Visual servo control with artificial neural network

Abstract

In this paper a robotic manipulator system is presented, which uses visual feedback to control both the position and orientation of the manipulator. In this system, visual data from the camera are directly used in the servo control process of the manipulator, without calculations of the inverse kinematics. This approach is achieved by the use of an artificial neural network (ANN) to learn the relationship between the position and the orientation of the manipulator and the joint angles. As there are no pre-stored parameters of the manipulator, the system is adaptable to slight changes in the kinematics of the manipulator or the working environment. Hence, this technique should able to extend the application of robotic technology in some special industrial areas. The results of computer simulations of this system are also presented in the paper.