Actuator Fault Estimation Based On Adaptive Iterative Learning Algorithm For Nonlinear Periodic Systems

A actuator fault estimation algorithm based on adaptive iterative learning is proposed for a class of nonlinear periodic systems with actuator fault and disturbances. Firstly, the proposed method designs a state observer and a fault estimator based on adaptive iterative learning. The Lyapunov stability theory and the optimization theory are employed to prove the convergence of the method. Then the LMI toolbox is utilized to obtain the results of the observer gain matrix and fault estimator parameter matrix. The method combines the iterative learning strategy with the adaptive algorithm, and utilizes not only the state error signals of the last iteration learning, but also the output error signals of this time. It makes the estimated fault approximate to the real fault gradually and realizes the accurate estimation of the time-varying fault and the intermittent fault. Finally, the validity and accuracy of the proposed method are verified by the simulation of a single-link robotic arm mold.