Research On Running Motion Control Of Quadruped Robots

Running motion generally refers to the periodic state changes of the limbs when animals move their positions at high speed in nature.In the field of robotics,unlike wheeled robots,bionic foot robots have multiple gaits to achieve their movement.For bionic quadruped robots,the most prominent gait that can achieve high-speed motion is the bound gait.This article will mainly study the controller of the quadruped robot in the bound gait.The following briefly introduces the work related to the running control of the quadruped robot in this paper.According to the principle of bionics,simple quadruped robot design guidelines were formulated about the body structures of different quadrupeds.Based on these guidelines,the simplified mechanical structure parameters of the robot in the virtual prototype were determined.Based on the kinetic energy and potential energy of the robot in the physical world,Lagrangian dynamics is used to derive the dynamic equation of single-leg motion,which provides a basis for impedance control.Finally,the problems that need attention in the centroid trajectory of the running state are analyzed,and the centroid trajectory planning plan of the robot when running is determined.Using the dynamic equation of the joint space,the impedance control based on the impedance model is designed for the swinging leg of the running robot.With this controller,the flexibility of the foot swing and the stiffness of the leg at the moment of touching the ground can be changed arbitrarily.After the controller is designed,the corresponding foot trajectory is designed for the foot end of the swing leg by using the Bezier curve.The virtual model controller is analyzed,collecting the body states information.The gait generator switches the foot controller based on the touchdown information.According to other feedback data and program preset or expected acceleration information input by the handle,the controller makes the foot end generate corresponding virtual foot end force.The supporting phase controller part linearizes the dynamics of the robot fuselage so that the problem to be solved in the later optimization becomes convex,and the required optimal solution is quickly obtained within the simulation time step corresponding to the control frequency.In the optimization part,we use the quadratic programming to solve the optimized foot end force,and then convert the foot end forced into the required joint torque through the Jacobian matrix and send it to the joint motor driver,which controls the joint motor to generate the corresponding torque.Use the Bullet simulation platform to establish a virtual prototype model and perform dynamics simulation on the simulation platform.The simulation tests the omnidirectional running motion of the quadruped running robot in the horizontal direction and tests the running stability of the robot under certain external interference or terrain interference.The established and deduced mathematical model of the quadruped robot,the correctness of forward and inverse kinematics,dynamics,planned centroid trajectory,and foot end trajectory,the feasibility of various gaits generated by the gait generator,and impedance,the controller and dynamics optimize the stability of the controller.The quadruped running robot studied in this paper can achieve forward high-speed movement at the desired speed and has an excellent performance in lateral movement and lateral impact ability tests.The designed dynamic optimization controller and gait generator not only achieve bound gait,but It can also run omnidirectionally in the horizontal direction and can switch to various other gaits to achieve high-speed movement.This paper provides a reference for the running motion of the legged robot.