A Load Balancing Cross Clustering Approach in Wireless Sensor Network

From last decade the extensive use of Wireless Sensor Networks in different application areas and environmental conditions where human intervention is not very economical and favorable has led to an increased demand of energy efficient algorithms. Due to the use of random deployment strategy in such areas, the load must be balanced equally among the sensor nodes for efficient utilization of energy and extending the network life time. Hence, an efficient load balancing algorithm is proposed for grid-based clustering where merging and splitting of clusters is devised for sparse and dense clusters respectively and cross clustering is applied for averagely dense clusters. We have used multi-hop communication among the Cluster Heads (CHs) which reduces the energy consumption as well. The decision of cluster members to join adjoining cluster depends on its distance from both its current CH and adjoining CH. Thus communication of data, that requires the highest amount of energy, will be done for a short distance. Hence, the amount of energy spent by cluster members in communicating is also reduced. The overhead comparison with an existing algorithm depicts that the proposed algorithm being centralized in nature generates asymptotically the same amount of control packets as a distributed approach does. Our algorithm can perform better in terms of energy consumption and load balancing which eventually increases the network lifetime and minimizes creation of energy holes.