Research And Design Of Indoor Location System Based On WiFi

With the development of mobile Internet and the popularity of mobile smart terminals, the application of wireless location technology has been a brand-new era. Especially in complicated indoor environment, we often need to know the position of smart terminal or the holder. The rapid development of WiFi technology and high coverage of WiFi hotspot indoor makes it possible to locate by widespread WiFi network indoor. Thus, it makes up the drawbacks of GPS in indoor location.An indoor positioning technology solution based on RSSI is presented in this paper, which focuses on three components of the positioning system. Firstly, data acquisition and processing. In view of the RSSI data may have big errors, so data should be smoothed. Gaussian filter is used to reduce the impact of accidental error of RSSI data by analysing and comparing of two flitering models. Secondly, distance estimation. Several RSSI ranging models in common use are analyzed in detail, and then the parameters are optimized based on maximum likelihood estimation method in the simplified model of empirical model. The third component is corrdinate computation. The multianalysis has contrasted NN and KNN algorithm of RADAR system. On this basis, an improved weighted-based algorithm called NNL is put forward which subjoins different weight values to anchor nodes in order to improve the position accuracy. Then, this paper shows an improved Taylor series expansion localization algorithm based on NNL algorithm by combining the advantages of NNL and the Taylor series expansion localization algorithm, named NNL-Taylor. Taylor series expansion localization algorithm is initialized by localization result of NNL as initial iteration value. When reference nodes is less, a rather big error would occur for NNL. And when initial value error is large, Taylor series expansion localization algorithm may not work. NNL-Taylor algorithm overcomes the disadvantages mentioned above. The simulation result from matlab shows that the position accuracy of NNL is higher than NN and KNN. Because of the high accuracy of NNL-Taylor algorithm, the iteration number is reduced accordingly.This paper also shows design of a WiFi tracking tag which works well with stable performance. It can accurately collecte the RSSI data of nearby hotspots, and upload the various parameters from AP to PC server where they could be analyzed in order to get the final location result of WiFi tag. Moreover, it implementes the functions of data acquisition as well as network transmission for temperature, humidity and light intensity.