Data Aggregation with Principal Component Analysis in Big Data Wireless Sensor Networks

Abstract

In wireless sensor networks (WSNs), numerous sensors can produce a significant portion of the big data. It remains an open issue how to timely gather and transmit such large amount of data while minimizing data latency through wireless sensor networks (WSNs). On the other hand, spatially correlated sensor observations lead to considerable data redundancy in the network. To efficiently eliminate data redundancy and improve energy efficiency, in this paper, based on the fact that the more similar the measure data are, the smaller the amount of data after aggregation is, we first develop a new distributed clustering algorithm which can categorize sensor nodes with high similarity into a cluster for data aggregation, while ensuring uniform energy consumption within the cluster. Then, we propose a data aggregation algorithm based on principal component analysis (PCA) which can be executed in the cluster head (CH). Finally, our experimental results demonstrate that the amount of data transmission can be significantly reduced based on our proposed clustering and data aggregation algorithm.