Localization of a Swarm of Mobile Agents via Unscented Kalman Filtering

Abstract

This paper deals with the application of Kalman filtering (KF) techniques to the localization of a swarm of mobile agents in a wireless sensor network (WSN). In particular, both extended (EKF) and unscented (UKF) Kalman filters have been investigated referring to a typical urban scenario with energetic and resource constraints. A cooperation strategy among sensor nodes, based on a virtual diversity scheme, has been introduced allowing the swarm tracking under severe propagation conditions. The effectiveness of the proposed solution has been assessed by means of simulations concerning a squad of robots moving in realistic scenarios. It has been shown that UKF achieves a higher accuracy and reliability than EKF in localizing the barycenter of the robot squad. Further, the proposed solution provides advantages in terms of measurement update frequency and, hence, of energy saving.