Enhancing Human-to-Robot Skill Transfer: A Framework Integrating Movement and Variable Impedance Based on EMG

Skill transfer and impedance control is gaining increasing prominence as a solution to interactive tasks which are challenged through pre-programming. In this paper, a framework for skill transfer that integrates both trajectory and variable impedance skill transfer is proposed. Firstly, EMG signals are retrieved from the human operator while demonstrating the task via teleoperation. A method is introduced to establish the relationship between the human operator's arm stiffness and the robot's stiffness for both contact and non-contact tasks. Then, the dynamic movement primitives are employed to simultaneously encode the trajectories and stiffness profiles. Finally, the skill transfer method proposed in this paper is validate through simulation experiment with a UR5 robot.