Error estimation in wireless sensor networks

We present an analogy between the operation of a Wireless Sensor Network and the sampling and reconstruction of a signal. We measure the impact of three factors on the quality of the reconstructed data, namely, the granularity of the process under study, the spatial distribution of sensors, and the protocol for clustering and data aggregation. In order to quantify this influence, a Monte Carlo study is performed for estimating the error introduced by the observation process. The phenomenon being observed is described by a Gaussian random field with varying scale, the distribution of sensors is modeled by a new point process and two protocols are assessed: Leach and Skater. We show that Skater performs better than Leach, at the expense of using the sampled data on the clustering stage. Categories and Subject Descriptors C.2 [Computer-Communication Networks]: General; G.3 [Probability and Statistics]: Stochastic Processes General Terms Wireless Networks Keywords Estimation, signal proce...

• Type: Journal
• Year: 2008
• Where: SAC
• Authors: Alejandro César Frery, Heitor Ramos, José Alencar Neto, Eduardo Freire Nakamura
• Keywords: Applied Computing, Random Fields, SAC 2008, Sensor, Wireless Sensor Networks