Adaptive trajectory modeling of humanoid robot 3-DOF arm using inverse neural MIMO NARX model

Abstract

In this paper, a novel inverse adaptive neural MIMO NARX model is used for modeling and identifying the inverse kinematics of the humanoid robot 3-DOF arm system. The nonlinear features of the inverse kinematics of the industrial robot arm drive are thoroughly modeled based on the inverse adaptive neural NARX model-based identification process using experimental input-output training data. This paper proposes the novel use of a back propagation (BP) algorithm to generate the inverse neural MIMO NARX (INMN) model for the inverse kinematics of the humanoid robot 3-DOF arm. The results show that the proposed adaptive neural NARX model trained by Back Propagation learning algorithm yields outstanding performance and perfect accuracy.