Representing and discovering the configuration of Conro robots

Abstract

A Conro reconfigurable robot is formed by joining set of self-contained modules in a particular configuration; the actions of the robot are the result of the coordinated actions of its modules. These actions can be controlled using a master-slave approach only if the master can map the particular configuration of the robot to one that it already knows how to control. We discuss how to describe this configuration using graphs, how to discover a robot configuration and how to identify it as a particular known configuration. The methodology used is very general and can be applied easily to other modular robots. Experimental results for Conro quadrupeds and snakes are presented.