Satellite propagation channel modelling using photogrammetry and hidden Markov model approach (English)

In a typical satellite-earth propagation path, the field strength of the received signal level changes with respect to time and space according to the changing morphology of the environment. It has been shown that the switching between different attenuation levels can be represented as a Markov stochastic process using the assumption that the channel properties are quasi-stationary in small periods of time. A hidden Markov model is introduced to replace the often used channel fading models, such as Rice, Loo and Rayleigh distributions. Three sets of parameters are trained to simulate the state transition and observation conditional probability density function among three basic propagation states that are categorised according to the environmental images. The simulated data were compared to the real RF measurement of satellite signals in several propagation properties, such as the cumulative fading distribution, level cross rate and error probability performance, at typical user environments. With the help of observation sets and their conditional probability densities, the proposed propagation channel simulator can model the satellite propagation properties with better accuracy than three-state Markov transition models.