A dynamic feedback neural model for identification of the robot manipulator

Abstract

Robot manipulators are very powerful industrial systems. These systems are used in many different industrial applications. Since the derivation of the mathematical model of a robot manipulator has complex processing load, a suitable neural model can be designed. In this paper, both inverse and forward kinematics equations of a six degree of freedom (DoF) robot manipulator are given and also adapted to MATLAB environment with a graphical user interface to identify behaviors of the robot manipulator. At the same time, a multilayer artificial neural model is proposed to provide the robot manipulator identification. Back-propagation learning algorithm is used to train dynamic feedback neural model based on NARX network structure. The experimental results are presented to show the performance of the dynamic feedback neural model.