Global distributed clustering technique for randomly deployed wireless sensor networks

Abstract

Wireless sensor network applications are composed of a vast number of inexpensive battery-powered sensors. One of its primary applications is environmental monitoring for physical phenomena in rigid areas such as forests and volcanoes. In such applications, a large number of sensors are randomly scattered by aircraft over the area of monitoring. These applications mainly depend on clustering to arrange nodes into groups to facilitate their communication. Previously proposed clustering techniques are classified into two types, which are distributed or centralized techniques. Each of these types has advantages as well as some flaws. In this paper, we propose a globally distributed clustering technique. This technique depends on some global information about the network to allow each node to decide its role in the produced clusters locally. This information is assumed to be known by default by the BS for any communication or topological control activities. Simulation results show that the proposed technique achieves less power consumption and therefore longer network lifetime when compared with other clustering techniques.