Indoor Positioning System Design And Implementation Based On Android

At present, the applications of Location Based Service(LBS) are throughout all aspects of people life and production; it has a huge market prospects and development space. T he location technology that is closely related to LBS has also been developed rapidly. Reliable indoor positioning technology is an important prerequisite for achieving LBS. Especially the popularization and application of the Internet technology and the Android platform, provide technical advantage for the design of indoor poisoning.First indoor positioning technology is studied in this paper, include indoor and outdoor positioning, and the research status at home and abroad is analyzed, the characteristics of the various positioning technology are analyzed, the classification of the positioning technology is researched, the Android development platform is introduced, the relevant technology involved are studied. Then research is focus on the research of location fingerprint positioning algorithm, including nearest neighbor(NN), weighted KNN algorithm, K neighbor method, the positioning method based on compression perception, and the important algorithm based on KNN and least squares support vector machine(LSSVM), and simulation experiments are conducted to compare the performance of several methods.Next, this paper designed and implemented indoor positioning system based on the Android platform. T his system collect Wi Fi signal with the mobile terminal, the terminal contains Android system, generate fingerprint database by computer terminal, by the client and the server interaction to achieve indoor positioning. When users have positioning demand, sent request signal, read the signal strength from different wireless router, the server use the combination of the nearest neighbor method and least squares support vector machine to locate, and the positioning information return to mobile terminal. After the system test, the average error of the positioning system designed in this paper is 1.8 m; the maximum error is 5 m, the accuracy within 3 m is more than 80%.