Online teaching of robotic arm by human–robot interaction: end effector force/torque sensing

Abstract

Human–robot interaction HRI is one of the most important research areas in robotics. A novel approach for HRI on a robotic arm is proposed by using online teaching to eliminate the effect of tool inertia. The position error is reduced in repeated motion. A multi-axis F/T sensor is attached to Denso robotic arm to measure six components of force and torque. A new controller structure is introduced by modifying the virtual spring control with tool inertia effect compensation. The human hand force and torque are transformed to the desired position/orientation (P/O) through the instantaneous matching between the direct human guidance and the robot response. The motions according to the experimental results have been compared with different teaching control variables. The results are shown significant improvement in teaching performance. The repeated motion errors are obviously reduced.