A Bluetooth Localization Algorithm Based on Map Path Calibration and Time Series Filtering

摘要

With the gradual progress of the urbanization process of modern society and the continuous improvement of residents' living standard, people's demand for IPS (Indoor Positioning System) is more urgent than that of the past, higher requirements for positioning accuracy, positioning robustness and positioning energy consumption are put forward. Low-power Bluetooth technology has the characteristics of low-cost, easy deployment and low power consumption, sup?porting all mainstream smartphone operating systems, and attracting extensive attention from academic and industry field. In this paper, aiming at the situation that current Bluetooth low energy positioning accuracy has large fluctuation under the circumstance of multipath propagation, external disturbance and other factors, and has poor user experience, a Bluetooth localization algorithm based on map path calibration and time series filtering is proposed (MCTF). This algorithm compre-hensively utilizes a post-processing method based on the average sliding window, time-series based spatial restraint post-processing method and path information based location calibration method, effectively eliminating the large fluctuation of positioning results caused by multipath propagation and other external wireless signal interference in complex indoor environment. By using map information, based on the fact that position cannot be transient, utilizing feedback filtering, positioning results can get path correction and direction correction respectively, solving the problem that the positioning coordinates jump between adjacent narrow aisles and jump back and forth on the same path during the positioning process. Extensive experiments performed in shopping mall, building office, exhibition center and other typical indoor environments demonstrate that the average positioning error of the proposed algorithm is within 1.5 m, which has good robustness and ensures the smoothness, accuracy and stability of the positioning result.