Comparative Study and Analysis of Hierarchical Clustering Methods in Wireless Sensor Networks

Hierarchical Clustering is a procedure of cluster analysis which aims to construct a hierarchy of clusters. There are two kinds of hierarchical clustering i.e. Agglomerative, which is a bottom – up approach, where all the observations start in its own cluster, and pairs of clusters are merged moving up the hierarchy, and the other one is divisive, which is a top down approach, where each observation starts in one cluster, and splits up recursively while moving down the hierarchy. The main problem is Shortage of Network Lifetime, Presence of less Residual Energy, Cost of building the clusters and the issue of Dead Nodes, which occur very frequently. In the earlier work, only three parameters were considered i.e. Battery level, Node density and Proximity to Base station for the Effective utilization of the Cluster by Fuzzy inference Engine. They optimized the clustering process, Cluster Head Election and decreased the number of dead nodes. A lot of work has been done in Fuzzy based System Simulation for Cluster Head Selection in Wireless Sensor Networks and Fuzzy System Based Cluster Selection ( FSCS ) Technique has been proposed by taking two Fuzzy logic controllers, and by using parameters i.e. Remaining Battery Power, DCC, Degree of Neighbor Nodes, Feedback sensor speed, Sensor speed, and Probability of Cluster Head Selection for the Effective utilization of the Cluster and decreased the Probability of the nodes, Controlled RPS, and increased the Controlled Feedback Speed. The objective is to cluster the nodes in a hierarchical way, by taking as many parameters as we can in order to decrease the number of dead nodes, save of cost of creating new clusters, increase the necessary residual energy, and enhance the network lifetime. The methodology applied is to select a cluster among the network randomly and calculate their weight functions. Cluster head Election is done on the basis of weight functions and weight function of the next hop is calculated. Among all the nodes, the optimized Next hop is calculated and Hierarchical Routing is performed using Fuzzy Inference Engine. The threshold time is calculated for the cluster head, if it is achieved then again the cluster head election will take place in the same cluster. By applying this method, the residual energy and network lifetime can be increased, whereas the cost of creating the clusters and number of dead nodes can be decreased. Keywords—Hierarchical Clustering; Residual Energy; Network Lifetime, Battery level, Fuzzy Inference Engine, Cluster Head Election, Next hop.