Physical Intelligence on the Edge: A Vision for the Decade Ahead

Abstract

Join us in exploring the hidden potential and added value of Physical Intelligence on the Edge.

Physical Intelligence on the Edge (PIE) is an emerging paradigm in which raw sensor streams are transformed into real-time, safety-critical intelligence that can act in the physical world. This talk traces the evolution from centralized to distributed and edge architectures and argues that PIE marks a qualitative shift: the edge becomes the primary platform for tightly integrating sensing, reasoning, and actuation under stringent real-time constraints. Dr. Shi will introduce five research thrusts—embodied spatial reasoning, embodied temporal reasoning, edge-native customization, symbiosis, and sustainability—and use a hypothetical PIE scenario to highlight a core gap in today’s systems: while current edge platforms can run components of perception and inference, they still cannot autonomously close the sense-think-act loop with certifiable guarantees on timing and safety. Finally, he will talk about the recent development of D-STAR, a live and evolving testbed for vehicle computing on the STAR campus at the University of Delaware.