Control Design of Robotic Manipulator Based on Quantum Neural Network

Abstract

In this study, a control system was designed to control the robot's movement (The Mitsubishi RM-501 robot manipulator) based on the quantum neural network (QNN). A proposed method was used to solve the inverse kinematics in order to determine the angles values for the arm's joints when it follows through any path. The suggested method is the QNN algorithm. The forward kinematics was derived according to Devavit–Hartenberg representation. The dynamics model for the arm was modeled based on Lagrange method. The dynamic model is considered to be a very important step in the world of robots. In this study, two methods were used to improve the system response. In the first method, the dynamic model was used with the traditional proportional–integral–derivative (PID) controller to find its parameters (Kp, Ki, Kd) by using Ziegler Nichols method. In the second method, the PID parameters were selected depending on QNN without the need to a mathematical model of the robot manipulator. The results show a better response to the system when replacing the traditional PID controller with the suggested controller.