Research On Indoor Positioning Algorithm Based On Visible Light Communication

As one of the indoor wireless communication technologies,visible light communication(VLC)has gradually become a research hotspot in various countries due to its green,low-carbon,good confidentiality,no electromagnetic interference,and fast transmission rate.In recent years,all walks of life have increasingly demanded high-precision indoor positioning technology.As a result,the indoor positioning technology based on visible light communication came into being and has attracted widespread attention.Conducting a comprehensive research on the basic issues of the visible light indoor positioning system,such as the system model and channel characteristics,and discusses the influence of various factors in the indoor environment on the positioning performance under the NLOS channel.The RSS indoor visible light location algorithm based on K-means clustering and the location fingerprint location algorithm based on the combination of bilinear interpolation and K-means clustering are proposed.The specific work content of this paper is as follows:(1)The indoor VLC system model and channel model are established,and the illuminance distribution,received light power distribution and signal-to-noise ratio distribution of a 5m*5m*3m indoor room are simulated and analyzed.Several common indoor positioning algorithms and three commonly used indoor positioning accuracy evaluation indicators are introduced.(2)Aiming at the problem that RSS positioning accuracy is greatly affected by non-line-of-sight transmission,an RSS indoor visible light positioning algorithm based on K-means clustering is proposed.The clustering algorithm is used to screen out the points that are greatly affected by a reflection,and perform the weighting process improves the positioning accuracy of the edge area of the room,thereby improving the positioning performance of the entire receiving plane.(3)The location fingerprint method and three classic matching algorithms are introduced,including nearest neighbor-algorithm(NN),K nearest-neighbor algorithm(KNN),and weighted K nearest-neighbor algorithm(WKNN).Aiming at the problem of manual selection of K value in KNN and WKNN algorithms,a dynamic K proximity algorithm is studied,which improves the positioning performance by adaptively selecting a suitable K value,and performs simulation analysis under NLOS channel.Based on the dynamic WKNN algorithm,the influence of the LED layout and the number of LED on the positioning accuracy is explored to provide a reference for the actual LED deployment.(4)The influence of fingerprint density on the positioning performance is studied.In order to improve the positioning accuracy and reduce the computational complexity of the algorithm,a position fingerprint based on the combination of bilinear interpolation and K-means clustering algorithm is proposed.The core idea of the algorithm is clustering first and then interpolation.Finally,a matching algorithm is selected to achieve the final positioning.The simulation results verify the feasibility and superiority of the algorithm.