An Efficient Data Compression Model Based on Spatial Clustering and Principal Component Analysis in Wireless Sensor Networks

Sensors

Yihang Yin¹  Fengzheng Liu¹  Xiang Zhou¹  Quanzhong Li¹ 

[1] School of Data Science and Computer, Sun Yat-Sen University, Guangzhou 510006, China; E-Mails:

关键词: data compression;    principal component analysis;    cluster;    wireless sensor network;   

DOI  :  10.3390/s150819443

来源: mdpi

【 摘 要 】

Wireless sensor networks (WSNs) have been widely used to monitor the environment, and sensors in WSNs are usually power constrained. Because inner-node communication consumes most of the power, efficient data compression schemes are needed to reduce the data transmission to prolong the lifetime of WSNs. In this paper, we propose an efficient data compression model to aggregate data, which is based on spatial clustering and principal component analysis (PCA). First, sensors with a strong temporal-spatial correlation are grouped into one cluster for further processing with a novel similarity measure metric. Next, sensor data in one cluster are aggregated in the cluster head sensor node, and an efficient adaptive strategy is proposed for the selection of the cluster head to conserve energy. Finally, the proposed model applies principal component analysis with an error bound guarantee to compress the data and retain the definite variance at the same time. Computer simulations show that the proposed model can greatly reduce communication and obtain a lower mean square error than other PCA-based algorithms.

【 授权许可】

CC BY   
© 2015 by the authors; licensee MDPI, Basel, Switzerland.