MAE Seminar: Safe and Interactive Autonomy - Learning and Control for Multi-Agent Interactions

Zoom Link: https://uci.zoom.us/j/95085392231

Abstract: To transform our lives, robots need to interact with other agents in complex shared environments. For example, autonomous cars need to interact with pedestrians, human-driven cars and other autonomous cars. Autonomous delivery drones need to navigate in the aerial space shared by other drones, or mobile robots in a warehouse must navigate in the factory space shared by robots and humans. The goal of my research is to develop control algorithms that enable safe and efficient robotic interactions in such multi-agent settings.

In this talk, I will first focus on interactive planning for robots. To reach intelligent robotic interactions, robots must account for the dependence of agents' decisions upon one another. I will discuss how game-theoretic planning and control enables robots to be cognizant of their influence on other agents. I will then talk about safety of robots. I will discuss how risk-aware planning methods can ensure safe and efficient interactions in the presence of uncertainty. Finally, I will talk about ensuring safety of robots when learning from offline data. I’ll present an algorithm for ensuring convergent behavior of a robot when its policy is learned from offline data by leveraging theoretical guarantees from contraction theory.

Bio: Negar Mehr is an assistant professor in aerospace engineering at the University of Illinois Urbana-Champaign. She is also affiliated with the Coordinated Science Laboratory and Department of Electrical and Computer Engineering at UIUC. Previously, she was a postdoctoral scholar at Stanford University's Department of Aeronautics and Astronautics from 2019 to 2020. She received her doctorate in mechanical engineering at UC Berkeley in 2019 and received her bachelor's degree in mechanical engineering from Sharif University of Technology, Tehran, Iran, in 2013. Her research interests lie at the intersection of control theory, robotics and game theory. Specifically, she is interested in developing control algorithms that enable safe and intelligent multi-agent interactions. Mehr was recently awarded the NSF CAREER award. She was the recipient of the IEEE ITSS best Ph.D. dissertation award in 2020. She won the best student paper award at the International Conference on Intelligent Transportation Systems, 2016.