Unequal Distributed Spatial Correlation-based Tree Clustering For Approximate Data Collection In WSNs

Wireless sensor network (WSN) has involved into our real life, such as the Medical Treatment, Traffic, Environment, Military, Smart Home and so on, rather than limited in the theoretical research. WSN is different with the traditional wireless network, including network range, data-centric, limited energy, based on application, etc. Because of the constrained energy, the design of energy-efficient routing protocol becomes a top priority.According to the different classification criteria of routing protocols, it can be divided into three categories:data-centric, based on the geographical position, based on cluster structure. In the third category, there are some routing protocols which are based the spatial-temporal correlation of data. These protocols pay more attention on the data content; they combine the node residual energy, data similarity with location to partition the whole network into clusters such as the typical routing protocols DSSC, ASAP, EEDC, DClocal and DCglobal. On the basis of the DSSC agreement, this paper proposes a new protocol UDSCTC (Unequal Distributed Spatial Correlation-based Tree Clustering for Approximate Data Collection in WSNs).UDSCTC is used to solve the problem of same competitive radius and the directly communication between the clusters and sink node in DSCC protocol. We carry out the innovation from the following aspects:(1)We modify the radius of node competition depending on the distance of the nodes to the sink node to make close to the sink node of the cluster radius decreases;(2) realize the compromise of energy-consumption between intra and inter clusters. At the same time, we make the clusters’data transmitted to the sink node by multi-hops. The method can enlarge the network area.(3)In order to get the simulation results of the improved protocol, we use MATLAB to compare the UDSCTT with those classical agreements, such as DSCC, ASAP. We pay more attention on the number of cluster heads, the quantity of forced-cluster, the variance of the size of cluster, the consumption of energy, the average residual energy of cluster nodes, and the average similarity of all the clusters. The simulation results show that UDCSTC is better in saving energy, prolonging the lifetime of network than the classical protocols.