A Neural Network Algorithm Based on Computed –Torque Method for Robot Almega 16 Control

Abstract

This paper presents a problem of building up a control algorithm for the Robot Almega 16 using neural network based on computed torque method. To control the Robot Almega 16, the problems with unknown kinematic and dynamic parameters is solved by implementing neural network based on computed torque method to estimate the parameters and compensation the disturbances from the sample model. The simulation results in the control requirements: steady-state errors of robot joint angles quickly converge to zero and transient time is short.