Compressed Data-Gathering Method based on Spatiotemporal Correlation Clustering in Wireless Sensor Networks

Abstract

To solve the problem of high energy consumption in traditional data-gathering protocols and to balance the network load, a spatiotemporal correlation-based clustering method for compressive data gathering (SCCM-CDG) is proposed. First, we present a mathematical model to measure the spatiotemporal correlation of neighborhood nodes. Second, a spatiotemporal correlation-based clustering method (SCCM) is proposed and then applied to the data-gathering protocol. Sensor nodes within a cluster send a small number of linear projections to cluster heads using compressive sensing theory, and then cluster heads send sample data along the shortest square distance spanning tree among cluster heads and the sink. Results of the simulation and experiment verify the accuracy of the SCCM algorithm, revealing that the SCCM-CDG algorithm can substantially reduce energy consumption, prolong network lifetime, and promote improvements in data recovery at the sink compared with existing compressive sensing-based data-gathering schemes.