Energy optimization algorithm based on data density correlation in wireless sensor network

Abstract

It’s importance to send the typical data from sampled data to the sink node in wireless sensor network. Compared with actual data, the representative data are always imprecise. Moreover, the energy consumption is huge. In order to minimize the energy consumption and improve the data accuracy, this paper presents the data fusion model PCCDNCD (correlation degree base on the Pearson correlation coefficient, the distance factor and the number of neighbor nodes, PCCDNCD). The correlation degree formula which can characterize the node from three aspects and classify nodes into three types precisely, is based on the Pearson correlation coefficient, the distance of nodes and the number of neighboring nodes. Nodes are classified into typical、ordinary and isolated nodes. In addition, the typical and isolated nodes are responsible for transferring data, while ordinary nodes are not required. The results show that the typical data achieved by the PCCDNCD method have higher degree of accuracy than the data from PCC (the Pearson correlation coefficient, PCC) and DDCD (the data density correlation degree, DDCD) methods. Meanwhile PCCDNCD algorithm has a low energy consumption.