A New Method for Indoor Positioning Based on Integrating Wireless Local Area Network, Bluetooth Low Energy, and Inertial Sensors

Abstract

Increasing the accuracy of indoor positioning is still a challenging issue. In this paper, we propose a novel integration structure for indoor positioning using a wireless local area network, Bluetooth low energy beacons, and inertial sensors to increase the positioning accuracy. The main steps of this method are initial and relative positioning. Wireless local area network fingerprinting and database filtering using Bluetooth low energy are applied to calculate the initial location. Relative location is computed using inertial sensors data and the pedestrian dead reckoning method. In order to increase the accuracy of pedestrian dead reckoning, two sources of information, wireless local area network, and Bluetooth low energy are used. This new correction method is performed using a double Kalman filter. Extensive experiments were conducted in a smartphone and under two indoor environments. Our correction structure using a double Kalman filter outperforms previous pedestrian dead reckoning structures in terms of accuracy. Moreover, experimental results show our correction structure achieves an average accuracy of 1.7 meters.