Social Odometry in Populations of Autonomous Robots

Abstract

The improvement of odometry systems in collective robotics remains an important challenge for several applications. In this work, we propose a localisation strategy in which robots have no access to centralised information. Each robot has an estimate of its own location and an associated confidence level that decreases with distance travelled. Robots use estimates advertised by neighbouring robots to correct their own location estimates at each time-step. This simple online social form of odometry is shown to allow a group of robots to both increase the quality of individuals’ estimates and efficiently improve their collective performance. Furthermore, social odometry produces a successful selforganised collective pattern.