Abstract
We consider the problem of target tracking under communication constraints between the sensor nodes, a problem that has recently received significant attention. We have considered a random sensor network where the entire network acts as a set of distributed datasets. Each of these datasets has its local temporal dataset along with spatial data and the geographical coordinates of a given object or target. In this paper a new algorithm is proposed which mines global temporal patterns from these datasets and discovers the nonlinear trajectories of moving objects. The main goal here is to process the data contained in the various datasets in decentralized fashion to discover global spatio-temporal patterns while keeping the communication cost as minimum as possible. We present the algorithm and analyze it in terms of the communication costs.