Comparison of Adaptation Mechanisms on MRAC

Model reference adaptive control (MRAC) could be designed using different adaptation mechanisms. An adaptation mechanism could affect the Model Reference Adaptive System (MRAS) performance. High adaptive gain increases the transient response performance while decreasing the robustness of the MRAS even under perfect conditions (no time delay, without unmatched uncertainty and absence of disturbance) and increasing control signal oscillation in classical MRAC. Classical MRAS do not perform satisfactorily on such time-delayed adaptive signals. This study presents and compares two different adaptation mechanisms in MRAS, which are the low-frequency learning method on MRAC and MRAC with PI adjustment mechanism. We have proposed a modification to the low-frequency learning method on MRAC to get a transient response as good as MRAC with a PI adjustment mechanism on the time delay system. The best performance is obtained from MRAS with PI adaptation mechanism.