Sensor-assisted Monte Carlo localization for Wireless Sensor Networks

Localization in Wireless Sensor Networks (WSNs) denotes the procedure of a single sensor node to determine its geographical position in space. As these nodes are limited in computational power, battery lifetime and communication range, there is the requirement for efficient localization algorithms which is an ongoing topic in research. Nearly all algorithms are based on the usage of seed nodes which are aware of their location and help other nodes approximating their own position. In this paper we extend an existing Monte Carlo particle filter approach (Monte Carlo Localization, MCL) to account for situations where the degree of seed nodes is low, i.e. the location estimation of a node cannot be updated. For this purpose we make use of comparatively cheap sensors to determine the movement direction and velocity of a node. With this obtained information we can update a nodes recent position estimation even in the absence of seed nodes. We simulate our approach and compare our results to the originally proposed algorithm, MCL.