A Robust RSSI Fingerprint Localization Method in Wireless Local Area Network

The Received Signal Strength(RSSI) based location fingerprint positioning technology under the wireless local area network(WLAN) has become a research hotspot. For indoor positioning, using wireless networks for location calculation is cost effective and easy to deploy with high positioning accuracy. However, there are still some problems in the indoor fingerprint positioning technology based on RSSI signal: in the offline stage, on the one hand, the indoor environment is complex and has undergone changes, the received signal has spike noise during the acquisition process, which will make the reliability of the collected fingerprint collection change over time. As a result, it is difficult for users to detect the RSSI signal transmitted by the Access Point(AP) when they are online, which reduces the accuracy of indoor fingerprint positioning. It is also different, and even the status of some APs will change when they are online, reducing the robustness of fingerprint positioning. In the online stage, the traditional fingerprint positioning method needs to traverse the entire database when calculating the coordinates in the positioning stage, which in turn brings a lot of redundant calculation overhead and mismatches, decreasing positioning accuracy. Therefore, this paper designs a prototype system DeLoc for denoising. An AP entropy based on Gaussian detection is proposed for the offline selection and online filtering of APs. The average positioning error of DeLoc under different grid sizes and different AP numbers is verified by simulation experiments, and compared with traditional methods, DeLoc, designed in the paper, is robust and stable.