Research On Motion Planning And Control Of Quadruped Robot Based On Optimal Distribution Of Foot Contact Forces

The legged structure is very common in the body of animals.Due to its good adaptability to complex terrain,the legged animals have almost spread all over the mainland.Now experts have developed a new mobile robot based on the leg foot structure,which is called quadruped robot.Quadruped robot has the advantage of good terrain adaptability,which makes quadruped robot can complete some tasks instead of human,especially in some dangerous places,so as to ensure the safety of people's lives.But at present,the research and development technology of quadruped robot is not mature.Most quadruped robots have the problems of poor stability and terrain adaptability.But at present,the research and development technology of quadruped robot is not mature.Most quadruped robots have the problems of poor stability and terrain adaptability.In order to solve these problems,the motion planning and control algorithm of quadruped robot is studied,and the work is divided into four aspects: kinematics modeling,state and terrain estimation,motion planning and control,simulation analysis.Firstly,the mathematical model of the quadruped robot platform is established by analyzing the structural characteristics.In order to obtain the relationship between the joint angle and the feet's position of the leg,the forward and inverse kinematics is solved by using the geometric method.The workspace of the foot is also solved,which provides constraints for the optimization of the foot position.Secondly,the state information of the quadruped robot and terrain environment are estimated.The state estimation of the quadruped robot is realized by a Kalman filter,which combines the acceleration of the body,the attitude angle,the encoder of the legs,the touchdown information and the kinematics.This method can get the estimated values of the center of mass displacement,velocity and the foot position of the quadruped robot in three-dimensional space.In order to perceive the terrain environment and adjust the control strategy of the quadruped robot on the slope terrain,the local terrain of the quadruped robot is fitted into a space plane by using the least square method,and then the desired attitude angle of the quadruped robot is solved by using the fitting ground information.Then the State Planning and motion control algorithm of the quadruped robot is established.As for the control of the stance phase,the virtual model is used to calculate the expected acceleration of the fuselage,and the force is distributed to the stance foot through the optimal solution.As for the control of swing phase,the expected landing point of swing leg is planned by using the capture point theory of linear inverted pendulum.The inverse kinematics is used to convert the trajectory control of the foot to the joint space control.The state planning is based on ZMP(zero moment point)model,and the constraint conditions between force and state are established.The optimal distribution of foot force is used as the cost function to predict the state of the quadruped robot in the future.Finally,a series of simulation tests are carried out by using ADAMS and MATLAB co-simulation.The simulation includes quadruped robot attitude control,centroid position control,trot gait walking on flat ground and slope.The simulation results show that the motion of the quadruped robot is stable,and the simulation data is basically consistent with the state of the planning.In general,the control planning algorithm proposed in this paper is feasible.