A Novel Approach to NLOS Identification for UWB Positioning Based on Kernel Learning

Non-line-of-sight (NLOS) propagation is one of the key challenges in radio positioning. Significant attentions have been drawn to the identification and mitigation of the NLOS in recent years. This paper focuses on the identification of NLOS signals based on the feature attached to the signals themselves. We propose a novel NLOS identification technique employing K-Means Clustering (KMC) and kernel KMC methods, which are two of the most popular techniques in the field of data mining. Our work is based on the fact that the features used, i.e., the kurtosis, the mean excess delay spread, and the root-mean-square delay spread are well modeled by log-normal random variables. Numerical simulations are performed to demonstrate the validity of the methodology.