Research And Implementation Of WLAN Indoor Positioning Technology Based On Location Fingerprinting

In recent years, much attention has been paid to the development of new indoor location service. Besides WLAN deployment is widespread, plenty of resources and low cost, intelligent terminal can access and capture the receive signal strength easily, thus the WLAN indoor positioning method based on location fingerprinting is a potential method and attracts the attention of enterprise institute and numerous scholars colleges. For solving the problem of low position fingerprint accuracy caused by RSS instability and the problem of limited positioning efficiency caused by fingerprint data volume growth rapidly, this paper does the research in detail and carries out a prototype system. The research content is shown as below:At first in RSS property analysis, all RSS samples were collected in teaching building, which were analyzed in aspect of the time correlation and spatial correlation. Respectively, in the aspects of RSS time distribution, the AP signal correlation and RSS sampling number correlation between illustrates the RSS, the strong time-varying volatility had been indicated. In the aspect of RSS space distribution, detect the AP number and RSS toward the correlation the RSS was explained a certain but not stable ability to depict the space.Secondly, this paper put forward a kind of indoor positioning mechanism based on RSS-Probability Fingerprint, which adopts RSS probability distribution as the fingerprint and effectively improves the Radio Map characterization ability. At the meantime, according to the similarity measurements of different kind of location fingerprint are different in this paper, the Bhattacharyya distance is used as similarity measurement value and the metric of NN algorithm to calculate the target location.Thirdly, if the entire search space fully participate in online phase, computational complexity will be increased and low-efficiency, so a clustering method based on AP weight was presented in this paper. The weight of each AP in each reference point was calculated to divide reference points into clusters, which was verified to narrow the scope of search space and presented an effective positioning performance.Finally, this paper applied the above algorithm to design and implement a WLAN indoor positioning prototype system(LocNeedle) based on location fingerprinting, which was applied in real scenario and achieved acceptable positioning accuracy and high efficiency.