Anding Xu∗,∗∗ , Huishen Zhu∗,∗∗,∗∗∗ , Ke Su∗,∗∗ , Yuqi Zhou∗,∗∗ , Jin Guo∗,∗∗ , Hongwu Wang∗,∗∗ , Zhiyong Jiang∗,∗∗ , and Yifan Liu∗,∗∗
Humanoid robot, virtual model control (VMC), model predictive control (MPC), robustness, active flexibility
In order to realise the stable walking of humanoid robot, a multi- model fusion complex control strategy based on the idea of task decomposition is proposed. Firstly, the virtual model control control algorithm is used to control the robot’s swinging phase, which not only ensures the robot’s flexible touch to the ground, but also improves the accuracy of the robot’s single-leg trajectory tracking. In order to further improve the robustness of the system, feed-forward correction is used to compensate. Secondly, the model predictive control algorithm is used to control the standing phase of the robot, which realises the stable adjustment of the fuselage attitude and the soft contact between the foot and the ground, thus further improving its anti-interference capability. Finally, finite-state machine is used to integrate the two control models, and MATLAB/Simulink is used to compare the control algorithms. The results show that the robot has good robustness and active flexibility under the control strategy robustness.
Important Links:
Go Back
