Fault-Tolerant Nonlinear Control of Robotic Manipulator Using Neural Network

This research aims to present a systematic approach to design a robust fault-tolerant nonlinear controller for two-link robotic manipulators to track the desired trajectory in the presence of modeling uncertainties and unknown faults. A conventional nonlinear controller is utilized as a baseline control law to design an intelligent control scheme with fault detection, dynamic estimation, and robust tracking capabilities. An adaptive scheme is developed to estimate the states under modeling uncertainties. The stability analysis of the adaptive scheme is provided to ensure the tracking error convergence and boundedness. In addition, an adaptive threshold is designed to detect the nonlinear fault that occurs in the robotic system dynamics, and an estimation model is developed using an online approximator to estimate the nonlinear fault. The proposed control system is constructed and simulated in MATLAB/SIMULINK to validate the design approach. The simulation results show that the intelligent control system provides accurate tracking performance, maintains the system stability, and accommodates the nonlinear fault under modeling uncertainty.