A novel approach for fingerprint positioning based on spatial diversity

Recently, fingerprint positioning systems for urban areas using the existing wireless-LAN (WLAN) play important role in ubiquitous applications. The most challenge in fingerprint positioning is how to reduce the number of training samples and high time consumption with no significant degradation in location accuracy. To deal with this problem, we propose a micro-cell-based radio map construction method based on spatial correlation of collected signal samples and spatial diversity. We present how to apply these properties to improve positioning accuracy and reduce time consumption of constructing a metropolitan-scale radio map. To further enhance the positioning accuracy and reduce latency for real-time positioning, we propose an unequal weighting fast matching algorithm to exploit temporal-spatial diversity and spatial continuity of the objects' movement trajectory. The practical implementation of positioning mobile device at different speeds in urban environments is described and conducted. Experimental results show that our algorithms can significantly improve the performance of WLAN-based fingerprint positioning system.