January 31, 2024

David Fridovich-Keil, an assistant professor in the Department of Aerospace Engineering and Engineering Mechanics at The University of Texas at Austin, was selected to receive a National Science Foundation (NSF) Faculty Early Career Development Program (CAREER) award for 2024.

Fridovich-Keil was awarded for his proposal, “Game Theoretic Models for Robust Cyber-Physical Interactions: Inference and Design under Uncertainty.” He will use the award over the next five years to develop flexible modeling frameworks and efficient algorithms for cyber-physical systems – autonomous systems that interact with people, the environment, and other similar systems.

Almost everyone interacts with some type of cyber-physical system (CPS) and often, even several times a day. Examples of a CPS include air and road traffic systems, urban mobility networks and guidance systems, like Google maps. Currently, many of these systems are being designed and analyzed separately from human interaction and the environment. This can create unanticipated consequences that weren’t originally planned for in the CPS.

For example, a person might be directed through a quiet neighborhood by a map guidance system, creating traffic congestion in unanticipated or unsafe locations. Or worse, a data-spoofing attack might lead to rerouting traffic in the wrong direction. 

highway traffic game theory model graphic
This graphic represents the performance of Fridovich-Keil's recent game-theoretic objective estimation scheme in a five-care highway driving scenario.

Fridovich-Keil, a game theory expert, aims to improve models like these by studying current traffic systems and identifying their weaknesses due to the limited knowledge of how players, in this case, humans, interact with the CPS. Ultimately, the goal is to develop new and improved models that use long-term strategies by predicting the interactions and behavior of the players, instead of only reacting to a problem after it has arisen.

One example is to develop a tolling system that is designed to work around traffic flow during different times of day, which could help avoid congestion during rush hour.

“We want to model the way people will interact with these systems like they are playing a game,” said Fridovich-Keil. “For example, with tollways, we can set the tolls one way in the morning, and then change them in the early afternoon to preemptively clear the routes we think people are going to take. It’s very similar to planning ahead several moves in the game of chess.”

The overall goal of this research is to answer fundamental questions in dynamic game theory within cyber-physical transportation systems. The algorithms and theoretic models developed by Fridovich-Keil and his team will provide fundamental building blocks for the future development of transportation models and other cyber-physical applications.

More broadly, Fridovich-Keil said that his team’s efforts will provide tools that can be used in urban planning and regulation to help design new and improved transit systems. This can help reduce greenhouse gas emissions, improve road safety and minimize time wasted due to traffic congestion and flight delays.

Fridovich-Keil joined the department in fall of 2021. His research spans optimal control, dynamic game theory, learning for control and robot safety. He is currently investigating the role of dynamic game theory in multi-agent interactive settings such as traffic. Fridovich-Keil’s work also focuses on the interplay between machine learning and classical ideas from robust, adaptive, and geometric control theory. In addition, he has worked on problems in distributed control, reinforcement learning and active search. Learn more about his research on his website.

The NSF CAREER award is among the most prestigious offered to junior faculty, providing up to five years of funding to those who exemplify the role of teacher-scholars through outstanding research, excellent education and the integration of education and research within the context of their organizations' missions.