IRIS-00-385

Primitive-Based Movement Classification for Humanoid Imitation

Odest Chadwicke Jenkins, Maja J Mataric, and Stefan Weber

Motor control is a complex problem and imitation is a powerful mechanism for acquiring new motor skills. In this paper, we describe primitives, a biologically-inspired notion for a basis set of perceptual and motor routines. Primitives serve as a vocabulary for classifying and imitating observed human movements. We demonstrate how a model of imitation can be implemented using such primitives. Furthermore, we present approximate motion reconstruction from a set of visual data taken from typically imitated tasks, such as aerobics, dancing, and athletics.