Self-organization of reaching operation in a robot: a bio-mimetic approach

Recently a new approach, called, behavior based robotics, has been proposed to overcome the difficulties of the traditional AI approach to robotics. The approach has two goals: an engineering goal of building a flexible autonomous robot, and a scientific goal of understanding human cognition. As one of behavior based robotic approaches, this paper deals with the reaching operation development in a simulated robot. Based on the importance of the contraction of muscles controlling body parts such as binocular eye balls, a neck and an arm, the reaching in the robot is modeled and learned via a learn-by-doing manner. And the learning result is stored in a memory as a form of experience of the muscle contraction, which is considered to be useful as a critical cue for space perception by the vision system and for arm reaching