This is a hybrid event with in-person attendance in Levine 307 and virtual attendance via Zoom. This week’s presenter will be in-person as well.

ABSTRACT

A key challenge to the widespread deployment of robotic manipulators is the need to ensure safety in arbitrary environments while generating new motion plans in real-time. This talk describes a technique that constructs a parameterized representation of the forward reachable set that it then uses in concert with predictions to enable certified, collision checking. To improve computational speed, this talk describes how to represent this parameterized reachable set using a neural implicit representation without sacrificing any safety guarantees. This approach, which is guaranteed to generate safe behavior, is demonstrated across a variety of different real-world platforms including ground vehicles, manipulators, and walking robots.