The Research Of Data Fusion In Wireless Sensor Network

Wireless sensor networks have an important role in the field of computer networks, as a set of information collection, information transmission, information processing in the integrated intelligent information systems, which can be used for military, space exploration, health care, environmental monitoring, traffic management, urban traffic management, etc.First, the system structures of the WSN networks were studied and discussed in detail. Then a sort of fresh system model was put forward. The formerly system architecture offers the network transmission support for various kinds of application program. Nodes only for data transport instead of processing the data. But the layered architecture is focus on sensing data, which the aim is to obtain more accurate characteristic information of the perceived targets. So the maximum extent possibly streamlines the traditional secondary functions, and combined the model network layer and transport layer into a transport layer. Between adjacent layers achieves a certain degree of self-adaptive adjustment, while data fusion techniques applying to the stack, this can reduce energy consumption in the process of network traffic and prolong the life of the network.In-depth study of information fusion in wireless sensor networks theory of knowledge, data fusion theory and sensor networks, combining the characteristics of the various hierarchical done analysis and research, In order to solve the energy limited wireless sensor networks to enhance the accuracy of the information collected and improve the efficiency of information gathering and so on. Relying on routing protocol established structural model of distributed information fusion for the final chapter.The data fusion approach in wireless sensor networks is studied deeply,a kind of data fusion algorithm based on self-adapting weight are presented. According to the size of the weights, the sensor data of cluster node carries out a reasonable assessment and get more accurate results after treatment. The algorithm can effectively solve the conflict in network information, reducing redundancy of data within the network; simulation results show that it has highly efficient implementation and scalability.Finally, a robust form of information fusion kalman filter algorithm is raised, studying the structural model of distributed information fusion in the WSN. This fusion algorithm uses in the distributed network cluster dealing with the information fusion. Only transmitting convergence the results, reducing the amount of data transmission, while it reduces network conflicts and congestion. Simulation results show that this algorithm having lower communication cost than the traditional centralized fusion algorithm has better results. The lifetime of networks gets long.