Human-Robot Collaboration Based on Motion Intention Estimation

In this paper, adaptive impedance control is proposed for a robot collaborating with a human partner, in the presence of unknown motion intention of the human partner and unknown robot dynamics. Human motion intention is defined as the desired trajectory in the limb model of the human partner, which is extremely difficult to obtain considering the nonlinear and time-varying property of the limb model. Neural networks are employed to cope with this problem, based on which an online estimation method is developed. The estimated motion intention is integrated into the developed adaptive impedance control, which makes the robot follow a given target impedance model. Under the proposed method, the robot is able to actively collaborate with its human partner, which is verified through experiment studies.