Key Pose-Based Dynamic Humanoid Motion Generation Using Parallel Multi-Fidelity Model Predictive Control

Abstract

This paper proposes a method that realizes dynamic motion of humanoid robots from reference key poses. The proposed control method runs two types of model predictive controllers with different fidelity and time scale in parallel; one performs long-horizon prediction by making use of a closed-form solution of the centroidal dynamics, and the other performs short-horizon prediction based on the whole-body dynamics. In dynamical simulation of 32-DoF humanoid robot, the controller was able to perform challenging motions including toe contact and jumps over unlevel surfaces in real time computation speed without any offline optimization.