Latest News / July ‘24 / We are experiencing a surge of new members coming in. The latest are OSRA, Kudan, Synkrotron, REKA and AutoCrypt!

The Autoware Challenge 2023

Author: ADMIN
The Autoware Challenge 2023

The Autoware Challenge 2023 is the software development challenge hosted by the Autoware Foundation and TIER IV aimed at developers of all levels, including students to professionals.

The challenge began on May 15, 2023, and continues until November 1, 2023.

This challenge is sponsored by TIER IV

Overview

Autoware is the world’s first “all-in-one” open-source software for autonomous driving hosted under the Autoware Foundation. TIER IV has been a leader in the development of Autoware and has conducted multiple real-world tests based on Autoware worldwide.

​As one of the main contributors of Autoware, TIER IV has faced many complex challenges through development. TIER IV would like to sponsor a challenge to solve such engineering challenges. Any researchers, students, individuals or organizations are welcome to participate and submit their solutions to our proposed challenges.

Through the challenge, participants will choose one of the challenges that are proposed from AWF / TIER IV and present their solution to the selected challenge. TIER IV will examine the presentation and choose a winner who came up with the best solution that improves Autoware functionality.

Awards

First Place: 15,000 USD
Second Place: 7,000 USD
Third Place: 3,500 USD

*Possible Additional Special Awards

Challenges

The following slides share some of the challenges that AWF is facing. Note that these are only an example challenges provided from AWF, and the participants may also come up with their own unique challenges that they think are beneficial to Autoware development. Then, the participants are to propose the solution to the challenge to improve the performance of Autoware.

Submission

The following are the steps for submitting a solution for the chosen challenge.

​The first step is to submit an abstract on which challenge the participants will be working and an overview of the solution that participants would like to propose. We expect an amount of about a page of A4 PDF or 5-6 presentation slides. TIER IV will then give a reply within 1-2 weeks with feedback on whether the proposal seems worthwhile for us to proceed with the final submission. This is to ensure that participants won’t spend a whole challenge period on topics we are not interested in.

​The second step is to submit a final proposal for the challenge. This is a free-to-choose format. It can be a PDF documentation, a video, a slide deck, or a repository of a working code to explain the proposal’s scope.

​Participants will do a presentation based on their final proposal, and the TIER IV judges will evaluate the presentation and choose a winner, a runner up and a third place. The main criteria TIER will take into account is:

  • How much it improves the capability of Autoware
  • Ease of integration with Autoware
  • Feasibility to run hardware with low computational power
  • Robustness against vehicle/sensor configuration and different ODDs

Timeline

Call for Participation: May 15th

Registration/Abstract Deadline: August 1st September 1st

Proposal Submission Deadline: November 1st

Online Presentation: November 8th

Announcement of the Results: December 1st


Submission Forms


Results

We would like to express our gratitude for your participation in the Autoware Challenge. The presentation underwent a careful evaluation by the judges from AWF and TIER IV, leading to the following rankings:

1st Place

  • Title: Detection of light-states of surrounding vehicles via Light-Skeleton-Detection
  • Team Name: BeIntelli

2nd Place

  • Title: FlexMap Fusion
  • Team Name: TUM FTM Mapping

3rd Place

  • Title: Data-driven system identification optimal control algorithm
  • Team Name: Clothoid

Honorable Mention

  • Title: Enhancing initial pose for localization through Visual Place Recognition and Object geo-localization
  • Team Name: BeIntelli
 


 

Organized By

Ryohsuke Mitsudome

Yukihiro Saito