An adaptive human-robot system using CMAC and over damping

Abstract

The aim of this paper is to propose an adaptive human-robot system to control a robot by the operator's hand. In the proposed system, a leap motion (LM) is used to track the motions of the operator's hand so that the robot can be controlled after the hand motions are translated as robot control commands. However, the measured motions of the operator's hand obtained by the leap motion sensor have inherent error and tracking noise, which results in inaccurate robot manipulation. Thus, two cerebellar model articulation controllers (CMACs) are applied to process the unstable position and orientation signals so as to filter the influence of the measurement error, respectively. What's more, instable factors such as the operator's hand trembling may occur during the manipulation, which leads to the motions of the robot manipulator may change greatly. Therefore, an over damping method (ODM) is introduced to eliminate the influence of error signals, so that can improve the accuracy and stability of the robot manipulation. The proposed system has been used for related experiment in a lab, and the results has been verified the effectiveness of the proposed system.