Populating virtual access points for localization in sparse wireless networks

Even though many GPS-free localization algorithms have been proposed to provide reliable positions with high levels of accuracy, most of them, such as centroid, fingerprint, or their variant versions, should require three or more Access Points (APs) to provide sufficient signals for the device. Thus when enough number of APs do not exist in the proximity of the device, the level of position accuracy is deteriorated. In this paper, we present a localization framework of populating virtual access points for localization in sparse wireless networks that can estimate the device position in the environment where the device cannot obtain GPS signal nor sufficient number of AP beacon frames. The proposed framework first gathers beacon frames of adjacent APs periodically and stores the frame information into a device database, then it chooses an adequate number of adjacent APs from the database, and finally it performs a localization algorithm with the chosen APs. To verify the accuracy and robustness of the proposed system, we conducted an experiment in an indoor environment where the deficient number of APs provide their own information, and compared the proposed framework with the original localization algorithm. The results indicate that the proposed framework can pinpoint the current location with acceptable errors in wireless networks with sparse coverage.