Controlling a 6DOF Serial Manipulator Using Parallel Robot by ANN

Abstract

Nowadays, robotic arms have many uses. Among the equations for robotic arms, we can consider direct kinematics and inverse kinematics as the two most important equations. Direct kinematics equation converts the angles of the joints to end effector position and orientation, whereas inverse kinematics converts end effector position and orientation to the angles of the joints. Inverse kinematic equations are very difficult to solve for particular robots and also parallel robots. However, these equations can be easily solved using neural networks. In this research, we intend to control an articular arm via a parallel joystick or robotic arm. In our proposed method, this is achieved by solving both the direct and inverse kinematic equations mathematically, and the neural network is not used to connect these two arms to each other. Even if one arm intends to imitate another arm, the control of a serial arm by a parallel one is very complicated. So, in this paper we investigate the control of the serial arm using neural networks without knowing the final position of the parallel arm. In this investigation, we implement the most difficult scenario, i.e. controlling a serial arm using a parallel control handle and finally check the results.