AImotive is misleading the industry and using my words improperly to do it
As I have stated before, the current simulation systems being used in this industry have value and the visual rendering engines are fantastic from a visual and basic physics POV. However, the gaming architecture and modeling approaches will have model fidelity and real-time issues as they are used for increasingly complex and loaded scenarios. (More in my earlier article below.)
“Physics based”, “Digital Twin” and “Real-time” Simulation Terms can be Misleading
I have been watching the simulation companies in the autonomous vehicle space add to and morph their use of buzzwords for well over a year. In most cases using my words incorrectly to mislead the industry versus fixing their technology. Miraculously these folks discovered “determinism” and even “federation” recently after adopting and ramping up use of “digital twin”, “real-time”, “physics” etc. I have been using most of these words for almost 4 years. I tried to get most of these companies to use the right tech and approach almost 4 years ago, offered to help them do so and they refused. The reasons for that varied from not wanting to admit they took the wrong approach, their customers would have to redo work, they would have to replace what was out there at their cost or carry two baselines. (That last reason was predominantly from OEM simulation companies.) I was actually told by some of them they would fix their systems when their customers figured out the tech was flawed and paid for it. Since that likely will not be until a real-world tragedy occurs, I decided to take this on myself and created Dactle. Having said all of this, as well as what I have been saying about the use of public shadow and “safety driving” over the same period, it should be obvious my issue here is not a competitive one. It is an ethical one. I do not want my words or POV to be hijacked and bastardized by people who have not only have no interest in creating and using the better technology, they have no experience with it. Instead, they want to mislead people into thinking they do and using and getting buried in their inadequate technology. Which will eventually cause significant problems, rework and harm people. I have questioned folks here and there on this over the past year. However, AImotive has gone too far. Six days ago, I wrote the article I posted a link for above. Just prior to that I announced I received a MegaGrant from Epic Unreal to continue our efforts to supply a legitimate “digital twin” that will facilitate FULL L4/5 development. Just today, and quite coincidentally, AImotive posted their LinkedIn entry and YouTube video. I posted a response to those items on LinkedIn today and they deleted it within 5 minutes versus responding to my questions and concerns directly.
Here are several relevant links AImotive links
YouTube video — https://www.youtube.com/watch?v=y_Vste1e35Y&feature=youtu.be
Here are the quotes I questioned and objected to
“Deterministic environment” — That is not a deterministic architecture (Federated models were never mentioned. I expect that any moment now.)
“Precise high-fidelity simulation” — “Every aspect of the virtual world creates a perfect digital twin” — “Exact physics-based modeling” — From YouTube video description “This highly accurate virtual environment makes it possible to properly simulate all major sensor modalities including LiDAR and radar with proven correlation to their real-world counterparts.”
· How “precise”? Exactly what sensor vendor models did you model? How was this proven? Will you supply the performance curves and data to prove this? Did you model several of the same radars in a parking garage with all single and multiple cumulative reflections and interference?
ISO26262 and ASIL-D certified simulator
· What do these men exactly? (Keep in mind I created the SAE ORAD Simulation Task Force to create a J-doc to assist the industry in understanding what levels of modeling and fidelity existed across all relevant industries, what use cases or scenarios they would map to and ways to identify their depth and breadth of fidelity. This based on my time in DoD/aerospace working on systems to a certified FAA Level D. As well as in this industry.)
I want to close with a final comment. Regardless of what any of us says I suggest you verify the depth, breadth, and accuracy of whatever we produce compared to the exact original. Not something close or similar. I guaranty most folks will not produce the data or it will be woefully incomplete. (I provide guidance in my article above.)
More relevant articles
The Autonomous Vehicle Industry can be Saved by doing the Opposite of what is being done now to create this technology
Using the Real World is better than Proper Simulation for Autonomous Vehicle Development — NONSENSE
My name is Michael DeKort — I am a former system engineer, engineering and program manager for Lockheed Martin. I worked in aircraft simulation, the software engineering manager for all of NORAD, the Aegis Weapon System, and on C4ISR for DHS.
Key Industry Participation
- Lead — SAE On-Road Autonomous Driving SAE Model and Simulation Task
- Member SAE ORAD Verification and Validation Task Force
- Stakeholder for UL4600 — Creating AV Safety Guidelines
- Member of the IEEE Artificial Intelligence & Autonomous Systems Policy Committee
- Presented the IEEE Barus Ethics Award for Post 9/11 Efforts
My company is Dactle
We are building an aerospace/DoD/FAA level D, full L4/5 simulation-based testing and AI system with an end-state scenario matrix to address several of the critical issues in the AV/OEM industry I mentioned in my articles below. This includes replacing 99.9% of public shadow and safety driving. As well as dealing with significant real-time, model fidelity and loading/scaling issues caused by using gaming engines and other architectures. (Issues Unity will confirm. We are now working together. We are also working with UAV companies). If not remedied these issues will lead to false confidence and performance differences between what the Plan believes will happen and what actually happens. If someone would like to see a demo or discuss this further please let me know.