An enhanced indoor positioning method based on Wi-fi RSS fingerprinting

Abstract

In WiFi-based indoor positioning, the received signal strength (RSS) measurements are commonly used to estimate the mobile user location. However, these measurements significantly fluctuate over time and are susceptible to human movement, multipath and Non-Line-of-Sight (NLOS) propagation, which reduce the location accuracy. In this paper, an enhanced positioning method based on the nearest neighbor algorithm is proposed. The set of the RSS samples recorded from several Access Points (APs) is used rather than their average, for reducing the location errors introduced by the RSS variations and the multipath problem. The proposed algorithm, named the Nearest Kth Nearest Neighbor (NK-NN) is experimentally evaluated and compared to other powerful methods. The results show that the proposed method outperforms these methods.