Abstract: In this paper, we proposed a novel Hybrid Reinforcement Learning framework to maintain the stability of a biped robot (NAO) while it is walking on static and dynamic platforms. The ...