Indoor Localization with Low Complexity in Wireless Sensor Networks

Autonomous localization of nodes in wireless sensor networks is essential to minimize the complex self organization task consequently enhancing the overall network lifetime. Recently, precise indoor localization is impeded by multi path propagation of signals due to reflections at walls or objects. In this paper we partly overcome some of these problems by methods like frequency diversity and averaging multiple measured data. Received radio signal strength (RSS) in combination with weighted centroid localization, featuring low communication overhead and a low complexity of O(n), is our basis of a localization on the energy constrained sensor nodes. We first analyze the RSS-characteristics on a specific sensor node platform in different rooms. Next, we describe methods to improve these characteristics to reach best localization results at minimized complexity. Finally, in a practice indoor localization we achieve a small localization error of only 14% for 69% of all test-points that was enhanced to at least 8% in average by simple optimizations. For that, no hardware modifications as well as time consuming RSSI-maps or complex signal propagation models are required.