Reinforcement Learning endowed with safe veto policies to learn the control of Linked-Multicomponent Robotic Systems | Publicación