Adaptive robust non-singular fast terminal sliding mode control for a two-link robotic manipulator with NN-based ESO

Abstract

A novel adaptive non-singular fast terminal sliding mode control (NFTSMC) scheme is developed for a highly nonlinear two-link robotic manipulator (TLRM) system subject to system uncertainties and unknown disturbances. First, to compensate for the system uncertainties, neural network (NN) is employed to approximate the unknown nonlinear dynamic function. Subsequently, to address the challenges of disturbance compensation and unmeasured joint angular velocity, a nonlinear NN-based extended state observer (NN-based ESO) is introduced. The NN-based ESO facilitates accurate disturbance estimation and compensation, while also enabling simultaneous estimation of the joint angular velocity. Through the integration of these techniques, the closed-loop system’s stability is proven using the direct Lyapunov method, demonstrating remarkable tracking performance within finite time. Simulation results validate the effectiveness of the NFTSMC with the NN-based ESO, showcasing superior performance compared to alternative control strategies.