DISTRIBUTED LOCOMOTION ALGORITHMS FOR SELF-RECONFIGURABLE ROBOTS OPERATING ON ROUGH TERRAIN
Abstract: In this paper, we describe a set of distributed algorithms for self-reconfiguring modular robots that allow them to explore an area in parallel. The algorithms are based on geometric rules that each module evaluates independently relative to its local neighborhood. This paper concentrates on developing algorithms within this framework to enable travel over the widest
variety of terrain. In particular, we show how to perform straight-line motion, turning while on obstacles, climbing over tall obstacles, and tunneling under overhangs, all of which work for groups of arbitrary size. This last feature is important, as it also allows a large system of self-reconfiguring modules to divide up into several groups of various sizes, each of which is equally capable of motion and participation in the overall group task. We also discuss implementations and ways to improve efficiency and switching between tasks.