Humanoid Robot Gait Planning And Control Research

Humanoid robot has made great progress in recent years. It represents the world’s most advanced automatic control technology. The humanoid robot receives high attention because of its humanoid structure and function. As the humanoid robot has the characteristics of nonlinear, high dimension and strong coupling, the research of humanoid robot walking stability is difficult. How to deal with the interference is also a popular research area of humanoid.This thesis focuses on the walking gait of humanoid robot. First of all, the robot walking gait is planed based on linear inverted pendulum model. Based on the ZMP stability criterion, the walking gait in the forward plane and lateral plane is planned. Arm swinging is added to improve the stability of robot walking gait. After that, the gait of humanoid robot NAO is planned. This planning method uses three function curve smoothing and plans the trajectory of ZMP and centroid, which avoids the problem of ZMP trajectory connection. A ZMP gait compensation algorithm is presented, which is based on the table-car model. Feedback controller is designed for dynamic compensation of walking gait, which improves the stability of the walking robot.According to the problem of robot’s static balance, this thesis studies three balance strategies: respectively, hip joint strategy, ankle joint strategy and step strategy. These balance strategies can make robot resist the interference of external environment in the process of walking.