Research on adaptive impedance control strategy for humanoid walking in unstructured dynamic environment

Abstract

Adaptability and robustness are the important expressions of intelligent walking ability of humanoid robots. However, they may be in an unstable state due to the huge impact contact forces produced by foot instant landing. This paper is concerned with the problem of dynamical biped walking and robust control of humanoid robots under ground reaction forces (GRF). In order to imitate human’s muscles to absorb the landing force, the robotic system is modeled as a mass–damp–spring model. The novelty of the article lies in the use of impedance control based on ground reaction forces, which deals with the complicated optimization problem subjected to both equality and inequality constraints. A feedback controller is designed to utilize inertial damping to generate the desired motion trajectory of the robot. The constructing autonomous evolution mechanism is mentioned to realize adaptive optimization of walking model. It ensures that the impact of GRF and reinforce stability during transition from single support phase to double support phase. Finally, the effectiveness of the proposed method is verified by simulations.