Fingerprint Based Positioning Algorithm In WLAN Indoor Localization

As a large number of smart terminals flourish and indoor positioning technologydevelop continuously, the demand for location services is more and more. The WLANbased indoor localization is widely used for its advantages of high positioning accuracyand requiring no additional equipment and manpower material resources. Thefingerprint localization algorithm is widely used as the WLAN indoor positioningalgorithm. The positioning process can be mainly divided into offline and online phases.This paper focuses on the fingerprint localization algorithm and improved algorithmswhich are aimed at practical problems and improving positioning accuracy are proposedin this paper.The progress of science and technology makes a great diversity of mobileterminals, which can include various notebook computers with different brands, typesand configurations and also the various types of smart phones. Electronic products arechanging people’s life and become an indispensable part of people’s life. However,indoor positioning often appears such problem that the mobile terminals users use are ofvarious kinds and are often incompatible with the terminal used to create the databaseoffline, which will lead to a large matching error during the online phase and seriouslyaffect the positioning accuracy and even failure. There are not only large resourcesinvested in the process but also the existing methods not conducive to ensure real-timepositioning, thus the improvement of location accuracy is limited. Due to theseproblems, this paper proposes an improved algorithm based on KNN with goodapplicability. At the same time the location algorithm based on compressive sensingwhich is in the initial stage will be popularized and applicated in this paper. And thispaper also compares the positioning performance of different reconstruction algorithmand its improved algorithm.First, this paper has carried on a summary about the fingerprint localizationalgorithm and a detail introduction of the positioning principle of each algorithm. Dueto the device diversity, this paper introduces several kinds of typical solution to theproblem which provides the theoretical foundation for subsequent research.Secondly, on the one hand this paper analyzes signal strength characteristics andthe nature reason that causes the device diversity. Then we put forward the improvedKNN algorithm based on cosine similarity and analyze the positioning mechanismagainst device diversity. On the other hand, this paper further studies the locationalgorithm based on compressive sensing, and different reconstruction algorithms areused for the comparison of the positioning performance. We conclude that SL0 reconstruction algorithm is superior to other algorithms, and an improved algorithm ofSL0based on conjugate gradient method is proposed to further improve the positioningprecision.Finally, we use the real measured data to conduct the experiments which furtherverify the validity of the algorithm. Experiment results show that the improved KNNalgorithm ensures the real-time positioning and the positioning accuracy is improved atthe same time. Further more, the improved SL0algorithm is better than the traditionalalgorithms.