The Autonomous Working Group members recently reached a significant milestone, which paves the way to the reference solution of a safe system architecture for self-driving vehicles. We interviewed some of the members to dig into the details.
Launched in June 2021, The Autonomous Working Group Safety & Architecture brings together diverse companies and academia to define the state-of-the-art system architecture for safe self-driving cars, and more precisely, for an SAE Level 4 Highway Pilot.
The Working Group members recently reached a significant milestone, which paves the way to the expected reference solution of a safe system architecture for self-driving vehicles: the finalization of the first report increment, containing the following three channel topics:
- Reference autonomous driving (AD) use cases
- Assumptions and Requirements
- Dependability goals and KPIs (metrics) for architecture evaluation
In this post, hear from three of our Working Group Members, representing different organizations in the automotive space, who strongly believe in the power of collaboration to create the necessary change and solve the greatest safety challenges for autonomous mobility together.
It’s time to highlight their work and perspectives.
You’ve been contributing to The Autonomous first Working Group for nine months and each one of you have been so far responsible for specific channels. Can you tell us more about your channel topic and its relevance to the final public report and reference solution?
Moritz Antlanger (System Architect, TTTech Auto, responsible for the “Reference AD use case” channel): As part of this particular channel, we outlined what an “SAE Level 4 Highway Pilot” AD feature could look like. The process included describing the feature from a user perspective and what environmental and traffic conditions it would support as well as how it would act when it encounters a failure. We, therefore, discussed many safety-related requirements the AD system will need to fulfill. This channel serves as a reference for defining specific dependability goals and KPIs.
Udo Dannebaum (Lead Principal Engineer Automotive Applications at Infineon Technologies, responsible for the “Dependability Goals and KPIs” channel): Automated Driving (AD) systems for vehicles need to demonstrate very high dependability. This means such a system needs to be robust (AECQ-100), safe (ISO26262), and secure (ISO21434) to gain public trust and confidence. The KPIs are vital to assess an AD concept before vehicle production.
Martin Törngren (Professor and Center Director at KTH Royal Institute of Technology, responsible for the “Assumptions and Requirements” channel): As part of this channel, we all agreed that when working with architectures (i.e., design) of complex systems such as ADS, it is essential to clarify and make assumptions explicit, including clarifying rationales and underlying principles.
What has been the biggest challenge so far for your particular channel?
Moritz Antlanger: We encountered multiple challenges in outlining the reference AD use case. There is currently no “Highway Pilot” AD feature on the market, and the introduction of such features may involve gradually expanding functionality. We, therefore, chose a rather conservative starting point, i.e., the minimal initial functionality. AD features will also likely differ between OEMs, so we tried to stay at a sufficiently high level that is likely to apply to most.
Udo Dannebaum: In my case, the biggest challenge was to find the right perspective. As many of you know, I come from the semiconductor industry, which means component/item-level. While working on this channel, I had to free myself from this level; otherwise, remaining there would mean that the concept is already in a solution space, which is not intended here.
Martin Törngren: For my channel, there haven’t been any significant challenges so far. However, there was a lot of effort needed to harmonize views and align the various contributions with the rest of our work. Moreover, the topics we tackle are incredibly challenging, so exciting work lies ahead of us.
How did you work as a team to overcome the challenge?
Moritz Antlanger: We started from an initial proposal and then iterated on it several times. The different perspectives came in handy not to make the outline too specific and thus less applicable to various implementations. We also profited from the multiple suggestions regarding the best formalism and standards we could refer to.
Udo Dannebaum: We found a suitable assessment perspective for the KPIs in several team discussions.
Martin Törngren: We carried out alignment meetings and worked on scoping the goals for the first iteration.
Can you tell us a little more about the learnings you’ve gained from working with different companies and institutions to solve a common challenge for the autonomous driving sector?
Moritz Antlanger: It was exciting and challenging at the same time to realize that each participant in the Working Group has their own focus and point of view. This can sometimes lead to ideas and proposals that sound very different from a theoretical perspective but are often compatible or even similar in the context of a more detailed architecture or design. That’s why we are currently working out a standard format for describing, discussing, and comparing different architectures for an SAE Level 4 Highway Pilot.
Udo Dannebaum: As an engineer at a semiconductor vendor, I am used to being very focused on a single component. The discussions with different companies and institutions allowed me to overcome this and think out of the box.
Martin Törngren: There has been a valuable and constructive discussion climate, including cross-reviews and collaborative work towards the first report. In the upcoming months, I look forward to engaging even more partners in our work and recently suggested an “open call” for architecture proposals.
Thanks for reading and getting to know some of the Working Group Safety & Architecture members.
This milestone achievement will build the basis for the core activity of our Working Group, which is the discussion of various architecture candidates for the given problem statement. We are currently preparing an OPEN CALL asking for your contribution. Stay tuned for more information coming soon!
If you already want to contribute to The Autonomous Innovation Stream and become a member of our Working Group to help accelerate the market readiness and series development of safe autonomous vehicles, don’t hesitate to reach out to us.
To learn more about our first Working Group, check out this article by Working Group Chairman Christoph Schulze.