A Clustering Algorithm Based On Energy Prediction For Wireless Sensor Network

Wireless sensor networks collect the advantages of the embedded control technology,sensor technology, data processing and network communication technology. Through theuser’s control of wireless sensor networks work complete the information collection of workarea, environmental monitoring, and other tasks. Because of its low cost, high efficiency,small volume, it has a broad prospect in many applications.Wireless sensor nodes are powered by batteries with limited energy. When they work,the batteries can not be recharged. Under the circumstance that the quality of wireless sensornetworks can’t be affected, saving energy in wireless sensor nodes is the long-term security ofthe network. It’s also critical for wireless sensor network research. The researchers’experiment suggests that effective clustering of the network can significantly save energy andprolong network life cycle.This paper analyzes the significance and background of research subjects, and energyconsumption of wireless sensor network. It introduces the components of wireless sensornetwork briefly. On the basis of the above contents, it introduces some typical clusteringalgorithms. And then analyzes the advantages and disadvantages of these algorithms.After detailed studying of a variety of algorithms, this paper learns their advantages andimproves their disadvantages. It proposes CATSP algorithm. The algorithm takes fullconsideration of the node residual energy and node degree when the cluster heads rotate, sothat the node which remains more energy and node degree is elected to be cluster head node.Because the network data flow of the cluster head node is predicted, the cluster head node canbe full protected. It can avoid the early death of the cluster head node due to communicationoverload. Finally, the algorithm is carried out by NS-2simulation software. Simulation resultsshow that, CATSP algorithm can save more network energy be better able to balance thenetwork load, and more effectively extend the network time than LEACH algorithm.