Robust Adaptive Neural Network Finite-Time Tracking Control for Robotic Manipulators Without Velocity Measurements

Abstract

This paper proposed a robust finite-time tracking controller with adaptive neural networks for uncertain robotic manipulators without velocity measurements. A simple output feedback controller plus a nonlinear filter is designed to achieve satisfied performance, such as high accuracy, and fast response, which is more convenient and lower cost for robotic manipulators in practice. The adaptive neural networks with finite-time convergence are designed to compensate the uncertainties, which effectively further improve the robustness. The Lyapunov stability theory and geometric homogeneity technique are employed to prove the practical finite-time stability of the whole closed-loop system. Simulations on two-degree robotic manipulators show the effectiveness and robustness of the proposed control strategy.