Neural-Network-Based Adaptive Fault-Tolerant Control for Nonlinear Systems: A Fully Actuated System Approach

Abstract

The tracking issue is studied for nonlinear uncertain fully actuated systems in the presence of the actuator’s potential loss of effectiveness fault and bias fault. In contrast to the existing results, this paper takes uncertainties, including totally unknown system dynamics and actuator faults, into consideration. Neural networks are utilized to approximate the unknown dynamics. The adaptive technique is used to update the networks’ weight vector and estimate the unknown bounds of the actuator efficiency factor and bias fault in order to avoid the detrimental effect brought on by uncertainties. Then, a fault-tolerant control method is given to ensure all system’s signals are bound. Finally, a practical example is considered to demonstrate the validity of the main results.