SPATIAL CORRELATION MODELS ANALYSIS FOR WIRELESS SENSOR NETWORK

Abstract

Wireless sensor networks (WSN) are deployed very densely so that accurate monitoring can be achieved in many applications. For example, the deployment density can be as high as 20 nodes/m3[1]. As sensor nodes are deployed with high density, the information observed and transmitted by sensor nodes in an area at the time an event is occurred is spatially correlated. It means there is spatial correlation in the sensors observations. The spatial correlation along with collaborative nature of the Wireless sensor networks (WSN) has potential advantages in designing energy-efficient models of Communication protocols. There exists a spatial correlation function[7], which describes correlation characteristics of information sensed by sensor nodes, which is related to location and coverage model of sensor nodes. This paper presents a study of various cluster head selection algorithms and proposes an energy efficient model for wireless sensor network by exploiting spatial correlation among various sensor nodes of the network. Firstly, study the behavior of wireless sensor network without spatial correlation. Secondly, we divide entire wireless sensor network into clusters assuming spatial correlation among the observations within the nodes of a cluster. One of the nodes of the cluster is considered as Cluster Head (CH) on random selection and sends the observation to Base Station (BS). In last part of analysis we make one of the Cluster Head (CH) as super Node (SN) that aggregates observations from all Cluster Head (CHs) and sends information to Base Station (BS). Finally, based on the above analysis and their results we propose an energy efficient model for wireless sensor network.