A new clustering scheme for wireless sensor networks

Abstract

Energy consumption has always been the most serious issue to consider while deploying Wireless Sensor Networks (WSNs). Sensor nodes are limited in power, computational capacities and memory so reporting the occurring of specific events, such as fire or flooding, as quickly as possible using minimal energy resources is definitely a challenging issue. In this work, we propose a novel, reactive and energy-efficient scheme in order to report events. In our algorithm, nodes that detect a certain event will organize themselves into a cluster, elect a clusterhead that will collect data from the cluster members, aggregate it and forward it to the mobile sink. Our approach considers the signal strength received from the mobile sink as well as the residual energy of sensor nodes. It is worth to mention that our scheme is completely reactive, so no pre-configuration phase is needed, which will lead to a high energy saving. In order to implement and evaluate our algorithm, we used Opnet simulator. A new Opnet sensor node model, where the network layer is implemented from scratch, is also proposed in this paper; our scheme is implemented using Opnet Process Modeling Methodology (PMM).