Data-driven control for systems with constraints

My core research has been focused on hybrid control systems which evolve in continuous- and discrete-time according to (generalized) nonlinear differential and difference equations constrained on sets. I have been interested in using advanced hybrid tools for engineering applications where hybrid phenomena play an essential role and the desired properties can be guaranteed by Lyapunov(-like) methods. These applications include satisfaction of linear temporal logic, obstacle avoidance, and friction compensation.

My research here will be focused on data-driven control, that is, a finite amount of data collected from a dynamical system is used to directly design control laws, and carries sufficient information to give guarantees for the properties enforced by these control laws. By taking directly into account the actual system behaviour, data-driven control can provide a less conservative design in the presence of complex dynamics (e.g., nonlinear and high-dimensional) and complex environments (e.g., with constraints and partially unknown), which are typical scenarios in robotics and mechatronics applications.