A VLC-Based 3-D Indoor Positioning System Using Fingerprinting and K-Nearest Neighbor

In this paper, a three-dimensional (3-D) positioning system using fingerprinting and K-Nearest neighbor is proposed. The implementation of the proposed positioning system based on visible light communication (VLC) is discussed. On-off keying (OOK) modulation, Manchester coding and time division multiplexing (TDM) are utilized to obtain both lossless data and accurate received signal strength indications (RSSIs) without sacrificing the comfort of indoor illumination. The proposed 3-D positioning algorithm is conducted in three processes. Firstly, we quickly traverse all sampling points in the fingerprint database to find the nearest neighbor and choose candidate sampling points around the nearest neighbor. Secondly, we conduct an iterative search among candidate sampling points to find K-Nearest neighbors. Finally, we obtain estimated coordinates using weighted average method. Experimental result shows that, the proposed system possesses fast location and high accuracy.