Robust walking stabilization strategy of humanoid robots on uneven terrain via QP-based impedance/admittance control

Abstract

A robust walking stabilization strategy of humanoids on uneven terrain via a QP-based impedance/ admittance control is addressed in this paper. The core idea is combining the following two strategies. The first is to reduce the effect of an unexpected contact force on the centroidal momentum dynamics, and the second is to adjust post-contact reference for the swing foot with which its pose is regulated on the obstacle. The former can be achieved by replacing the task of the trajectory tracking control for the swing foot with its task-space impedance control, and the latter follows by employing the hybrid admittance control combining the admittance control with resetting the post-contact reference. In addition, an optimal set of parameters used for the admittance control is computed via the Taguchi optimal design method. The proposed algorithm is embedded into the momentum-based whole-body control (WBC) framework and verified its validity by multiple simulations with the physics engine.