Research on Indoor Visible Light Positioning Algorithm Based on K-means Clustering

In order to improve the indoor positioning accuracy and reduce the impact of a reflected light on the positioning performance, based on the analysis of the existing RSS positioning algorithm, an improved positioning algorithm based on K-means clustering is proposed. According to the indoor visible light communication system model and optical power distribution, 10*10 receiving points are uniformly selected on the receiving surface of 5m*5m. Through the collection of the received optical power of each receiving point, the K-means clustering algorithm is used to classify 100 receiving points: the first cluster of points is less affected by a reflected light, and the second cluster of points is greatly affected by a reflected light. The second cluster of receiving points is weighted after classifying, and the simulation results show that the algorithm significantly reduces the room edge positioning error, and the positioning performance of the entire room is improved by 28%.