Motion Planning And Control For Jumping And Running Of Quadruped Robot

The quadruped robot has strong adaptability to various environments where wheeled robots cannot adapt and has a bright prospect in applications in unstructured environments.In these environments,fast and flexible motion control can enhance the capability of quadruped robot.Hence,it is valuable to study the dynamic and stable motion control of quadruped robot.This thesis studies the control algorithms of continuous and stable motion of trotting and bounding gaits on the quadruped robot platform "Chitu".The main research work include following points:1.Robot modeling.According to the Virtual Model Control algorithm,the relationship between the torques of the joints and the ground reaction forces of each leg is derived where the body and leg of the robot is modeled separately.In addition,the Kalman filter is used to estimate the velocity of the robot to improve the performance of speed control.2.Trotting motion control.A ground force distribution method is proposed to satisfy the underactuated constraints,as well as a trotting motion control strategy is given to realize continuous and stable trotting motions on the real quadruped robot.3.Bounding motion control.A "Virtual Vertical Spring" control law and a"Motion Phase" calculation&feedback algorithm are proposed to control the posture of the robot.Then,a bounding motion strategy is implemented to maintain the stable motion cadence based on these algorithms.The experiments are conducted both in co-simulation platform and the real quadruped robot.