A novel reduced-order model reference adaptive control approach with high-frequency oscillation suppression

High-frequency oscillations may occur in the control signal when the adaptive gain is too large, which will adversely affect the robustness of the model reference adaptive control (MRAC) law and potentially induce system instability. To tackle this challenge, we introduce a reduced-order MRAC approach based on a modified scalar update law that includes a modification term to filter out high-frequency content for the multi-input multi-output (MIMO) nonlinear dynamical systems. This modification significantly reduces the high-frequency oscillations, allowing a controller with a high adaptive gain to achieve fast adaptation. In addition, with this reduced-order adaptive control approach, only two parameters need to be updated online, thus substantially reducing the overall computation burden. Furthermore, we demonstrate that the proposed reduced-order MRAC approach guarantees the Lyapunov asymptotic stability of the system error dynamics. Simulation results are conducted to validate the electiveness of the proposed method.