Evaluation of Personal Radiation Exposure from Wireless Signals in Indoor and Outdoor Environments

Authors: Ruijie P, Sali A, Li L, Mohyedin MZ, Qahtan S

Year: 2025

Category: Exposure Assessment, Wireless Radiation, EMF Safety

Journal: IEEE Access

DOI: 10.1109/ACCESS.2025.3579085

URL: https://ieeexplore.ieee.org/stamp/stamp.jsp?tp=&arnumber=11031433&isnumber=6514899

Abstract

Overview

With the rapid growth of wireless technology, concerns have risen regarding the health effects associated with increased public exposure to electromagnetic fields (EMF). This study thoroughly evaluates personal EMF exposure from wireless signals within various indoor and outdoor micro-environments throughout Malaysia. Four different types of micro-environments, influenced by key factors such as population density and base station proximity, were selected for analysis.

Findings

  • A specialized radiation exposure meter (ExpoM-RF 4) was used to measure electric field strength across these environments.
  • Machine learning models—including Fully Connected Neural Network (FCNN), eXtreme Gradient Boosting (XG Boost), and Linear Regression (LR)—were employed to model and predict electric field strength (RMS and maximum exposure levels).
    • LR performed well with simple data sets.
    • XG Boost and FCNN excelled with varied or complex data sets.
    • FCNN provided the most accurate predictions, especially in urban and suburban environments where extreme values occurred.
  • Measured and predicted data were compared against public exposure limits set by ICNIRP, MCMC, and FCC.

Key Results

  • Personal radiation exposure usually remained lower than the exposure limit (61.4 V/m), consistent with previous studies.
  • In high-density regions with many base stations, maximum exposure could approach 56.7365 V/m—very close to the exposure threshold, highlighting a potential health risk from EMF in such areas.
  • The highest personal radiation exposure was found in outdoor urban environments, especially areas with:
    • High population density
    • Numerous, closely spaced base stations
  • Parks with dense vegetation showed markedly lower exposures—trees naturally attenuate EMF by absorbing and scattering the waves.
  • Indoor environments generally had reduced field strengths compared to outdoors, attributed to structural shielding and fewer high-power sources.

Conclusion

Linear Regression (LR) gives better predictions with simple data, while XG Boost and FCNN are superior for analyzing complex, multi-type data. FCNN is especially effective in urban situations with high/variable exposures.

Although most personal exposures are below the international safety limits, areas with dense populations and heavy wireless infrastructure can produce exposures approaching regulatory thresholds—signaling a tangible link between increased EMF exposure and elevated health risk potential. Vegetation and structural barriers help reduce risk, but caution and monitoring remain warranted in high-density urban areas.

electromagnetic fields wireless signals personal exposure health risk prediction modeling machine learning indoor and outdoor environments
← Back to Stats