2P2-G16 成長に伴う身体特性の変化を活用した歩行獲得学習の安定化(進化・学習とロボティクス)

Abstract

The control and mechanical systems of an embodied agent should be tightly coupled so that useful functionalities such as adaptivity can emerge. This indicates that the mechanical system as well as the control system should be capable of a certain level of "computation" for generating the behavior. In order to explicitly indicate this type of "embodied" computation to be embedded in the mechanical system, Pfeifer et al. have recently coined the term called "morphological computation", which is expected to be an indispensable concept for building adaptive agents in various time scale such as here-and-now, ontogenetic, and evolutionary time scales. In this study, we focus on ontogenetic time scale of "morphological computation", and intensively discuss how the change of some body properties through growth influence stable learning to achieve bipedal walking. The preliminary simulation results derived indicate that the change of body parameter through the growth from child to adult plays crucial role in achieving efficient dynamic walking.