Jumping Locomotion of Quadruped Robot During Running Based on Multiple Model Fusion

Continuous jumping motion is an indispensable part of the flexible movement of quadruped robot, and jumping is often an effective way to break through the dilemma when the robot crosses the complex rough or uneven road surface. In this paper, a quadruped robot control framework is proposed to continuously jump during running on a flat terrain. The framework switches motion modes by using a central Pattern generators (CPG)-based finite state machine(C-FSM). Multiple model fusion (MMF) method is used to realize bound gait movement of quadruped robot. The four legs of the robot are controlled by the CPG to generate periodic motion, the spring loaded inverted pendulum (SLIP) model is used to achieve flexible control and avoid disadvantages of position control. The foot force is calculated by offline optimization to control the robot jumping. In the simulation experiment, the robot can run with bound gait of 1.75m/s and the height of the center of mass (COM) reaches 1.22m, and can run continuously after landing.