Research On Online Planning Algorithm Of Foot Trajectory Of Quadruped Robot

Quadruped robots have become the focus of research on robot technology due to their excellent motion performance and environmental adaptability.However,there are still great challenges to realize the environmental adaptability of quadruped robots like quadrupeds,and the basic theories and key technologies in the related fields still need to be further explored.Based on the kinematics modeling and analysis of the quadruped robots,this paper carries out research on the foot trajectory planning as well as the calculation and prediction of Fourier Coefficient of Joint Angle,focusing on the flexible transition of the foot trajectory.Finally,the online planning algorithm of robot foot trajectory is proposed by combining central pattern generator and Fourier Coefficient,whose effectiveness is verified by the numerical simulation and virtual prototype simulation.The main research work of this paper is as follows:(1)Robots kinematics and trajectory planning.The general design scheme of quadruped robots is put forward.On the basis of the overall design scheme,a simplified model is established,based on which,the forward and inverse kinematics analysis of a single leg is set up.According to the fact that the impact between the foot and the ground is small with no relative sliding,the foot trajectory planning is carried out to obtain the motion function of the joint in the leg.(2)Robots joint motion network prediction.Fourier Coefficient of the robot joint motion function is obtained based on the Fast Fourier Transform.Taking the step size S and height H of the robots and the Fourier Coefficient of its corresponding joint motion function as outputs,an online prediction model of the Fourier Coefficient of each joint motion function of the robots is established based on BP network.(3)On-line planning algorithm for robot foot trajectory.Combining CPG and Fourier series,a new on-line planning algorithm for robot foot trajectory is proposed.Firstly,the CPG model output is used to replace the coefficients of the Fourier series of the online prediction model to construct a new Fourier series.Secondly,Tu and Tv,the appropriate CPG model parameters,are selected.Then,the joint motion function corresponding to the foot parameters is constructed through the CPG model output to adapt to the step size S and height H,thus realizing the online foot trajectory planning.(4)Robots numerical simulation and virtual prototype simulation.Based on the above-mentioned algorithm,the experiments on numerical simulation and virtual prototype simulation are carried out,which show that the online planning algorithm of robot foot trajectory combining CPG and Fourier series proposed in this paper is effective.