Research Of Sensor Data Fusion And Fault Tolerant In Grid

Sensor grid is a new research area sprung up in recent years; it is a distributed system integrated wireless sensor network and the grid, which has the advantages of wireless sensor network and grid. In sensor grid, the grid can take advantage of computing power, communication bandwidth, and storage space to fuse sensor data for higher-value data; the grid can also access real physical information by wireless sensor network. Wireless sensor network is different from the grid, this also causes new questions, especially increased the difficulty of data fusion and system fault tolerant, therefore, the study sensor grid faces many difficulties and challenges.The data fusion is actually the process of using the minimal resource consumption to get the highest value of fusion results, which should not only satisfy the users, but not reduce resource consumption. It seems to be a critical and difficult issue to select the optimal fusion routing in the sensor grid. Among the data fusion algorithm, some works focus on energy consumption, some works focus on deadline constrained scheduling, but few works consider both energy and deadline as constraints. The grid is a dynamic, unstable environment, various errors may occur, in this thesis, we propose a data aggregation algorithm based on grid and adaptive genetic algorithm for wireless senor networks, which also involves the fault tolerant mechanism.The leading research and innovation of this thesis are as follow:Firstly, it analyzes the characteristics of sensor grid data fusion and the necessity of fault tolerant. A fault tolerant framework is proposed based on the characteristics of sensor grid, the feature of which are:on the one hand, errors are effective treated in the task processing; on the other hand, the data sensed by sensor node is the full integration, fusion results provide strong support for the user decision-making.Secondly, this thesis proposes a data aggregation algorithm based on grid and adaptive genetic algorithm for wireless sensor network to limited wireless sensor network energy issues. The proposed algorithm firstly runs the adaptive genetic algorithm in wireless sensor networks, finding the optimal gateway nodes, then mobile agents with minimal communication cost access gateway nodes and merge data, the algorithm aim is to obtain the optimal fusion results with minimal energy consumption, and prolong the network life time.Thirdly, a variety of errors will be produced inevitably in the task processing, in order to deal with burst errors, improve the quality of service, this thesis proposes a adaptive fault tolerance algorithm, which is the combination of the retry and alternative resource, it can be adaptively choose the appropriate treatment methods according to the error kind, to ensure that the tasks will be process successfully.Lastly, this thesis selects NS2 as the simulation platform, the proposed algorithm is simulated in sensor grid environment. The data aggregation algorithm based on grid and adaptive genetic algorithm for wireless sensor network is compared with an existing fusion algorithm, the simulation suggests that the proposed algorithm has good performance. In addition, the proposed adaptive fault tolerance algorithm is compared with four classic fault tolerance algorithms, the simulation suggests that the adaptive fault tolerance algorithm has better performance.This thesis is supported by National Natural Science Foundation of China (NSF) under grant (No.60773211),Open Fund of the State Key Laboratory of Software Development Environment (No:SKLSDE-2009KF-2-02),Program for New Century Excellent Talents in University, China under grant (NCET-08-0806),Municipality science and technology in the Wuhan offends pass item (No:201010621207),Fok Ying-Tong Education Foundation for Young Teachers in Higher Education Institutions of China (No:121067), and Independent Innovation Foundation of Wuhan University of Technology(No:2010-ZY-JS-004).