Locally Linear Embedding for Node Localization in Wireless Sensor Networks

RSSI gives an initial rough measure of the inter node distances at low cost without the need of additional equipment or complexity. This necessitates the need for a mechanism to obtain accurate node locations from the noisy distance estimates. Manifold learning techniques can be used for estimating locations, but their ability to localize node in the sensor network environment has not been benchmarked. In this paper, locally linear embedding (LLE) has been proposed for localization of nodes from noisy RSSI distance estimates by viewing the localization process as dimensionality reduction. The efficacy of centralized LLE technique to localize sensor nodes has been studied with respect to localization accuracy, communication and computation overhead. Simulation results show that, the centralized LLE is able to localize nodes with high accuracy. However, information collection on the sink node requires extensive message passing from all nodes to the sink.