Human Motion Mapping to a Robot Arm with Redundancy Resolution

Abstract

In this chapter the problem of mapping human arm motion to an anthropomorphic robot arm has been addressed using an Xsens MVN motion capture suite and a 7-DoF KUKA LWR. The desired end-effector trajectories of the robot are reconstructed from the human hand, forearm and upper arm trajectories in the Cartesian space obtained from the motion tracking system by means of human arm biomechanical models and sensor fusion algorithms embedded in the Xsens Technology. The desired pose of the robot is reconstructed taking into account the differences between the robot and human arm kinematics and is obtained by suitably scaling the human arm link dimensions. A Cartesian impedance control is designed to replicate, at the robot side, the human wrist motion and a compliant null-space control strategy is applied to solve kinematic redundancy exploiting the compliant behavior of the elbow to obtain suitable body reconfigurations.