Flow Simulation Based Energy Efficient Clustering in Wireless Sensor Network

The paper introduces a flow simulation based clustering algorithm for wireless sensor network. The algorithm is fully decentralized. The proposed algorithm is inspired by simulating flow in the network. The algorithm clusters the nodes that are at distance of at most k hops from the cluster head while balance the energy dissipation among sensor nodes. By grouping sensor nodes into clusters, the performance of the mobile network, such as power dissipation, can be optimized. The preliminary simulation results show that the proposed algorithm is efficient. Introduction In wireless sensor network, the energy dissipation e to transmit a packet to a receiver at distance d can be estimated by formula [5]: 1 2 c tx e k d k d = × + × , where c( 2 4 c ≤ ≤ ) and k are constants for a specific wireless system. To receive a data packet at distance d the energy can be estimated by formula: 1 rx e k d = × . From those two formulas, we can see that: – The energy used to transmit messages are larger than to receive messages. – The energy used is greatly depend on the distance between the source and the destination of the communication channel. To reduce the energy overhead and prolong the lifetime of the whole sensor network, itwas proposed to partition the network nodes into clusters [5,7,8]. Suppose thereis centroid node in each cluster that are closest to all of the clustered nodeswithin the same cluster, we call this node cluster head. In our study, we propose a flow simulation based algorithm to clusterthe wireless sensor node. The algorithm is inspired by [1] and [4]. The rest of thepaper is organized as follows. First, the paper describes the algorithms of flow simulation,thenprovides the details about the clustering algorithm in wireless sensornetwork. Finally, a preliminary simulation about this algorithm will be presented.