Information Aggregation Mechanisms in Social Odometry

Multi-robot exploration and navigation is a challenging task, especially within the swarm robotics domain, in which the individual robots have limited capabilities and have access to local information only. An interesting approach to exploration and navigation in swarm robotics is social odometry, that is, a cooperative strategy in which robots exploit odometry for individual navigation, and share their own position estimation through peer-to-peer local communication to collectively reduce the estimation error. In this paper, the robots have to localize both a home and a goal location and navigate back and forth between them. The way in which navigational information is aggregated influences both the efficiency in navigation between the two areas, and the self-organized selection of better paths. We propose three new parameterfree mechanisms for information aggregation and we provide an extensive study to ascertain their properties in terms of navigation efficiency and collective decision.