Nowadays, cities and towns all over the world are dotted with high-rise buildings, and floor positioning technology has gradually become the focus of indoor positioning research. Multi-level buildings with an atrium space structure mainly consist of shopping malls and hospital outpatient buildings. This kind of building is widely spread, and they usually have a lot of people and a high demand for location-based services. The specific spatial structure of the indoor atrium makes the propagation of wireless signals different, which leads to most of the wireless signal-based floor positioning methods not working well in such an environment. Aiming at this atrium structure environment, this paper proposes a high-precision multi-method adaptive weighted fusion algorithm based on Bluetooth Low Energy signals. In experiments, the real-time dynamic floor positioning accuracy reached 92.7%, while the floor positioning accuracy in the 5-s static state was as high as 100%. Besides, when using 3 s of test data rolled back, real-time effects could also be achieved, with an accuracy of 98.3%. This method can be used alone for real-time floor positioning, it is suitable for wireless fidelity signals, and it can provide a highly accurate initial floor reference for barometric or inertial methods. It could also be integrated with other methods to provide highly accurate floor locations for long periods of time. The fingerprint library acquisition in this method saves time and effort. The database volume is small and generalizable to a wide range of applications.
"Acta Geodaetica et Geophysica" – Springer Journals
Published: Jun 12, 2021
Keywords: Indoor positioning; Floor positioning; BLE signal; Wi-Fi signal; Multi-level atrium space