Swarm-Intelligent Localization

In wireless sensor networks we often want to know where individual sensor nodesare physically positioned in order to make sense of the data that they report. Theprocess of obtaining such position information for the nodes is known as localization.Simple solutions to the problem of localization are to either place the nodes manuallyat specified places, or to use some special localization hardware such as GPSreceivers. However, these solutions can be impractical or too costly, especially forlarge networks. Instead we can use some algorithm to try to compute the nodes'positions based on available data. We present a new distributed algorithm, which wecall Swarm-Intelligent Localization (SIL), for computing these positions. Our algorithmassumes that a fraction of the nodes, the so-called anchors, have an a prioriknowledge of their positions, and that noisy range measurements can be madebetween neighbouring nodes in the network. The average computational complexityper node running SIL is constant in the network size, and linear in the connectivity ofthe network. We evaluate the algorithm through simulations of different networktopologies with varying parameters, such as network size, range measurement errors,fraction of anchors and connectivity. The results of the simulations indicate that inmost cases SIL can successfully locate the majority of sensor nodes with reasonableaccuracy, even in the face of difficulties such as large distance measurement errors