Research Of Data Fusion Technology Based On Wireless Sensor Network

Wireless sensor network is constituted by abundance of cheap, low-power wireless communication nodes in monitoring the region’s, it is used to cooperative sensing, collecting and processing information within the network coverage area, and send the information to monitors. Data fusion system can process data fusion effectively, reduce redundant information transmission, which greatly reduce the energy consumption and increase the life time of network. According to poor accuracy of these algorithms, this paper works as follows:This paper puts forward several basic of data fusion algorithm(recursive estimation data fusion algorithm, adaptive weighted data fusion algorithm and Arithmetic average data fusion algorithm)with some analysis and testing, finding their poor accuracy. Then it raises a algorithm of Clustering-based Data Fusion and its improved algorithm. By means of the tool Matlab, we get the quality of the two algorithm. Experiments show that the proposed improved clustering algorithm can reduce the amount of data transmission and get better result in the final data precision.With the disadvantage of low efficiency and poor accuracy of manual measurement point by point in granary storage in our country, an intelligent granary monitoring system based on wireless sensor network is proposed in order to reduce the losses in granary storage, combined with embedded, wireless communication, sensor and embedded network technology. The system collects the data about temperature, humidity and the pests through wire sensor network, and transits these data to remote monitoring center by GPRS. The monitoring management program accomplishes the tasks of the real-time inquiry and gives an alarm if there are any over-standard parameters. It can not only guarantee the character and quality of granary, but also reduce the intensity of labor, so as to achieve real-time operation and intelligence of granary monitoring system.This paper brings with some solutions about hardware design, including the design of sensor node, sink node and the selection of sensor module. To realize the sink node, I use ARM7(LPC2106) as the CPU, FM25H20as the external memory, short-range technology for wireless communication, JTAG as the debugging interface and so on. In software program, it puts forward program flow chart of sensor node, the main program and Network coordinator. To design the monitoring platform, I use SQLserver2005and Visual C++. Finally we can realize a set of reliable monitoring plan to the warehouse environment for monitoring.