Abstract: Reinforcement learning (RL) is capable of sophisticated motion planning and control for robots in uncertain environments. However, state-of-the-art deep RL approaches typically lack safety ...
Some results have been hidden because they may be inaccessible to you
Show inaccessible results