A meta-module approach for cluster flow locomotion of modular robots

Modular robots consist numerous interconnected and often identical components called modules, through which flexibility and versatility are attained in their consigned tasks. This paper focuses on the locomotion of modular robots comprised of Atron modules, and proposes a new algorithm based on meta-module formation and planning in order to accomplish cluster-flow-type locomotion. Through an iterative process of meta-module formation to termination lifecycle, modules move three-by-three (as metamodules) and occupy the goal region, while trying to avoid collisions with obstacles and other modules, and minimizing the paths toward their final destinations. According to experimental results, the proposed algorithm can successfully be applied to modular robot locomotion in flat, obstacled, sloped, and holed surfaces. The runtime of the algorithm grows linearly with the number of modules, and therefore is fast. Comparisons with other methods showed that the proposed method can readily be implemented for cluster flow locomotion in real-world modular robots, since unlike existing methods-which have generally dealt with abstraction models of modular robots-realistic aspects of the environment, as well as physical and mechanical specifications of the modules, are fully considered.