Experimental Evaluation Of Intuitive Programming Of Robot Interaction Behaviour During Kinesthetic Teaching Using sEMG And Cutaneous Feedback

Modern applications related to service or production can nowadays benefit from the introduction of collaborative robots already available on the market and endowed with several advanced features such as precise torque control or safe physical interaction with operators. More importantly, collaborative robots allow operators to teach end-effector trajectories by means of physical interaction – known as kinesthetic teaching – which is one of the most intuitive programming-by-demonstration techniques. However, important functionalities provided by modern collaborative robots, like the possibility of performing smooth interactions, cannot be programmed intuitively with the available framework of kinesthetic teaching. In the present study, we propose and experimentally evaluate a robot programming framework for the simultaneous teaching of both trajectories by means of kinesthetic teaching, and robot interaction behavior by means of impedance shaping along the trajectory exploiting a wearable interface. Specifically, the wearable interface is designed to not affect the free motion of the operator, necessary to perform kinesthetic teaching, and it is based on the usage of surface electromyography (sEMG) and vibrotactile stimulation. In this way, we propose an intuitive robot programming framework for an offline robot trajectory and interaction behavior programming, according to which the operator will be able to plan interactions with the environment and humans. In this article, we report a preliminary experimental evaluation of the proposed system, in which an operator will teach a 7-degrees-of-freedom manipulator the execution of a task on a robotic wiring test-bed. In the experiment, the programming of requested compliance levels during the kinesthetic teaching of a trajectory is performed, and the reported results show that the provided wearable interface is successfully exploited by the operator. Finally, the experiment also demonstrates that the offline intuitive programming of trajectories and impedance levels can be exploited online for human-robot co-work.