Advanced Principal Component-Based Compression Schemes for Wireless Sensor Networks

Abstract

This article proposes two models that improve the Principal Component-based Context Compression (PC3) model for contextual information forwarding among sensor nodes in a Wireless Sensor Network (WSN). The proposed models (referred to as iPC3 and oPC3) address issues associated with the control of multivariate contextual information transmission in a stationary WSN. Because WSN nodes are typically battery equipped, the primary design goal of the models is to optimize the amount of energy used for data transmission while retaining data accuracy at high levels. The proposed energy conservation techniques and algorithms are based on incremental principal component analysis and optimal stopping theory. iPC3 and oPC3 models are presented and compared with PC3 and other models found in the literature through simulations. The proposed models manage to extend the lifetime of a WSN application by improving energy efficiency within WSN.