ECE Seminar

Details

Speaker: Calin Belta, University of Maryland

Title: Formal Methods for Safe and Interpretable Control

Abstract:
In control theory, complicated dynamics such as systems of (nonlinear) differential equations are mostly controlled to achieve stability and to optimize a cost. In formal methods, simple systems such as finite state transition graphs modeling computer programs or digital circuits are analyzed or controlled from specifications such as safety, liveness, or richer requirements expressed as formulas of temporal logics. Many current applications, such as dexterous robotic manipulation,
involve high-dimensional and partially known dynamical systems, which require machine learning techniques for motion planning and control. Ensuring safety and incorporating specifications given in rich, natural language within a learning-based system are challenging problems that received a lot of attention recently.
In this talk, I will show how techniques based on control barrier and Lyapunov functions can be combined with temporal logics and reinforcement learning to address these challenges. I will use examples from robotic manipulation and autonomous driving.