A UWB indoor positioning method based on triple filtering fingerprint matching algorithm

In recent years, ultra-wideband (UWB) has gradually become a research hot spot in the field of indoor positioning because of its various advantages. Although UWB has such excellent performance in normal environments, the non-line-of-sight propagation of signals in complex indoor environments and the multi-path effect caused by obstacles will affect its positioning accuracy. To solve this problem, we use the fingerprint positioning method and optimize the previously commonly used k-nearest neighbor algorithm in the online matching phase. In this paper, we proposed a triple-filtered k-nearest neighbor algorithm based on sample distance weighting (TFWKNN). Experimental results show that the mean calculation error of TFWKNN is 63.4% less than the k-nearest algorithm, and the proposed algorithm has better prediction stability than other commonly used machine learning regression algorithms.