An Efficient Load Balancing Clustering Scheme for Data Centric Wireless Sensor Networks

Clustering is an efficient approach to capitalize the energy of energy constraint sensor nodes in wireless sensor networks. Clustering schemes do not guarantee formation of clusters with equal number of nodes. So data frames transmitted by the nodes vary. TDMA schedule of nodes of smaller cluster is smaller than others that results more number of data frames and hence more energy consumption. The non uniform energy consumption of nodes affects the load balancing of network and these nodes are more prone to die earlier than others. In this paper, an improved scheme for cluster head selection is proposed. Clusters having variable frame slots for nodes are applied to E-LEACH and improved E-LEACH to make the cluster more load balanced. Simulation is carried out in NS-2 to analyze the performance of E-LEACH and improved E-LEACH with variable frame length. Variable frame slot scheme for clusters is also measured with the varying distance of base station from the field. Simulation results show that clustering with variable frame length has an improvement of 7% in node death rate over E-LEACH and an improvement of 9% in node death rate over improved ELEACH. Results suggest that variable frame length scheme improves the performance of clustering schemes for WSNs and have most significant result at base station located at 75m from the field.