Self-reconfigurable modular robots are usually composed of multiple building blocks of a relatively small repertoire, with uniform docking interfaces that allow transfer of mechanical forces and moments, electrical power, and communication throughout the robot. Numerous modular robotic systems have been developed in the past few years. The design goal for modular robots is supposed to be versatile, robust and low-cost. Modular robots should be able to adapt or be adapted to many different functions or activities, handle hardware and software failures, and also be cost-effective to be used in more cost-sensitive tasks. With these advantages, modular robotic systems are promising to do a wide variety of tasks. However, it is also a challenge to come up with planning and control algorithms to handle numerous modules. One key reason is that, the number of all possible configurations of the system increases drastically as the number of modules increases. Plus, each module may have multiple degrees of freedom and the configurations in topology become even more complicated. Hence, configuration recognition, namely matching a new modular robotic configuration to an existing configuration in a library and mapping each module in these two configurations, is an essential problem we need to solve for planning and control of modular robots. Automatic configuration recognition is the process by which a modular system can determine its own configuration without having it explicitly programmed and a variety of its uses are described in [3]. In addition to those applications, automatic configuration recognition is also necessary for fully autonomous modular robots.
展开▼
