Technical Steering Committee (TSC) Meeting #27 (2021/02/17) Minutes

2021/02/17 15:00 UTC

Attendees

  • Geoffrey Biggs (TSC Chair)
  • Jilada Eccleston (Tier IV)
  • Ryohsuke Mitsudome (Tier IV)
  • Stefano Longo
  • Akihiko Tsukuda (eSOL)
  • Shinpei Kato (AWF)
  • Seonman Kim (LGE)
  • Tsutomu Otake (Macnica)
  • Hiroshi Igata (ODDWG)
  • Josh Whitley (AWF)
  • Yang Zhang (Linaro)
  • Adam Dabrowski (Robotec.ai)
  • Yao Zhai (Ziiko)
  • Joe Speed (ADLINK)
  • David Van Guyen (AutonomouStuff)
  • Filipe Rinaldi (Arm)
  • Daisuke Tanaka (AWF)

Minutes:

Action items

Opening remarks and new member introductions (Board)

  • The board meeting minutes summaries will be shared with the TSC from now on.
  • New member: Navya

Confirmation of previous minutes

Minutes approved

Action items review

  • No action items

TSC Chair election

Ryohsuke Mitsudome has been elected as the new TSC chair.

Working group report: Autonomy Software

Autoware.Auto 1.0 will be released this Friday. Discourse and blog post from Josh and SPC to follow. Many thanks to everyone who made this possible. Cargo ODD development starts next week. Many members interested in contributing - should have a large development team for cargo and racing.

Project Manager has accepted the position - starts on or near March 1st. Will take some time to ramp up.

F1TENTH going well - started new hardware platform selection process. Working with developers recruited by Kat (Open Robotics) to get software working.

IAC going OK but slower than hoped. Clemson is still working on a demo lap with Autoware. AutonomouStuff to provide bench setup of all sensors for Josh to help configure and test. ROS2 sensor drivers are still a problem. ADLINK provided 2x computers to AWF for SiL configuration.

Working group report: Operational Design Domain

The issues with scenario tools have all been solved and I’ve set the license term to the Gitlab repository to make it easier for the members to contribute as per advice from a member.

As some of the WG members do not seem to have enough time to finish the assigned scenarios, some of the scenarios have been reassigned to new members.

We might need a couple of more weeks to complete the machine readable scenario set. I’m doing some follow ups, and may consider further reassignment as per necessity.

Tier IV is in the final process of open sourcing of the scenario tools and aiming to complete it by next week.

The ODD WG wiki page on Gitlab is being updated with the information of the achievements so far, for new members’ convenience.

I started to join some of the AWF meetings for racing projects to understand the expectation to ODD WG.

Cargo ODD demonstration capability from Ziiko

ADLINK moves to work with Ziiko on a joint marketing and demonstration opportunity.

Linaro seconds.

Updates on Racing ODD

Racing ODD work to be done:

  • MPC Controller
    • Needs significant improvements or replacement
  • NDT Localization
    • Needs significant improvements or replacement
  • High-level Route Planning
    • Good enough
  • Local Trajectory Planning
    • Needs improvements
  • Lidar-based Object Detection
    • Needs significant improvements
  • Visualization
    • Needs some improvements

Features to add

  • Flash LiDAR driver
  • Object Classification - Low Precision
  • Camera Processing Pipeline - High Precision
  • Radar Processing Pipeline - High Precision
  • Multi-modal Object Fusion - High Precision
  • Local Odometry from IMU Data + Wheel Encoders - High Precision
  • GPS Localization - High Precision