Motion Optimization Using Gaussian Process Dynamical Models

[ Information ]

Multibody System Dynamics, vol. 34, no. 4, pp. 307-325, 2015.

[ Authors ]

Hyuk Kang, and Frank C. Park

[ Abstract ]

We propose an efficient method for generating suboptimal motions for multibody systems using Gaussian process dynamical models. Given a dynamical model for a multibody system, and a trial motion, a lower-dimensional Gaussian process dynamical model is fitted to the trial motion. New motions are then generated by performing a dynamic optimization in the lower-dimensional space. We introduce the notion of variance tubes as an intuitive and efficient means of restricting the optimization search space. The performance of our algorithm is evaluated through detailed case studies of raising motions for an arm and jumping motions for a humanoid.