318 Acquisition of Adaptive Behavior for Virtual Modular Robot Using Evolutionary Computation

Abstract

There are many studies on hardware and software experiments to control a modular robot. The modular robot consists of homogeneous or heterogeneous simple modules. A feature of it is that the modular robot behavior is obtained by an each module movement based on local communications between linked modules. This study focuses on a computer simulation obeying physics laws for a virtual modular robot. This simulation aims at achieving given tasks autonomously for it in specific circumstances. Our goal is to acquire a control system in evolution for the virtual modular robot and to analyze and evaluate their behavior.