On the Application of the Baum-Welch Algorithm for Modeling the Land Mobile Satellite Channel

TL;DR

This paper adapts the Baum-Welch algorithm to estimate parameters of Markov chain models for land mobile satellite channels, improving model accuracy using measurement data.

Contribution

It introduces a method to identify Markov model states and parameters from observed data using an adapted Baum-Welch algorithm for satellite channel modeling.

Findings

Algorithm successfully estimates model parameters from test data.

Numerical results demonstrate the effectiveness of the approach.

Real measurement data validate the method's practical applicability.

Abstract

Accurate channel models are of high importance for the design of upcoming mobile satellite systems. Nowadays most of the models for the LMSC are based on Markov chains and rely on measurement data, rather than on pure theoretical considerations. A key problem lies in the determination of the model parameters out of the observed data. In this work we face the issue of state identification of the underlying Markov model whose model parameters are a priori unknown. This can be seen as a HMM problem. For finding the ML estimates of such model parameters the BW algorithm is adapted to the context of channel modeling. Numerical results on test data sequences reveal the capabilities of the proposed algorithm. Results on real measurement data are finally presented.