The Research Of Data Fusion Algorithm And Model Based On Wireless Sensor Network

The Wireless Sensor Network is a monitoring network which is made of large numbers of little sensors and was set up through wireless communication technology. Due to the limited node’s energy, the WSN can only finish its monitoring works within a short lifetime. For the more accurate and efficient acquisition and transmission to the data from the sensors in the network, we use data fusion technology into the WSN. The data fusion technology will get all the data from networks and then de lete the redundant information, so the remaining information is important, in this way we will reach the point of saving energy. In this paper, we search the S-LEAC H routing protocol and the improved BP neural network fusion algorithm based on genetic algorithm to build the data fusion model. The main work and content is as follow:(1) Use the data fusion thought into the WSN and research the relevant knowledge of the WSN and data fusion technology. Then we analysis the common routing pro tocol and data fusion algorithm of the WSN.(2) Based on the insufficient way on the selection mechanism and transmission mode of the cluster heads on the traditional protocol, we then put forward a S-LEACH algorithm based on the clustering thought in the WSN. This algorithm implement the cluster heads selection in the heterogeneous network by thinking of the remaining node’s energy, node’s location and node’s density.And use the single jump and jump way based on the hierarchy tree in the transmission of cluster heads and gathering nodes. Then make a simulation on this algorithm, from the simulation result we can see that the S-LEACH algorithm greatly reduce the energy of the network and also improve the life cycle of the network.(3) As the rate of convergence for BP neural network fusion algorithm is slow, and the average error of the fusion result is bigger, we then put forward the improved BP neural network fusion algorithm based on genetic algorithm. The improved algorithm use the optimal initial weights and threshold selection in the genetic algorithm to the BP neural network, then improve the rate of convergence. The BP neural network use the momentum term improvements to cancel the oscillation phenomenon. Then make a simulation of this algorithm, the result show that the execution efficiency of the improved BP fusion algorithm has improved a lot.(4) Through the research of S-LEACH routing algorithm and the improved neural network fusion algorithm, we build the data fusion model of WSN. In this paper, we also build a model of the first leve l fusion in the clusters and the secondary level fusion between the clusters based on the clustering feature.