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

Abstract

Overview

This study quantifies the emissions from a 5G mobile phone during various typical application uses, emphasizing their role as the primary source of human electromagnetic field (EMF) exposure. The research investigates time variability, peak exposure levels, statistical distributions, and time-frequency spectrogram features.

Findings

• Significant differences were found in amplitude probability densities, complementary cumulative density functions, and repeatability of emissions depending on mobile application and modulation scheme.
• Real-life measurement shows that the phone’s uplink exposure can be up to ten times larger than downlink exposure.
• Exposure during mobile use exhibits highly variable instantaneous power levels—variations as high as 19 dB (approx. 8000x), and crest factors can exceed 20 dB.
• Different mobile applications (upload, download, video, streaming) result in markedly different mean and peak emission powers, sometimes differing by two orders of magnitude.
• Studies cited reveal that pulsed signals from mobile phones can activate molecular, electronic, and protonic cellular components at different thresholds, potentially enhancing non-thermal radiofrequency effects.

Conclusion

While uplink EMF levels remain below current regulatory thermal safety limits, the observed extreme time-variability and short, intense exposure peaks demand careful scrutiny due to potential links to non-thermal biological effects. The pronounced short-term peaks and large crest factors may present unforeseen health risks, reinforcing concerns over the biological impacts of EMF exposure variations in 5G mobile communications.