Asia-Pacific Network Operations and Management Symposium
Comparison of NVA and RLOWESS Algorithms in Indoor Positioning System
Hung-Huan Liu, Wei-Chien Chen, Pin-Chun Hsu,
PDF download (440.7KB)
In the Wi-Fi-based indoor positioning system, by using signals of Wi-Fi access points (APs) but without the position of APs, the scene analysis method has better positioning accuracy. In the previous study, we proposed the quick radio fingerprint collection (QRFC) and neighboring vertices averaging (NVA) algorithm as a way to collect radio fingerprints. In this study, we compared NVA and RLOWESS, a well-known filtering algorithm from the point of view of positioning accuracy. A cluster AP problem which causing large positioning error when using Euclidean distance formula to estimate the position in a corridor is also discussed. From our experiment, positioning accuracy of NVA and RLOWESS are similar, which evidence the NVA is usable and can be used as a smoothing method of radio fingerprint.