Robust Adaptive Control of Switched Systems

In this thesis, robust adaptive controllers are developed for classes of switched nonlinear systems. Switched systems are those governed by differential equations, which undergo vector field switching due to sudden changes in model characteristics. Such systems arise in many applications such as mechanical systems with contacts, electrical systems with switches, and thermal-fluidic systems with valves and phase changes. The presented controllers guarantee system stability, under typical adaptive control assumptions, for systems with piecewise differentiable bounded parameters and piecewise continuous disturbances without requiring a priori knowledge on such parameters or disturbances. The effect of plant variation and switching is reduced to piecewise continuous and impulsive inputs acting on a Bounded Input Bounded State (BIBS) stable closed loop system. This, in turn, provides a separation between the robust stability and robust performance control problems. The developed methodology provides clear guidelines for steady-sate and transient performance optimization and allows for parameter scheduling and multiple model controller adjustment techniques to be utilized with no stability concerns. The results are illustrated for various systems including contact-based robotic manipulation and Atomic Force Microscope (AFM) based nano-manipulation. Thesis Supervisor: Kamal Youcef-Toumi Title: Professor