Mobile Phone Emissions in 5G FR1: Using Statistic Inferences and Deep Learning for Empiric Features Extraction

As the primary source of human exposure to electromagnetic fields, the emission of a 5G mobile phone is quantified for several mobile applications use. Following the time variability of the exposure, its peak levels and their statistic distributions, together with the time-frequency change in the spectrograms, we emphasize the peculiarities of the exposure in connection to modulation scheme used and to the application. The results show notable differences in amplitude probability densities, in complementary cumulative density functions and in repeatability of time-frequency features of the emissions.