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

Abstract

Mobile Phone Emissions in 5G FR1: Using Statistic Inferences and Deep Learning for Empiric Features Extraction Miclaus S, Deaconescu DB, Vatamanu D, Buda AM. Mobile Phone Emissions in 5G FR1: Using Statistic Inferences and Deep Learning for Empiric Features Extraction. 2024 IEEE International Symposium on Measurements & Networking (M&N), Rome, Italy, 2024, pp. 01-06, doi: 10.1109/MN60932.2024.10615263. Abstract 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. Excerpts It has been recently proved that the main exposure source of a person is the mobile phone itself [5], [6]. The study results of [7] confirmed that the human exposure in a 5G network is dominated by the uplink, and can be ten times larger than the downlink exposure.... Conducting measurements in real-life situations contributes to true assessment of the field exposure impact. It is essential to interpret all the findings on the background of dosimetric determinations and based on present regulatory guidelines [16], [17]. However, studies like [18] - [20] clearly pointed out that “cell phones emit a succession of pulse trains of different durations, the instantaneous fields of such signals allow activation of molecular, electronic, and protonic components within cells that have different thresholds and relaxation times, enriching the non-thermal effects of radiofrequency fields”. In the current work we aimed to analyze the exposure situation generated by a 5G mobile phone in a static position, while emitting signals in a 100 MHz bandwidth centered on 3.58 GHz. While using four different mobile applications running on the phone ... An Iphone 14pro (model A 2890, Apple, Zhengzhou, China) was the source of the emissions. It was maintained in a stationary position and connected to the 5G mobile network - Orange Romania operator. The central frequency was f=3.58 GHz, the bandwidth was 100 MHz and Time Division Multiple Access (TDMA) scheme was used.... ... the phone was situated at the cell edge and it received a good downlink signal. The receiving system was composed of a signal analyzer model FSW from Rhode & Schwarz, with a real time bandwidth of 160 MHz connected to a receiving antenna model Aaronia Omnilog 30800 (Fig. 1).... ... video streaming conducted to the highest mean emitted powers, while upload provokes the highest peak powers. Crest factors are the largest when a file is uploaded, and they can exceed 20 dB.... Variations of 19 dB around mean power is equivalent to a factor of approximately 8 x 10**3. Such a large variability during so short time (ms), may pose challenging biological response. The trends in the excessive exceedance of the mean power depend at least on mobile application type and on the modulation scheme. From this perspective, the tail features of time-variability of the exposure level may trigger unexpected consequences.... CONCLUSIONS A mobile phone connected to a 5G-NR network in FR1 band was investigated for its emitted EMF level and the associated time-variability characterization (during 250 ms), in four situations of usage: during file-upload, file-download, video call and Internet streaming. The measurement set-up consisted in a receiving antenna mostly planar and parallel to the phone’s display connected to a vector signal analyzer with a capability of analysing bandwidth larger than the communication bandwidth (100 MHz). The antenna collected the field of the phone at 30 cm distance from it, and had an omnidirectional pattern in azimuth.... All the results showed that the uplink EMF level is far below the safe limit based on thermal effects of human exposure. However, other important features evolved and they may be further linked to the non-thermal/specific effects of microwaves.... When it comes to the prevalence of peaks in the exposure, it resulted that crest factors varied mostly in function of the mobile application used but also, in a lower measure, in function on the modulation scheme. The largest crest factors were more than one order of magnitude higher than the smallest ones, overall. Mean and peak powers of emissions could deviate from each other with approximately two orders of magnitude, for the same mobile application. Such short and pronounced extremes should be carefully investigated from the biological endpoint of view.

AI evidence extraction

Main findings

Measurements of an iPhone 14 Pro connected to a 5G-NR FR1 network (100 MHz bandwidth centered at 3.58 GHz) showed notable differences in emission characteristics across four usage scenarios (file upload, file download, video call, internet streaming), including differences in amplitude probability densities, complementary cumulative density functions, and repeatability of time-frequency features. The authors state that the uplink EMF level was far below the safe limit based on thermal effects, while emphasizing pronounced short-term variability and peaks (e.g., crest factors exceeding 20 dB) that they suggest warrant further biological investigation.

Outcomes measured

• Measured emitted EMF level from a 5G-NR FR1 mobile phone
• Time variability of emissions (peak levels, statistical distributions)
• Time-frequency features/spectrogram characteristics
• Crest factor differences by application/modulation scheme
• Comparison to safety limits based on thermal effects

Limitations

• No health outcomes measured; study focuses on emission characterization and comparison to thermal safety limits
• Single phone model and single operator/network context described (iPhone 14 Pro on Orange Romania)
• Phone maintained in a stationary position; real-life use variability may differ
• Measurement geometry fixed (antenna at 30 cm, mostly planar/parallel to display)
• Short observation window for time-variability characterization (250 ms)

Suggested hubs

• who-icnirp (0.32)
  Mentions interpretation against present regulatory guidelines and thermal-effect safety limits.

{
    "study_type": "exposure_assessment",
    "exposure": {
        "band": "RF",
        "source": "mobile phone (5G-NR FR1)",
        "frequency_mhz": 3580,
        "sar_wkg": null,
        "duration": "250 ms (time-variability characterization)"
    },
    "population": null,
    "sample_size": null,
    "outcomes": [
        "Measured emitted EMF level from a 5G-NR FR1 mobile phone",
        "Time variability of emissions (peak levels, statistical distributions)",
        "Time-frequency features/spectrogram characteristics",
        "Crest factor differences by application/modulation scheme",
        "Comparison to safety limits based on thermal effects"
    ],
    "main_findings": "Measurements of an iPhone 14 Pro connected to a 5G-NR FR1 network (100 MHz bandwidth centered at 3.58 GHz) showed notable differences in emission characteristics across four usage scenarios (file upload, file download, video call, internet streaming), including differences in amplitude probability densities, complementary cumulative density functions, and repeatability of time-frequency features. The authors state that the uplink EMF level was far below the safe limit based on thermal effects, while emphasizing pronounced short-term variability and peaks (e.g., crest factors exceeding 20 dB) that they suggest warrant further biological investigation.",
    "effect_direction": "unclear",
    "limitations": [
        "No health outcomes measured; study focuses on emission characterization and comparison to thermal safety limits",
        "Single phone model and single operator/network context described (iPhone 14 Pro on Orange Romania)",
        "Phone maintained in a stationary position; real-life use variability may differ",
        "Measurement geometry fixed (antenna at 30 cm, mostly planar/parallel to display)",
        "Short observation window for time-variability characterization (250 ms)"
    ],
    "evidence_strength": "insufficient",
    "confidence": 0.7399999999999999911182158029987476766109466552734375,
    "peer_reviewed_likely": "unknown",
    "keywords": [
        "5G-NR",
        "FR1",
        "mobile phone emissions",
        "uplink exposure",
        "3.58 GHz",
        "100 MHz bandwidth",
        "TDMA",
        "time variability",
        "spectrogram",
        "crest factor",
        "amplitude probability density",
        "CCDF",
        "vector signal analyzer",
        "real-life measurements"
    ],
    "suggested_hubs": [
        {
            "slug": "who-icnirp",
            "weight": 0.320000000000000006661338147750939242541790008544921875,
            "reason": "Mentions interpretation against present regulatory guidelines and thermal-effect safety limits."
        }
    ]
}

Comments

Log in to comment.

No comments yet.