Locomotion of Miniature Catom Chains: Scale Effects on Gait and Velocity

Scaling down the module size of a self-reconfigurable robot will have a profound effect on the module's characteristics, e.g. strength to mass ratio. In this paper we explore how the characteristics of chains of modules, specifically locomotion velocity and best gait type, might change with the scale of those modules. The simulated experiments we report on here examine module sizes from (11 mum to 698 mum radius) and chain lengths from 3 to 30 modules. All gaits tested were based on central pattern generators optimized using a genetic algorithm and hill climbing. Our results show that scaling affects both the preferred type of gait as well as a chain's overall performance (average velocity). In summary, there is a tradeoff where larger scales face the challenge of overcoming gravity, while smaller sizes face the challenge of staying in contact with the ground and the friction it provides. We show that in between these two extremes lies a "best" module size for given environmental, physical, and engineering constraints