases:

Happy competition!

AI-DO 3 Summary

by

In case you missed it AI-DO 3 has come and gone. Interested in reliving the competition? Here’s the video.

We had a great time at NeurIPS hosting the Third Edition of the AI Driving Olympics. As usual the sound of Duckies attracted an engaging and supportive crowd.

Racing League - AWS DeepRacer

The competition began with the Racing Event, hosted by AWS DeepRacer. They ran their top 10 submissions and selected the winner by who could complete the fastest lap.

Racing Event Winner
Ayrat Baykov at 8:08 seconds

Advanced Perception League - nuScenes by aptiv

The winners of the Advanced Perception Event hosted by APTIV and the nuScenes dataset were announced. Luckily a member of the winning team was present to accept the award.

Rank 3
CenterTrack – Open and Vision

Rank 2
VV_Team

Rank 1
StanfordlPRL-TRI

Urban Driving League - Duckietown

The competition culminated with Duckietown’s Urban Driving Event, where the top submissions for each of the three challenges were run on the competition tracks.

 

Lane Following 
JBRRussia1: Konstantin Chaika, Nikita Sazanovich, Kirill Krinkin, Max Kuzmin

Lane Following with Vehicles
phmarm

Lane Following with Vehicles and Intersections
frank_qcd_qk

Congratulations to all the winners and thanks for participating in the competition. We look forward to seeing you for AI-DO 4!

AI-DO 2 Winners!

by

The second edition of the AI-DO was held in May 2019, in Montreal, Canada at the International Conference on Robotics and Automation (ICRA).

The competition comprised of one league, Urban Driving, run by the Duckietown Foundation. 

There were 3 challenges in the league. 

Lane following (LF) where a single Duckiebot drives within the lane in a closed road loop without intersections nor obstacles. 

Lane following with vehicles (LFV) where vehicles were added to the road and the Duckiebot running the competitors algorithm avoids making contact while still lane following. 

Lane following with vehicle and intersections (LFVI) where the competing Duckiebot navigates a road map that includes intersection while also avoiding contact with other driving vehicles. 

The finals were kicked off by a semifinals round, where we the top 5 submissions from the Lane Following in Simulation leaderboard. The finalists (JBRRussia and MYF) moved forward to the more complicated challenges of Lane Following with Vehicles and Lane Following with Vehicles and Intersections.

Results from AI-DO 2 live finals event on May 22, 2019 at ICRA

In the end JBRRussia came out on top for all three challenges!

If you couldn’t make it to the event and missed the live stream on Facebook, here’s a short video of the first run of the JetBrains Lane Following submission.

Thanks to everyone that competed, dropped in to say hello, and cheered on the finalists by sending the song of the Duckie down the corridors of the Palais des Congrès.