Robust model reference adaptive controller design

Model reference adaptive controllers with different adaptation mechanisms or structures have been developed in this thesis study. Control signal and performance of these controllers were investigated. Standard model reference adaptive controller (MRAC), MRAC with PI adaptation mechanism (MRAC-PI-AM), MRAC with PIFD adaptation mechanism (MRAC-PIFD-AM), MRAC with error compensation (MRAC-EC), MRAC with closed reference model (CRM-MRAC), MRAC with fractional integral adaptation mechanism (MRAC-FOI-AM) and MRAC with fractional PI adaptation mechanism (MRAC-FOPI-AM) methods were used in this study and their stability is shown via candidate Lyapunov functions. The stability of the MRAC-FOI-AM structure according to different approaches has been analyzed using candidate Lyapunov functions according to the transfer functions. The approximations discussed and compared here are the Oustaloup approximation and the refined Oustaloup approximation. In order to increase the robustness of MRAC-PI-AM and MRAC-PIFD-AM structures and to solve the parameter drift problem, their sigma modified versions have been developed and their stability have been proved over candidate Lyapunov functions. The performances and control signals of MRAC structures are compared for both linear systems and nonlinear wing rock systems. How these given structures affect the control signal is examined and their resistance to time delay is demonstrated through simulation studies. The required MRAC parameters for systems with time delay in the input signal were obtained using teaching-learning based optimization(TLBO). Finally, a new objective function is designed to reduce the oscillation of the control signal in MRAC for a system with time delay in the input signal.