A novel centralised clustering algorithm for energy efficient wireless sensor networks. (4th August 2008)

• Record Type: Journal Article
• Authors: Colombo, Stefano; Chilamkurti, Naveen; Zeadally, Sherali
• Is Part Of: International journal of autonomous and adaptive communications systems. Volume 1:Number 2(2008)
• Page Start: 242
• Page End: 261
• Publication Date: 2008-08-04
• Subjects: communications -- clustering algorithms -- energy efficiency -- wireless sensor networks -- WSNs -- wireless networks -- networking -- performance -- protocol -- energy efficient routing; Adaptive computing systems -- Periodicals; Wireless communication systems -- Periodicals; Computer networks -- Periodicals; 004.6
• Languages: English
• ISSNs: 1754-8632

Abstract

Wireless Sensor Networks (WSNs) consist of a large number of low-power nodes sensing data from the environment. The sensor nodes are equipped with small and often irreplaceable batteries with limited power capacity. One of the areas of sensor networking that has attracted a lot of attention recently is the prolonging of the life span of sensor networks by exploiting energy-efficient routing schemes. We propose a novel energy-efficient centralised clustering algorithm for WSN. The proposed algorithm generates a set of possible clustering alternatives from which it selects the optimal clustering. We also present a performance evaluation of the proposed algorithm using two performance metrics Max-min and Max-sum. We found that Max-sum yields superior performance. In addition, our proposed algorithm improves the system lifetime performance over Low Energy Adaptive Clustering Hierarchy-C by almost 15%. We also compare performance results obtained with the chosen metric (Max-sum) with those obtained using other strategies (such as random and configuration 0). The Max-sum metric outperforms the random and configuration 0 strategies by almost 8.5 and 11%, respectively, for system lifetime. Moreover, we found that the total amount of data collected using the Max-sum metric is at least 9% higher than with random and configuration 0 strategies.