A Robust Adaptive Cancellation of Unknown Sinusoidal Disturbance

Abstract

A robust adaptive controller that consists of a well-known PI controller and a robust frequency estimation in parallel was presented to cancel unknown sinusoidal disturbance in a stable linear time invariant system. The convergence of frequency estimation, the accurate cancellation of disturbance and the asymptotic stability of the overall feedback control system are analyzed by slow integral manifold and averaging method. The key advantages include asymptotically convergent cancellation for disturbance with much low frequency, the asymptotic stability of the whole error feedback system independent to the parameters of the plant and robust applicability in engineering. Simulation results validate the transient convergence performance and the steady-state anti-noise performance as well as the action of parameters of the proposed algorithm.