Gait balance of biped robot based on reinforcement learning

摘要

The study on biped walking control using reinforcement learning is presented in this paper. The Q-learning algorithm makes a robot learn to walk without any previous knowledge of dynamics model. The research topic is mainly focused on how the robot keeps balance with one leg. This balance control way that utilized the motion of robot arm and leg to transfer the Zero Moment Point (ZMP) of the robot would maintain the ZMP in a stable state. Hence, the proposed method which integrated this balanced algorithm with the balance control way applied on biped walking on the plain or seesaw, it makes the biped walk more stable. Finally, there are several simulations that demonstrate the feasibility and effectiveness of the proposed learning scheme.