Active flexibility and stability control for hydraulic quadruped robots

In this paper, we propose an impedance control (IPC) combined with a model predictive control (MPC) scheme to achieve stable walking and active exibility for quadruped robots in contact with complex environments. The support phase is controlled by MPC, which uses a simpli_ed single rigid body model to solve the optimal ground reaction force and adjust the body to the desired position in a long time range; the swing phase is controlled by IPC based on dynamics compensation, which ensures the quadruped's leg exibility and improves the accuracy of single leg trajectory tracking by reducing the impacts at the end of the foot. And the control errors of the robot's centre of mass position and centre of mass velocity are quantified by using the root mean square error (RMSE). Secondly, the RMSE is used to quantify the control errors of the robot's centre of mass position and centre of mass velocity and select the optimal gait. Finally, MATLAB/SimScape is used to simulate the results, which show that the quadruped robot has good robustness and exibility when walking in a trot gait in complex environments under this control strategy.