강화학습을 이용한 휴머노이드 로봇의 안정된 보행

This paper proposes a stable walking for humanoid robot with gyro and acceleration sensor using learning algorithms. In the case of walking on the rough surface, gait is the most critical factor for stable walking. There have been many studies about stable walking. Generally, humanoid walking solution is to use inverse kinematics, dynamics, inverted pendulum and ZMP(Zero Moment Point). ZMP uses FSR sensor in most cases but an amount of data from FSR sensor is not sufficient. The reason is that the response speed of FSR sensor is slow. This paper compares the performance between FSR sensor and gyro and acceleration sensor. Finally, swing value(Shift X) is obtained to adjust data of gyro and acceleration sensor to Q-learning algorithm. We focused on hardware design research for stable walking.