Vision-Based Segregation Behaviours in a Swarm of Autonomous Robots

摘要

The accelerating development and usage of autonomous robots has led to increased interest in decentralised systems and the cooperation of individual elements within a swarm, in particular the ability to self-organise and form patterns or self-segregate. Such segregation is readily exhibited in nature, as highlighted by prior work looking at the collective sorting of brood items by ant colonies [1,3]. This work focuses on the ability of robots to spatially segregate themselves. In particular, it takes inspiration primarily from the Brazil nut effect [6], which is responsible for the emergent striped patterns observed when a container holding particles of different sizes is agitated.