The Research Of Information Fusion Algorithm Based On Wireless Sensor Network

Wireless sensor networks integrate sensor technology, embedded computing technology and wireless communication technology. Because of their wide applications, wireless sensor networks were considered as one of the most important technologies in the 21st century. At the same time,, the sensor's energy, storage and capacity were limited, so information fusion technology based on wireless sensor networks caused a great concern.Information fusion technology, which first appeared in the 1970s in the 20th century, was a highly integrated multi-disciplinary research field. In 1980s information fusion became an important topic in the field of high-tech research and reached rapid development in military. In this paper, based on the characteristics of wireless sensor networks, we gave a distributed hierarchical Kalman filtering information fusion algorithm and a heterogeneous RBF neural network information algorithm. Main work and content are as follows:(1) The basic conceptions,features and applications of information fusion and wireless sensor networks were analyzed in this chapter. In order to solve the limited capacity, enhance and improve the accuracy of the information collection, the idea of information fusion was used in wireless sensor networks.(2) The article analyzed the wireless sensor network's structure, the level of information fusion,the existing information fusion based on network layer and application layer. According to TEEN routing protocol we established a hierarchical level of information fusion model(3) A distributed hierarchical information fusion algorithm was proposed in wireless sensor network. In the source, the Kalman filtering and signal processing methods were used to reduce the power consumption and data collision. In the sink, the dynamic weighted fusion algorithm based on linear minimum variance sense was adopted. This method could minimize the overall power consumption and the probability of collision in the network as well. A simple example was performed to verify the effectivity and reliability of this method.(4) A heterogeneous RBF neural network information fusion algorithm based on wireless sensor networks was proposed. RBF neural network had a good network and a powerful real-time parallel processing capability. We used heterogeneous RBF Neural Network information fusion for cluster or sink. In cluster or sink we only transmit the results of fusion to reduce the amount of data transmission, and it could also reduce network congestion and conflict. A simple example was performed to verify the effectivity and reliability of this method.