cmWave through vegetation: Correlation of pixels and attenuation using UT and Bayes inference

This paper presents a method to model the radio propagation attenuation through vegetation, based on its correlation with image pixels using unscented transform (UT) and Bayes inference theorem. This model does not use the vegetation depth distance, but it is complementary to the physical modelling based on vegetation depth. The vegetation density based on pixels filtered from a satellite image of the measurement area is used as a statistic proxy to infer the attenuation. In a a 24GHz measurement scenario with densely-vegetated areas, the correlation between the attenuation and pixel values was 0.62, and the linear regression showed that 38% of the attenuation variance was due to the pixel values. The UT vegetation pixel and correlated attenuation distributions were used as prior and sampling distributions in the Bayes theorem to calculate the UT posterior vegetation attenuation distribution. This method can be useful for radio prediction and simulation tools.