Fingerprinting Based Indoor Localization Considering the Dynamic Nature of Wi-Fi Signals

Abstract

Current localization techniques in the outdoors cannot work well in indoors. The Wi-Fi fingerprinting technique is an emerging localization technique for indoor environments. However, in this technique, the dynamic nature of WiFi signals affects the accuracy of the measurements. In this paper, we use the affinity propagation clustering method to decrease the computation complexity in location estimation. Then, we use the least variance of Received Signal Strength (RSS) measured among Access Points (APs) in each cluster. Also, we assign lower weights to alter APs for each point in a cluster, to represent the level of similarity to Test Point (TP) by considering the dynamic nature of signals in indoor environments. A method for updating the radio map and improving the results is then proposed to decrease the cost of constructing the radio map. Simulation results show that the proposed method has 22.5% improvement in average in localization results, considering one altering AP in the layout, compared to the case when only RSS subset sampling is considered for localization because of altering APs.