Wednesday February 19, 2020 15:00
- Geoffrey Biggs (TSC chair)
- Cheng Chen (Autocore.AI)
- Kenji Funaoka (Tier IV)
- Shinpei Kato (AWF board)
- Nobuyoshi Nakajima (ITD Lab)
- Tsutomu Otake (Macnica)
- Dejan Pangercic (Apex.AI)
- Filipe Rinaldi (Arm)
- Paul Sastrasinh (TRI-AD)
- Joe Speed (ADLINK)
- Stephane Strahm (Kalray)
- David Vangeyn (AutonomouStuff)
- Holly Watsonnall (StreetDrone)
- Josh Whitley (AWF)
- Yang Zhang (Linaro)
Minutes Geoffrey Biggs
- Opening remarks and new member introductions (Board)
- Confirmation of previous minutes (All)
- Follow-up Action items from previous meeting (All)
- Admin Election of Technical Steering Committee chair
- Working group report Autonomy Software (@gbiggs, @esteve)
- Working group report Autonomy Hardware (@cheng.chen, @sstrahm)
- Working group report Maps and Routing (@Brian_Holt, @mitsudome-r)
- Announce suspension of Maps and Routing working group
- Provide thirty minutes minimum in the March TSC meeting for Tier IV to present their new
- Provide a detailed presentation on Tier IV’s new proposed architecture
- Provide detailed architecture design documentation to the TSC for evaluation with the above presentation or as soon as possible after the March TSC meeting
- Run an election for the TSC chair
- AWF board
- Provide suggestions for the next ODD to set as our goal after AVP
- All TSC members
Opening remarks and new member introductions (Board)
- New Autoware Foundation full-time staff member: Josh Whitley (software architect)
- New member company: Embotech (planning experts)
- New member company: ADLINK
Confirmation of previous minutes (All)
Action items from previous meeting (All)
None from previous meeting
Election of Technical Steering Committee chair
The Autoware Foundation charter requires that the Technical Steering Committee elect its chair once per year. We are now overdue for an election, so one will be held during early March.
- Autoware.AI development is barely moving.
- Large number of issues that were held back from 1.13 due to not being ready, and they are still in the same state. Most of them have not had any movement since then. Out of 28 MRs held back, only one has had new commits based on the review feedback.
- There have been a couple of new minor features made by AutonomouStuff, but no major feature additions.
- At the current pace of development, 1.14 will ship with a few bug fixes and no major new features.
- Lack of maintainers for Autoware.AI
- With most of the focus of the Autoware Foundation’s members and the Autoware community in general now on Autoware.Auto, Autoware.AI is lacking in developer manpower. We have project leads but we need someone who can act as a maintainer for Autoware.AI to assist Josh Whitley.
- Tier IV is developing significant new features for Autoware.AI, but they won’t make it into 1.14. These features are being developed in a closed-source fork. Tier IV wants to contribute these new features to open source. Tier IV wants to know if the features should be put into 1.15 or should be made Autoware.AI 2.0.
- The AWF would like to see design documents for the new architecture that Tier IV wants to propose, including architectural diagrams and design rationale.
- Tier IV will present their new architecture in the March TSC meeting. At least 30 minutes will be set aside for presentation and discussion.
- Tier IV will provide detailed design documentation for the new architecture, including rationale and architectural diagrams. This documentation should be made available as soon as possible after the March meeting, but ideally alongside it to aide in discussion.
- The TSC will make a decision on how it wants to move ahead with Tier IV’s proposal, and Autoware.AI in general up to the end of its life, in either the March TSC meeting or April TSC meeting.
- Work continues apace on the Autonomous Valet Parking capability milestone and associated demonstration.
- Autoware.Auto now has an NDT implementation.
- A new vehicle interface base has been developed with some basic safety features. An improved interface to LGSVL has also been developed.
- The LiDAR object detection pipeline has been re-enabled.
- The work continues to be coordinated in the weekly Autonomy Software WG meetings. Anyone interested in participating should join this working group.
- Tarun Belagodu has volunteered to help out with managing marketing of the hackathon and demo.
AVP2020_report_feb2020.pdf (2.9 MB)
- Q: Is the AVP demo going to use vision, or is all object detection done using LiDAR?
- A: It is entirely done using LiDAR due to the relatively simple response to objects. We hope that the next ODD will be sufficiently complex enough to bring in vision aspects.
Working group report: Autonomy Hardware (@cheng.chen, @sstrahm)
Working Group Situation and Plan
We have sent a note to the Board on Feb the 9th about situation in this group. Here it is.
The Hardware Autonomy Group is not functional.
Why has it not been working so far?
- Autoware Foundation is seen as an Open Source Foundation. Users are looking for software. Most of Open Source foundation are providing software, not hardware
- The maturity of the software and portability has not been explicit so far. In 2019, we have been focusing on functional software at prototyping stage. The port and optimization on other hardware is a need that we are foreseeing for 2020
- Above reasons result in lack of audience to the conference calls
- The ambition to define a specification was too early. Without support by members, it was unrealistic
What could be our 2020 objectives?
I would like to share here my suggestion to move forward.
- The positioning of Hardware Autonomy Group scope must be explicit
- This group should be a vector for AWF members to
- Promote their hardware solutions running Autoware software: autoware.AI or/and autoware.Auto
- Keep an active list of hardware in use by member and demonstrated (namely for the time being: TierIV platform, Autocore, APEX.ai, Kalray, and maybe others)
- Collect platforms documentations: specification, user’s guide (installation and execution of Autoware software)
- Point to platform provider (contact at hardware company)
- Help on most appropriate hardware configuration according to ODD
- Map the listed hardware to ODD
- Recommend and list sensors per ODD (supported, or way to support)
- Collaboration with Autoware.IO for sensors and drivers
- Q1 2020
- Interview (via discourse) of members for used platform
- Create list and organize information for classification preparation (ODD)
- Map hardware to ODD
- Q2 2020
- Collect specification and user’s guide for completing « hardware recommendation »
- Q3 2020
- Q4 2020
- Open visibility on official Website Autoware
Vision Statement Proposition
“The Hardware autonomy Group is helping Autoware adopters to identify, select, configure and deploy appropriate hardware for reduce time to development and time to market by promoting members’ solutions”
I think for following years, the group can bring more and especially
- Helping defining performance characterization for ADAS/AD system as we could leverage companies working on software and hardware
- Helping on approaching safety concepts by providing SW and HW assumptions of use within a system
- Driving sensors integration and fusions approaches with recommendations, example maybe up to interfaces, best practices…
- … not an exhaustive list
I hope we could spend time in February to adapt and agree on such objectives.
With such objectives, I would be glad continuing my role side-by-side with Chen in this context if you are willing to extend my role in 2020, if the Board and TSC agree.
With high motivation and strong expectations for AWF
It sounds like a reboot? It is. Let’s make it happen.
Next immediate Step
- Communicate plan to TSC and having assessment: Stephane on Feb 18th
- Contact known “hardware users” for collecting information on hardware in use
- Initiate request at TSC Feb 18th
- Open questionnaire in Discourse
- First known users: TierIV, NXP, Autocore, AutonomouStuff, Kalray
- Consolidate plan and communicate it
Working Group Meetings report
- Feb 5th
- Feb 18th
- Will be publish later today
- Cheng Chen (Autocore - Autoware WG)
- Josh Whitley (Autoware - Autoware Arch)
- Stephane Strahm (Kalray - Autoware WG)
- Kenji Funaoka (TierIV)
- The board is considering the situation and the proposal seriously. The board does not want to move too hastily, and ensure we solve the real problem that is causing low participation on the hardware side.
- Having visible, achievable milestones may help; it seems to have worked well on kickstarting the software side.
- Should the AHWG do the reference design?
- It is not possible to create a reference design without first making the problem statement, either drafting the ODDs that will be targetted. It is likely that we will need multiple reference designs targeting different ODDs.
- Move forward with the proposed Q1 2020 tasks.
- Revisit this topic in the March TSC to discuss what to do next.
Working group report: Maps and routing (@Brian_Holt, @mitsudome-r)
Since there is such a large overlap between the Maps and Routing working group and the Autonomy Software working group while we work towards the April demo, we propose that we suspend activity of the Maps and Routing working group until that demo is complete and then re-evaluate at that point.
- Accept the WG leaders’ proposal. The Maps and Routing WG will be suspended until the May TSC meeting.
What to do after the AVP demo
- Tier IV has access to demo space during the time frame of the Tokyo Olympics (July/August).
- Tier IV wants to give the AVP demo in this demo space, but would also like to show something more.
- One potential idea is to show the AVP demo on a variety of different hardware platforms.
- Tier IV would like to add some kind of public road demonstration.
- We need to choose a next step that is not too large a jump from the simple and closed environment of a car park.
- TSC members need to start thinking about the next ODD we want to tackle once the AVP work is complete.