Novel indoor fingerprinting method based on RSS sequence matching

Abstract

To estimate the location of a mobile user in an indoor environment, many researchers have been developing fingerprinting methods. Fingerprinting is one of the most widely used method for indoor positioning because this system can be built using already installed infrastructure and users can use this system with just their smartphones. However, fingerprinting is highly dependent on the quality of the infrastructure, it is difficult to always provide users with accurate locations. For example, if a very small number of access points or beacons are installed, the performance of fingerprinting is significantly degraded. In this study, we propose a novel fingerprinting method based on RSS sequence matching. The main characteristic of the proposed RSS sequence matching method is that the user mask to be compared with the radio map is created using the user trajectory and an accumulated RSS vector. Thus, an RSS spatial pattern of the indoor environment that the user has just moved is created. The radio map is created by storing the RSS distribution of each beacon in a 2D space. In the position-estimation step, the position with the highest correlation with the user mask was found on the radio map. To verify the performance of the proposed method, a field test was conducted using four Bluetooth low energy (BLE) beacons, and an average position error of less than 2 m was obtained.