Design and analysis of energy reference metric in a cluster based wireless sensor network

Abstract

Wireless sensor nodes are expected to make economical, energy efficient decisions in routing of data with the limited amount of battery-voltage that they have. To make overall energy-efficient decisions, an approximation of the system energy levels can be made available to the sensor nodes. The aim of this work is to propose an algorithm to get the cluster-wise global energy status of sensors' residual energies. The study and implementation is done in a cluster-based network and the computed energy status is provided to the sensors within a cluster. An energy message is proposed to assess the sub-system energy status for each cluster. Two aggregates are explored for the energy status assessment within the energy message. The first aggregate has measurement of number of sensors with defined intervals of voltage ranges. The second aggregate considers that the cluster is divided into sectors and calculates the average of residual energies of each sector. Pre-defined bits in the energy message are used to indicate the comparison results. It is seen that as the number of sensors increased, more energy-rich sensors were chosen (45.87% higher for a 9times9 sensors sized network) in the routing process, with the proposed energy message considerations than without.