ILO-based Actuator and Sensor Faults Estimation for Nonlinear Discrete-time Systems

This paper addresses the multiple faults estimation problem of actuator and sensor faults with using iterative learning observer (ILO) for nonlinear discrete-time systems. First, an augmented descriptor system including system state and sensor fault vectors is constructed to achieve multiple faults estimation purpose. Then, a novel iterative learning observer is proposed, and the nonlinear estimation error is deduced into a more compact system by using a reformulated Lipschitz property. Next, design conditions of the observers are derived into the LMI optimization problem where a slack variable and L
∞
. performance index is introduced to reduce the conservativeness and obtain a more excellent performance of fault estimation. Finally, numerical simulations of an example of a single-link robot arm are provided to illustrate the effectiveness of the proposed method and multiple fault estimation scheme.