Lembono, T.S., Mastalli, C., Fernbach, P., Mansard, N. and Calinon, S. (2020)
Learning How to Walk: Warm-starting Optimal Control Solver with Memory of Motion
In Proc. IEEE Intl Conf. on Robotics and Automation (ICRA), pp. 1357-1363.
Abstract
In this paper, we propose a framework to build a memory of motion to warm-start an optimal control solver for the locomotion task of the humanoid robot Talos. We use HPP Loco3D, a versatile locomotion planner, to generate offline a set of dynamically consistent whole-body trajectory to be stored as the memory of motion. The learning problem is formulated as a regression problem to predict a single-step motion given the desired contact locations, which is then used as building block for multi-step motions. The predicted trajectory is then used as warm-starts for the fast optimal control solver Crocoddyl. We have shown that the approach manages to reduce the required number of iterations to reach the convergence from ∼9.5 to only ∼3.0 iterations for the single-step motion and from ∼6.2 to ∼4.5 iterations for the multi-step motion, while maintaining the solution's quality.
Bibtex reference
@inproceedings{Lembono20ICRA,
author="Lembono, T. S. and Mastalli, C. and Fernbach, P. and Mansard, N. and Calinon, S.",
title="Learning How to Walk: Warm-starting Optimal Control Solver with Memory of Motion",
booktitle="Proc. IEEE Intl Conf. on Robotics and Automation (ICRA)",
year="2020",
pages="1357--1363"
}