This paper proposes to use deep reinforcement learning for the simulation of physics-based musculoskeletal models of both healthy subjects and transfemoral prostheses’ users during normal level-ground ...