Capturing synchronized vision, sEMG, and tactile data from real human work to train robot manipulation policies.
