Research On Energy Eficient KNN Query And Key Technologies In Wireless Sensor Networks

In recent years, man-made and natural disaster has been bringing great losses to the stability of society, human life and property. The relevant departments have established preventive strategy of security from the aspects of the system, preparedness, propaganda etc, but haven’t worked out proper solution to intelligently find hidden dangers or reduce losses brought by the security accidents from the technological point of view. Based on the good consideration of the intelligent selection system of emergency doors for fire and the intelligent alarm system of hidden danger in coal mines, the related theory and operating principle of wireless sensor are studied and analyzed. Namely, when the event occurs or the user request query of perception data, the wireless sensor network, regarded as a data centric database, will transmit the data that the user are interested in to the end-user through the efficient routing technology and data fusion technology, provide the above service for end-users, at the same time, balance the energy consumption, and then improve the life cycle of the wireless sensor network.In order to solve these challenging problems of the processing power, bandwidth, esp. energy restriction, the paper takes the k NN query technology as the main line, takes query technology, data fusion technology and dynamic network technology of the wireless sensor network as the theoretical basis, and focuses on the research and analysis of reducing data redundancy, resolving the “hole” problem, reducing energy consumption, and improving query accuracy rate, through the three stages of the k NN query boundary estimation, data collection, and data feedback.GDRk NN query technology of sensor nodes in different skewness distribution, DFi CI data fusion algorithm and energy balancing algorithm based on E-MCDS are proposed. The specific algorithm is given, the simulation experiments has been done and proved the consistency with the theoretical analysis and the specific algorithm. The main contribution of this thesis includes the following three aspects:(1) in order to improve the efficiency and accuracy of data query, the query technology of the route formed by the adjacent grid unit center- GDRk NN query algorithm is proposed based on the study of the existing data query algorithm of wireless sensor network, especially k NN query algorithm based on the line, at the same time, a large number of simulation experiments are carried out in the proposed algorithm in different skewness distribution and the number of sensors, the results shows that the GDRk NN method have good performance in the aspects of query accuracy, communication costs and reducing data redundancy.(2) In order to solve the problems caused by the direct perception data transmission of sensor nodes, such as the high data redundancy, serious energy waste, short net life, etc, the technology of multi-level & multi-aspect data information fusion based on the DFi CI algorithm is proposed, which include information coding, evaluation and combination in the process of related time-space perception data transmission, and it can reduce the transmission data of wireless sensor network. Simulation experiments shows that this method has good performance in reducing data redundancy and the energy consumption.(3) In order to prolong the life of wireless sensor network, the energy balance algorithm based on E-MCDS is puts forwards by researching and analyzing the energy consuming model and topology control of sensor node and wireless senor network, which can form a backbone network used to perceive the back-pass data of GDRk NN by increasing the node dump energy and the weight of distance factor with the sink node in dynamical topology, this method can reduce the network nodes energy consumption, and at the same time can achieve the balance of energy consumption of wireless sensor network.Finally, the author made a profound thinking on the shortcomings of the paper, put forward the prospect of wireless sensor network for future areas of research and application, such as semantic understanding perception data of and data mining etc.