One Of The Gait Planning And Walking Feedback Control Algorithm For Humanoid Robot

Humanoid robot can walk in all directions and adapt to various complex terrains due to its bipedal walking movement form.To this end,humanoid robot has shown a broad application prospect in the fields of service,medical treatment and complex terrain exploration in recent years,so,the developed countries such as Japan and Europe and the United States have also shown great enthusiasm for their research.The most basic human ability of a humanoid robot is bipedal walking,what's more,stable bipedal dynamic walking is the basis for the realization of other types of human movements.Achieving bipedal dynamic walking has become a hotspot and difficulty problem in the field of robotics.In this paper,the off-line gait planning and on-line stability control of humanoid robot are extensively studied,and the main work can be summed up in the following aspects.1.The gait planning method based on linear coupled oscillator model is analyzed and selected.To solve these problems that the partial parameters of the coupled oscillator model can only be controlled by manual adjustment,and the robot is prone to instability on X axial under high speed or strong acceleration,the centroid offset control is added to the oscillator model,and an algorithm to optimize the parameters of a partial oscillator is proposed to improve the stability of the robot's walking in the X axis.The ZMP stability margin of the X axial and the vibration amplitude of the center of mass are combined as the optimization targets,and the genetic algorithms are used as a solution tool,the aim is to solve the optimal parameters at different velocity inputs.The MATLAB simulation experiment data show that the proposed optimization method has a greater stability margin on the X axis,at the same speed input.2.The gait trajectory planned by off-line gait planning is very effective in the simulation environment as no disturbance,but the disturbance is unavoidable in the actual operation.When the pitching angle deviation occurs because of the disturbance,it is easy to lead to the front and rear oscillations of the centroid of the robot.To this end,this paper depth analyzed the causes of the shock formation,and put forward a method to dynamically adjust the time of support leg switching to suppress concussion based on the attitude feedback.Finally,the PD algorithm is used to adjust the stride and frequency,and the pitching angle deviation is reduced or eliminated.3.The powerful and widely used Robot Operating System(ROS)in the current robot system has been analyzed,on this basis,the gait frame based on the linear coupled oscillator model has been realized,according to the specific model of the Nimbro robot,and the platform for data simulation and visualization model simulation is also set up.The comparative experimental data of the simulation platform without interference further proves that the gait optimization algorithm in this paper can plan the better ideal gait curve with higher stability.The experimental data of solid robot interference show that the attitude feedback control algorithm can detect,respond,and eventually reduce or eliminate the deviation.