Events

ECE Seminar: AI Agents in the Physical World: Balancing Safety & Performance

Speaker: Rahul Mangharam

Location: 370 Jay Street, Room 825
Videoconference link: https://nyu.zoom.us/j/93976459621

Date: Thursday, November 21, 2024

Balancing safety and performance is crucial to deploying autonomous agents in physical environments. In particular, autonomous racing is a domain that penalizes safe but conservative policies, highlighting the need for robust, adaptive strategies. Current approaches either make simplifying assumptions about other agents or lack robust mechanisms for online adaptation. In this talk we explore the following research themes on learning-based perception, planning and control at the limits of performance:

(1) How to generate the most competitive agents who dynamically balance safety and assertiveness by using distributionally robust online adaptation and Game-theoretic planning
(2) How to be better-than-the-best using imitation learning with multiple imperfect experts
(3) Using invertible neural networks to solve inverse problems in localization and SLAM
(4) How to build the most efficient agents with multi-domain optimization across mechanical, decision and control designs;
We realize all our research in the https://f1tenth.org autonomous racecar platform that is 10th the size, but 10x the fun! In the Autoware Center of Excellence for Autonomous Driving we develop several EV and AV reference platforms