Amp-Phi: A CSI-Based Indoor Positioning System.
Tao-Yun Zhou,Bao-Wang Lian,Yi Zhang,Sen Liu
DOI: https://doi.org/10.1142/s0218001418580053
IF: 1.261
2018-01-01
International Journal of Pattern Recognition and Artificial Intelligence
Abstract:With rapid growth in the demand of location-based services (LBS) in indoor environments, localizations based on fingerprinting have attracted significant interest due to their convenience. Until now, most such methods were based on received signal strength indicator (RSSI), which is vulnerable to non-line-of-sight (NLOS). In order to realize high-precision indoor positioning, we propose a channel state information (CSI)-based Amp-Phi indoor-positioning system which exploits the amplitude and phase information of CSI at the same time to establish a fingerprinting database. Firstly, according to the characteristics of the raw CSI information collected at different positions under different environments, we build an NLOS mitigation model and a phase error mitigation model, respectively, to process the amplitude and phase of CSI. Secondly, we analyze the statistical characteristics of CSI carefully, including maximum, minimum, mean and variance. After being processed with the models, the CSI features can be used to distinguish different positions clearly, which provides a theoretical basis for the indoor positioning based on fingerprinting. Finally, we build a fingerprinting database based on the features of amplitude and phase, realize to locate the targets position with the K-Nearest Neighbor (KNN) matching algorithm. Experiments implemented in different situations show that Amp-Pi system is reliable and robust, whose position accuracy is higher than that of PhaseFi, Horus and machine learning (ML) systems under the same condition. It can be used in many scenarios, such as the localization of robots in our daily life, by doctors or patients in the hospital, for people localization in large supermarkets or museums and so on.