Particle positioning system for ad-hoc wireless networks

Geometric routing, data collection, node deployment and node tracking in wireless ad-hoc networks often require a form of position-awareness by the nodes. There exist several location sensing algorithms that use network topology (connectivity) to infer spatial information. Most of these rely on either centralised computation method, several well known anchor nodes/points or expensive message exchange between the nodes. We propose a particle positioning system (PPS) that uses node connectivity information collected by routing protocol to map node positions and hence avoid some of the aforementioned restrictions. PPS mainly uses a combination of particle and spring physics to simulate a system that calculates increasingly accurate relative positions with increased topology knowledge. If combined with angle of arrival information (AoA) or anchor nodes, PPS can map its relative coordinates to the real world positions. In this paper we show that PPS generates positions accurate to between 6-33% of the node's transmission range, while competitive location estimate techniques manage a 20% accuracy only in best-case scenarios.