The manuscript presents an overview of my recent work on realization theory for complex cyber-physical systems and some of applications of realization theory to system identification and model reduction. In my research, I focused on classes of dynamical systems (hyrid systems, linear parameter- varying systems, bilinear systems) which can be described by piecewise-linear equations. While simple, these systems can be used to model complex nonlinear dynamics. In my work I explored the relationship between observed behavior of such systems and their internal structure (realization theory). I then used the obtained results for learning unknown parameters of such systems (system identification) , and simplifying the system equations (model reduction).
defended on 26/01/2023