MINES ParisTech CAS - Centre automatique et systèmes

Game-Theoretic Modeling of Driver/Vehicle Interactions in Traffic

Thursday 14th January 2021, 4pm – 5pm.

Anouck Girard, Aerospace Engineering, University of Michigan, USA, https://aero.engin.umich.edu/people/anouck-girard/

Game-Theoretic Modeling of Driver/Vehicle Interactions in Traffic

Industrial, military, surgical, and housekeeping robots are commonplace. Self-driving cars are currently being tested on shared roadways. The promise of enhanced autonomy is that it will allow robots to become more capable, benefiting society, performing dull, dirty and dangerous missions, collecting scientific data, helping the elderly and the disabled. The danger is that certifying the performance of such systems is not a solved problem.
This talk will focus on verification and validation (V&V) for the improved capabilities brought about by increasing autonomy. V&V capability is important to guarantee safety and efficiency of the proposed robot schemes. It is estimated, for example, that certifying the safety of a self-driving car may take 1,000,000 miles of driving. In the face of such numbers, our focus will be on simulation-based V&V; if a significant portion of the million miles can be certified in simulation, this will greatly reduce the costs associated with certification. A new game-theoretic simulator that models fully interactive traffic with 30-50 cars has been made available online. We will discuss game-theoretic traffic models, interactions in highway and urban/intersections scenarios, and applications to autonomous vehicle control.

Anouck Girard (PhD, Ocean Engineering, University of California at Berkeley, 2002) is an associate professor of Aerospace Engineering at the University of Michigan, Ann Arbor. She is a vehicle dynamics and control engineer, whose interests lie in controlling advanced and increasingly autonomous vehicles operating in the space, air, ground and marine domains. These vehicles and vehicle systems exhibit complex nonlinear dynamics, and must function in uncertain environments with limited resources, while satisfying stringent constraints and counteracting the effects of disturbances. Mostly, she addresses problems of human-machine interaction, optimal decision-making, energy and information-aware trajectory optimization, and control of unusual vehicle configurations. She has co-authored the book Fundamentals of Aerospace Navigation and Guidance (Cambridge University Press, 2014). Dr. Girard has taught aircraft and spacecraft dynamics and control courses at Michigan, and has co-authored more than 200 papers on vehicle dynamics and control.

See Slides