Top Takeaways from AutoSens in Detroit

Our Detroit edition of AutoSens took place online in May, hosting all the latest industry discussion from global experts and our AutoSens Community. 

In-cabin sensing was a hot topic and it was great to see the progress that is being made there in terms of driver and occupant monitoring and the requirements for sensor fusion inside the car. 

We heard what the next generation in-cabin experience could be from the perspective of Volvo and their focus on the importance of full occupant monitoring. 

Plus, perhaps a new term for some, the promise of neuromorphic sensing was discussed in a panel led by Xperi, exploring how this novel imaging mode can be applied to interior sensing with great effect. 

One of the key questions for the industry continues to be whether Level 3 is possible with just LiDAR and cameras, or if there will need to be a larger suite of sensors, or a different sensor mix. Infra-red sensing continues to offer valuable additional sensor data, but pricing is still a challenge for automotive budgets. 

The long-debated question of how many sensors is enough does not have a clear answer, but this is because it still depends highly on the application. Cameras are increasing capabilities all the time, but we know they have challenges in adverse weather conditions. Perhaps the real question is, ‘Should we expect an autonomous vehicle to drive in ALL conditions?’. With a human driver, sometimes we have to pull over if the conditions are too bad and we accept this. Is it right that we expect robotaxis to just brush these low visibility conditions aside and still deliver us to our destination? 

A perspective that AutoSens has not featured before is that of the trucking industry and Dr. Fridtjof Stein from Daimler Trucks shared his perspective on the different challenges for autonomous trucks and how after his time working on perception for cars, he too is still in search of the right sensor combination. Dr. Stein’s presentation sparked many questions around sensor positioning, maintenance and dealing with harsh conditions. Trucking is an area to watch, and you’ll see more from AutoSens on this topic in the future. 

Alongside several LiDAR developments sessions also covered additional sensor modalities such as gated vision technology, and the use of sound sensors, which can be used in combination with computer vision to reduce false positives for ADAS systems.  

A presentation from Labsphere on advances in LiDAR test targets also generated a lot of interest in terms of the impact the characteristics of targets have on LiDAR range and more. 

Our panel on the state of AI for ADAS provoked discussion around the robustness of AI systems and whether regulations would accept an AI system any time soon, in addition to which NCAP scenarios would benefit from AI. 

We also had insightful panels on Design for Manufacturing with Ford, Valeo and others revealing the key challenges in long term maintenance of sensor systems, and the closing panel discussion with key innovators from the sensor ecosystem explaining more about their relative merits, with some lively technical debate and searching questions from our AutoSens Research host, Khasha Ghaffarzadeh. 

Check out the on-demand schedule for more great insights from our community and expert discussion of the latest technology trends and challenges.