Iterative-learning-based actuator/sensor fault estimation for a class of repetitive nonlinear systems with time-delay

To meet the demand for estimating actuator and sensor faults simultaneously in a class of repetitive nonlinear time-delay systems, this paper proposes a novel fault estimation strategy based on an iterative learning scheme. Firstly, an iterative-learning-based fault estimation law is designed to estimate actuator faults while system is free of sensor failures. Both the fixed initial shift and random one are taken into consideration. Secondly, a novel sensor fault observer is proposed based on an augmented state variable which consists of original system state and sensor fault signal; output compensation strategy is also provided to ensure the iterative-learning-based actuator fault estimation method is effective considering the existence of sensor failures. In addition, theorems based on lambda$$ \lambda $$-norm and linear matrix inequality are provided to determine values or ranges of gain matrices and parameters in proposed sensor fault observer and iterative-learning-based actuator fault estimation law. Finally, two simulation examples are provided to illustrate the effectiveness of the proposed methods.