The Research On Data Fusion Of Wireless Sensor Network In The Grid

The Wireless Sensor Network (WSN) is composed of lots sensor nodes, which have the capability of sensing, computing and communicating. The sensor nodes are deployed in the sensor area and form a self-organization network. The WSN can be used in environment monitor, medical treatment, city traffic, and military applications.The widely deployed WSN can get lots of environment data and make people know more about the environment. But the data processing capability of WSN nodes is very limited, and the storage, communication capability of WSN nodes are also restricted, this makes the processing of sensor data, which are captured from nature quickly, more complex and out of time.The gird is a distributed computing technology, which is very popular in recent years. The grid can make the computing capability, database, and storage all together, to give a uniform grid service.The using of gird to do the WSN data fusion work has some advantages. Firstly, the huge computation capability can be fully used to do the fusion work and let the WSN data deeply processed. Secondly, the grid can integrate many WSN and give a consistent interface of WSN to the user. Thirdly, the integration also can shield the different data structure of every WSN and simplify the data utilization.The feature, utilization, system software, and the data fusion algorithms of grid and WSN are presented in the first part of this thesis. The next part of this thesis proposed the grid-WSN integration framework, and used it in the forest fire forecast. The integration framework is composed of the WSN integration layer, the feature fusion layer and decision fusion layer. The WSN data fusion algorithm and D-S evidence theory algorithm are used to form the feature fusion layer, and the probability judgment algorithm is the decision fusion layer algorithm.The WSN system software TinyOS and the grid software Globus Toolkit 4 are used to construct the experiment. The TOSSIM and TinyDB are used in the experiment to obtain the WSN data. The data is then transmitted to the data fusion services, which are built on the Globus Toolkit 4. By the using of integration framework, the data transmission quantity is greatly decreased and utilization improved.