Stitching Virtual and Physical for AV Sensors Design, Verification and Validation


About this session

The promise of zero crashes, added to the business potential of shared autonomous mobility, is fundamentally changing how auto industry approaches systems and vehicle engineering. Verification and validation (V&V) of the autonomous driving system is central to every company’s efforts. Companies are investing heavily in simulations, semi-virtual testing such as shadow mode as well as real world testing with AV fleets in various cities around the world. However, significant challenges hinder the progress. How can auto OEMs test robustness of autonomous driving when the sensors ecosystem and supplier base is evolving rapidly? How can one validate sensor performance simulation enough to rely on those results? How can OEMs be sure of autonomous driving robustness in various cities that have significantly different city infrastructure and driving behaviors? Industry needs better collaboration between virtual (simulation) and real-world testing. 

In this talk, we will look at how this collaboration between virtual and physical can be established or can be made better. We’ll offer our perspective on research needs for the next 5 years, need for multifaceted collaboration and will showcase Siemens ongoing efforts in connecting sensors hardware design to data collection and physics-based performance modeling to city deployment in any traffic and weather condition.

This talk is preceded by opening remarks from Rob Stead and Patrick Denny.

Responses