Safety for Learning-Based Robot Control

Somil Bansal

Assistant Professor, University of Southern California, Los Angeles, Electrical and Computer Engineering Department & Computer Science Department

Seminar Information

Seminar Series
Dynamic Systems & Controls

Seminar Date - Time
February 17, 2023, 3:00 pm
-
4 PM

Seminar Location
EBUII 479, Von Karman-Penner Seminar Room

Somil Bansal

Abstract

The ability of machine learning techniques to leverage data and process rich sensory inputs (e.g., vision) makes them highly appealing for use in robotic systems. However, the inclusion of learning-based components in the control loop poses an important challenge: how can we guarantee the safety of such systems? For safety, we present a controller synthesis technique based on the computation of reachable sets, using optimal control and game theory. We present new methods for computing the reachable set, based on a functional approximation that has the potential to broadly alleviate its computational complexity and to quickly adapt reachable sets based on online information. In the second part of the talk, we will present a toolbox of methods combining reachable sets with simulation-based methods, to expose safety-critical failures of learning-driven, vision-based controllers.

Speaker Bio

Somil Bansal is an Assistant Professor in the Department of Electrical and Computer Engineering and the Department of Computer Science at the University of Southern California, Los Angeles. He received a Ph.D. in Electrical Engineering and Computer Sciences (EECS) from the University of California at Berkeley in 2020. Before that, he obtained a B.Tech. in Electrical Engineering from the Indian Institute of Technology, Kanpur, and an M.S. in Electrical Engineering and Computer Sciences from UC Berkeley in 2012 and 2014, respectively. Between August 2020 and August 2021, he spent a year as a Research Scientist at Waymo (formerly known as the Google Self-Driving Car project). He has also collaborated closely with companies like Skydio, Google, Waymo, Boeing, as well as NASA Ames. Somil is broadly interested in developing mathematical tools and algorithms for the control and analysis of autonomous systems, with a focus on bridging learning and control-theoretic approaches for safety-critical autonomous systems. Somil has received several awards, most notably the Eli Jury Award at UC Berkeley for his doctoral research, the outstanding graduate student instructor award at UC Berkeley, and the academic excellence award at IIT Kanpur.