Iterative fault estimation and fault‐tolerant control for a class of nonlinear variant time‐delay systems

Abstract

Fault estimation (FE) and fault-tolerant control (FTC) are remarkable techniques, have achieved great success in many applications such as robot, spacecraft, and industrial assembly line. This article aims to design an iterative-learning scheme based FE and FTC method for a class of nonlinear system with iteration-variant state delay and additive measurement noise. A Luenberger observer in iterative version is proposed to achieve the reconstruction of system state information, which consider the historical observation error in order to improve the observation performance in current iteration. To deal with bounded iteration-variant state delay, an iterative-learning scheme based fault estimator is designed and the convergence is proved. Compared with relevant methods which use system output observation residual to revise the FE result of last iteration, the proposed approach uses filtered system output observation residual in order to reduce the effect of measurement noise. Based on the FE result, FTC using signal compensation technique is employed. In addition, an improved particle swarm optimization algorithm is employed for parameters adaptive tuning. Compared with traditional manual adjustment of parameters, the proposed method can find the optimal parameters and save time of parameter tuning. Finally, three examples are provided to verify the effectiveness of the proposed approach.

Fault‐tolerant control design for stochastic distribution control systems based on k‐step fault estimation

“…The accuracy of FE was improved by designing the input of the observer. In Li et al [41], the FE information obtained from the observer was optimized using the iterative learning algorithm. Although the accuracy of FE can be improved by using the iterative learning algorithm, the computational burden and the number of design parameters are also increased.…”