Date Added: May 2009
This paper presents algorithms for efficiently detecting the variation of a distributed signal over space and time using large scale wireless sensor networks. The proposed algorithms use contours for estimating the spatial distribution of a signal. A contour tracking algorithm is proposed to efficiently monitor the variations of the contours with time. Use of contours reduces the communication cost by reducing the participation of sensor nodes for the monitoring tasks. The proposed schemes use multi-sensor collaboration techniques and non-uniform contour levels to reduce the error in reconstructing the signal distribution. Results from computer simulations are presented to demonstrate the performance of the proposed schemes.