[CPS-IoT Week Workshop] Provably Safe Perception in Cyber-Physical Systems

Deep Samal, Dung Tran, and Marilyn Wolf are organizing a new workshop at CPS-IoT Week this year: the International Workshop on Perception for Safety-Critical Cyber-Physical Systems (PerCPS ’23).  This workshop is intended to provide a bridge between machine learning for perception and cyber-physical systems.

Perception is a critical capability for autonomous cyber-physical systems.  Cyber-physical automatons need to be able to sense and understand their environment in order to be able to perform their functions.  Perception may require multiple modalities: vision, sound, touch, smell.  Vision itself may be necessary in multiple bands: visible, infrared, radar. Given the complexity of perception and the challenges inherent in understanding the natural world, the fact that perception is a major area in machine learning is not surprising.

Perception should not be treated as a separate topic from cyber-physical systems.  The dynamics of the perception system affect the dynamics of the overall cyber-physical systems—delay is a critical parameter in the design of machine perception.  Safety properties of cyber-physical systems are similarly intertwined with the properties of their perception systems. Perception needs to be integrated into CPS formal methods. The accuracy of a perception system should be evaluated within the context of the vehicle and control system dynamics.  The need to fuse multiple perception modalities should be considered in the context of planning and control algorithms. Computation places a significant load on the autonomy’s computational platform and that load needs to be managed to satisfy real-time and low-power requirements.  All of these efforts need to be supported by data sets and benchmarks.

The call for papers for PerCPS ’23 is here: https://sites.google.com/view/percps/. The submission deadline is February 28.  Please consider submitting your work and attending in San Antonio.

DisclaimerAny views or opinions represented in this blog are personal, belong solely to the blog post authors and do not represent those of ACM SIGBED or its parent organization, ACM.