|Abstract:||With the popularity and development of wireless communication technology, location-based services extended from outdoor to indoor. And with the extensive deployment of AP, WLAN fingerprint localization has become a hot research topic. Firstly, this paper analyzes the problems of the off-line fingerprint altas in WLAN fingerprint localization, such as the inefficient use of fingerprint, the heavy workload in collecting a large quantity of signals, hard to update the fingerprint altas and so on. Then the paper presents a method to improve an RSS sampling information utilization efficiency, collecting directional fingerprint signal, building the interpolation location fingerprint altas, and making a way to pretreat it. In the comparison of different interpolation algorithms, we get to know that the neural network can better simulate the real environment. After applying data mining method to clustering of the fingerprint, and the machine learning method are used to build the virtual RSS data, and made the fingerprint altas organized in a more efficient way. Finally, three different comparison experiments were carried out. The first is to compare the directional and directionless fingerprint altas, the second is to compare different virtual point establishment method’s performance for the effectiveness and improvement in real localization, and the third is to analysis the virtual point(VP) and sensitive virtual point(SVP)’s performance. The experiments’ results show that the method can significantly improve the utilization efficiency of RSS samples, reduce workload of RSS signal collection. The pre-organized fingerprint altas can accelerate the convergence of the online positioning algorithm and to better estimate the user's actual location. In terms of positioning error, pretreatment fingerprint altas can improve the positioning accuracy of 35% (from 1.89 m to 1.22 m), and the offline processing does not increase online positioning time and complexity, so the algorithm has good application value.|
Proceedings of the 29th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2016)
September 12 - 16, 2016
Oregon Convention Center
|Pages:||1831 - 1837|
|Cite this article:||
Yu, Min, Guo, Kaixuan, Li, YaQing, Zeng, Zhi, Tang, Rui, Guo, Hang, "The Improvement of Location Fingerprint Altas of WLAN Indoor Positioning Technology," Proceedings of the 29th International Technical Meeting of the Satellite Division of The Institute of Navigation (ION GNSS+ 2016), Portland, Oregon, September 2016, pp. 1831-1837.
ION Members/Non-Members: 1 Download Credit