Auto.AI 2022
Auto.AI 2022
The conference was organized in a very interesting way, as there were two conferences (dual-cpu) running at the same time at the adjacent rooms and once in on the third day of the conference the wall went between two halls somehow was dismounted merging the two audiences and spaces for a joint session. There were a lot of engaging talks and conversation on autonomous cars and trucks and a lot of sincere concern raised towards safety.
First day was opened by CEO of May Mobility who spoke about applying multi-policy decision making process (sort of gaming theory in my understanding) for the ADV. I still didn’t understand what is the advantage of this process vs. conventional and more than that I wonder if we can call any deterministic process a conventional process for ADV to steer, accelerate and brake.
Quite interesting talk from Plus.AI on the challenges and lessons learned from autonomous trucking
Oculi acquired by Ambarella was featuring a high resolution radar which processing unit was is integrated into into the CV board. CV board developed by Ambarella consists of multi-core ISP and CV-flow engine and featuring very high resource utilization scheduler/compiler combo for NN-deployment.
Networking
During networking session I got a chance to hear about the progress of truck platooning with Mike Bucala from Daimler. It was interesting to hear from the audience about scary experience of riding in a truck platooning which is going not in the head of the train but somewhere in the middle.
Tony Gioutsos from Siemens team told about sensor simulation at the wave level.
Its was pleasant to find about about a team from NY who opened a labeling business in Africa.
Mark Vernacchia from General Motors was a great facilitator of controversial conversation, really enjoyed how much value he was able to pull out during the 30 minutes.
Five 30-minute sessions with industry gurus
One very inspiring session was organized in unusual way: 5 industry experts were given different topics to discuss and each attendee was given a sequence of numbers representing in which order he or she would attend a session with each of the experts. After each session ended one should change the group and discuss next topic where the previous group has left it. It was great organization from the Auto.AI crew and example of very productive collaboration.
One great quote about transportation:
