| Wireless Sensor Networks(WSN) is composed of a large number of sensor nodes by the self-organized manner. and it is data-centric. The deployment of WSN does not require infrastructure, and laying lines.Nodes can form a network by the self-organized manner, which gives WSN a broad allocation space.WSN is energy constrained and energy factors determine the lifetime of WSN. Therefore in addition to providing the correct data forwarding path, routing protocol of WSN also need to consider the effectiveness of energy, which is including the effective utilization of energy and energy balance. Clustering routing protocol of WSN can effectively reduce the energy consumption of the network, but it can also cause the energy consumption un-balance. Because the cluster head is not only responsible for establishing the cluster, controlling cluster communication and data fusion of the cluster, but also is responsible for communicating between the sink node and the cluster.If the communication between the cluster and the sink node is mostly by multi-hop, then the cluster head which close to sink node also need to forward data from other clusters to the sink node. So the energy consumption of the cluster head is much more than energy consumption of member nodes.This paper study the un- balanced energy consumption referred above, focusing on clustering routing protocol of WSN, proposing Centralized-Uneven Clustering Routing Protocol Based on Residual Energy and Communication Cost(C-EEUC) and Multi Cluster-Head clustering routing protocol on the base of EEUC.On the clustering phase, C-EEUC takes all nodes as the candidate cluster-head, and defines the weight matrix P, element of P P[i,j] represents the appropriate level of node i as cluster head of node j, the element combines the communication cost between cluster-head and member node and node residual energy, defines the weight of node i is the sum of i-th row of P. Each time when selecting a final cluster-head from the set of candidate cluseter-head, the node has the largest weight in the set of candidate cluster-head. Other candidate cluster-head within the competitive range of the node would give up the competition, and then update the set of candidate cluster-head. The process continues until the set becomes empty. There are two groups of experimentals carried out in this paper. The first group is the experiment that 100 sensor nodes randomly deployed in a range of 100m×100m,and the coordinate for sink node is(150m,50m);the second group is the experiment that 200 sensor nodes randomly deployed in a range of 200m×200m,and the coordinate for the sink node is(250m,100m). The first group experiments,in C-EEUC the first node died at 917 th rounds, 20% nodes died at 1742 th rounds, 50% nodes died at 2019 th rounds; in EEUC, the first node died at 695 th rounds,20% node died at 1071 th rounds, 50% node died at 1522 th rounds. The second group experiments, in C-EEUC the first node died at 126 th rounds, 20% nodes died at 607 th rounds, 50% nodes died at 1031 th rounds; in EEUC, the first node died at287 th rounds, 20% node died at 561 th rounds, 50% node died at 1031 th rounds. So C-EEUC extend the network lifetime.In the Multi Cluster-Head clustering routing protocol, in which a cluster have two cluster head, namely the inner-cluster head and the inter-cluster head. The inner-cluster head is responsible for fusing data from member nodes and forwarding it to inter-cluster head. Inter-cluster head is responsible for establishing cluster, forward data from others clusters to sink node. In the end, this paper takes coal mine as the background, assuming the convergence of the network is a 500m×10m roadway. In Multi Cluster-Head,the first node died at 190 th rounds, 20% nodes died at 1535 th rounds, 50% nodes died at 1875 th rounds; In EEUC, the first node died at 221 th rounds, the 20% nodes died at 1341 th rounds, the 50% nodes died at 1433 th rounds. The experimental results show that the Multi Cluster-Head outperforms better than EEUC in prolonging network lifetime. |
PDF Full Text Request