Energy Efficiency and Quality of Data Reconstruction Through Data-Coupled Clustering for Self-Organized Large-Scale WSNs

Energy efficiency has been a leading issue in Wireless Sensor Networks (WSNs) and has produced a vast amount of research. Although the classic tradeoff has been between the quality of gathered data versus the lifetime of the network, most works gave preference to an increased network lifetime at the expense of the data quality. A common approach for energy efficiency is partitioning the network into clusters with correlated data, where the representative nodes simply transmit or average measurements inside the cluster. In this paper, we explore the joint use of in-network processing techniques and clustering algorithms. This approach seeks both high data quality with a controlled number of transmissions using an aggregation function and an energy efficient network partition, respectively. The aim of this combination is to increase energy efficiency without sacrificing the data quality. We compare the performance of the Second-Order Data-Coupled Clustering (SODCC) and Compressive-Projections Principal Component Analysis (CPPCA) algorithm combination, in terms of both the energy consumption and the quality of the data reconstruction, to other combinations of the state-of-the-art clustering algorithms and in-network processing techniques. Among all the considered cases, the SODCC + CPPCA combination revealed a perfect balance between data quality, energy expenditure, and ease of network management. The main conclusion of this paper is that the design of WSN algorithms must be processing-oriented rather than transmission-oriented, i.e., investing energy on both the clustering and in-network processing algorithms ensures both energy efficiency and data quality.