Abstract

Vision is one of the most important sensing modalities for robots and has been realized on mostly large platforms. However for micro robots which are commonly utilized in swarm robotic studies, the visual ability is seldom applied or with only limited functions and resolution, due to the challenging requirements on the computation power and high data volume to deal with. This research has proposed the low-cost micro ground robot Colias IV, which is particularly designed to meet the requirements to allow embedded vision based tasks onboard, such as bio-inspired collision detection neural networks. Numerous of successful approaches have demonstrated that the proposed micro robot Colias IV to be a feasible platform for introducing visual based algorithms into swarm robotics.