Research On Humanoid Robot Control Based On Inverted Pendulum Model

The stable walking of the humanoid robot is the basis for assisting humans or independently completing the work,so it has also become a topic that has received much attention and focus.The humanoid robot's movement is similar to human walking.It is very important to expand the scope of human activities and accomplish some high-risk tasks.At present,the research on the stable walking of humanoid robots at home and abroad is very extensive,and the inverted pendulum model is one of the more commonly used models in the gait research of humanoid robots.The preview control adjusts the current centroid trajectory according to the information of the future ZMP(zero moment point).The mathematical model is simple and the trajectory is smooth.However,the research on the inverted pendulum model combined with the preview control has the problem of insufficient ZMP tracking accuracy.In this paper,combined with the theory of predictive control on the inverted pendulum model,the algorithm is improved,and the model is compensated to improve the tracking accuracy of ZMP.Based on the research of many domestic and foreign literatures,this paper studies the gait planning and stability effects of humanoid robots,including the following aspects:Analyze the kinematics of the robot and establish a humanoid robot model.The factors affecting the stability of human walking are analyzed.The ZMP point calculation formula which can show the stability of the robot is used to obtain the conditions for judging the stability of the robot walking process.The establishment of the inverted pendulum model of the humanoid robot in the two-dimensional case is analyzed,and the inverted pendulum model in the three-dimensional space is extended,and the gait of the robot is planned by combining the predictive control method.For the problem of preview control,the ZMP tracking has a large error when the step size is large during the walking process of the humanoid robot.The preview control is improved to improve the ZMP tracking effect under the longer step motion state,and then the inverted pendulum is applied.Since the inverted pendulum as arobot model ignores too much detail,it will cause errors in the zmp calculation.In this paper,the ZMP calculation result of the seven-bar model of the robot is used as the reference input of the preview control to compensate the inverted pendulum model,thereby improving the gait planning effect.The virtual prototype of the robot was modeled by Adams,and the gait planning method of the inverted pendulum model was jointly simulated by Matlab and Adams mechanical dynamics simulation software.The main components of the lower limb movement process of the robot were observed from the post-processing module of the simulation results.And compare the experimental results before and after the improvement,the joint simulation results show that the improved system has better ZMP tracking effect.