Research on Indoor Positioning Method Based on Terminal Clustering

In the dynamic indoor positioning environment, the traditional indoor positioning model will increase the positioning error and computational complexity with the increase of the number of positioning targets. This paper proposes a terminal clustering method based on the minimum Mahalanobis distance, which can provide good positioning services for multiple terminals at the same time and improve the positioning efficiency of the indoor positioning system. At the same time, two-level precoding is used to effectively overcome the inter cluster interference and intra cluster interference in the terminal cluster, and improve the positioning accuracy. In addition, the spatial sphere weighted centroid method based on the channel covariance matrix makes full use of the channel information between each node and the terminal, which is more targeted for the channel environment (including geomorphic features and surrounding buildings, etc.), so as to improve the quality of service.