| Wireless sensor network (WSN) is multiple-hop network self-organized with lots of low-cost sensors with ablities of sensing, data processing and wireless communication. Wireless sensor network configured automatically is highly valuable in both civil aspect and military aspect. Carrying out some tasks, the accomplishment of task depends on the network performance, such as life time of network, the balance of load and stability etc. All of the capabilities relates to network hiberarchy. Hiberarchy is mostly adopted in WSN. As WSN is established, the network is divided into some clusters. Every cluster is made up of a cluster head(CH) node and some cluster members. The cluster head manages to distribute resource in cluster and communication of clusters. Cluster members are responsible for data collecting and processing. Because clustering has great influence on network capability, the research on clustering algorithms in WSN is very important for the advance of network performance, in practical application as well. The thesis expatiates the research aiming at some clustering algorithms in Wireless Sensor Network, and the central achievements is shown as follows:1. In the paper three clustering algorithms in Ad Hoc network are introduced into wireless sensor network. One of these algorithms, which called Automatic On-demand Weighted clustering algorithm, is complex. In order to reduce the complexity of that one, a new algorithm is presented which adopted threshold estimation , instead of numerical value account. The new algorithm indicates the simpler operation, the better performance, and better load balance.2.Current clustering algorithms of network only select one cluster head in a cluster. When the cluster head dies out, the network will be reestablished, and stability of the network will be reduced. So a new " Double Cluster Heads " clustering mechanism is put forward. The mechanism brings two cluster heads: a formal cluster head;the other is spare cluster head.When the first dies out, the second will become formal cluster head. It dispenses with network re-establishment, reduces network updating frequency, and improves network stability. The mechanism may be introduce into existing algorithms. Compared with old algorithms, both stability and Head Fairness of wireless sensor network have been greatly improved after the new mechanism is introduced into existing clustering algorithms. |
