Under The Environment Of Multi-source Data Fusion Skyline Query Application Research

The Internet of Things plays an important role in the new generation of information technology, and Wireless Sensor Network (WSN), as the key technology of the Internet of Things Application, is receiving more and more attention by experts and scholars. Currently, the features of WSN applications, such as high cost, data uncertainty and so on, largely restrict the widespread of WSN. Skyline Query is a new type of application based on WSN, which is widely used in greenhouse monitoring, forest’s fire preventing and other areas. It can meet users’need by getting information from a large number of sensor data. In this paper, we mainly research the methods to overcome the inherent shortcomings of WSN, such as the high implementation cost of WSN, low accuracy and reliability of sensor data, and high energy consumption problem in multi-subspace skyline querying and updating.At first, we introduce two key technologys of Internet of Things—WSN and RFID, and analyzes their advantages and disadvantages, then propose a new architecture called hybrid of ZigBee and RFID by combining advantages of the two technologies. This architecture can reduce the cost of WSN, and it has more transparent information flow and good locating ability. In addition, we analyze the multi-sensor data fusion principle, and puts forward a method:in the fisrt step, we apply Batch Estimate Theory on the multiple observations of single sensor, and in the second step, we use Random Set Theory to uniformly represent the uncertain information lile multi-sensor data and environment states, then in the third step, we adopt Gaussian Fitting Method to re-conceptualize multi-sensor as a global sensor, finally we use Likelihood Measurement Function to measure the matching degree between global sensor data and environment states for decision making. At last, we propose a query model called ES-CSC based on Extended Skyline (ES) and Compressed Skycube (CSC), and analyze updating algorithm on vary conditions when sensor frequencies are different. Experiment on simulated environment shows that the ES-CES model and updating algorithm can reduce energy consumption effectively and improve computational efficiency.