Time‐varying long‐range navigation signal channel modelling based on Dirichlet process mixtures method

Abstract

AbstractA model of the LOng‐RAnge Navigation signal can provide a mathematical model for the development of related technology research. The propagation variations of the signal exhibit non‐stationary and non‐linear characteristics due to the time‐varying nature of the atmospheric dielectric constant in the channel. A Dirichlet process mixtures method is used by the authors to model the statistical properties of radio wave attenuation, and the Gibbs sampling method is used to flexibly estimate the parameters of the model. The model as a non‐parametric model can be applied to capture the dynamics of time‐varying signals with the parameters being adaptively inferred. A receiver is used as a measurement device to obtain real propagation data to check the performance of the model. Kullback–Leibler Divergence (KL‐Div) was used to evaluate different datasets. The results showed that the model's statistical properties were consistent. Its flexibility allows it to represent complex distributions without prior knowledge.