Neural learning enhanced teleoperation control of Baxter robot using IMU based Motion Capture

Abstract

In this paper, we have developed a neural network (NN) control enhanced teleoperation strategy which has been implemented on the Baxter robot. The upper limb motion of the human operator is captured by the inertial measurement unit (IMU) embedded in a pair of MYO armbands which are worn on the operator's forearm and upper arm, respectively. They are used to detect and to reconstruct the physical motion of shoulder and elbow joints of the operator. Given human operator's motion as reference trajectories, the robot is controlled using NN technique to compensate for its unknown dynamics. Adaptive law has been synthesized based on Lyapunov theory to enable effective NN learning. Preliminary experiments have been carried out to test the proposed method, which results in satisfactory performance on the Baxter robot teleoperation.