This paper reports results from a large-scale, multi-phase measurement campaign in mainland France assessing changes in RF exposure associated with 5G deployment from 2020 to 2024. Using more than 24,000 on-ground measurements in direct view of 5G antennas, it finds small average increases in broadband exposure and increased contributions from 5G-related bands over time. The 3.5 GHz band contribution increased but remained a secondary contributor compared with legacy 800/900 MHz bands, and exposure during active downloading was higher than in idle mode.

This exposure-assessment study proposes a Voronoi-diagram approach to visualize RF-EMF exposure across a city using personal exposimeter measurements of RMS electric field at seed points. Most mapped areas corresponded to about 1.9 V/m, with a maximum reported value of 11.4 V/m, all below the cited ICNIRP guideline level. The authors conclude the method is useful for communicating spatial variability, while also noting broader literature discussing potential health risks from EMF exposure.

This engineering/monitoring paper describes enhancements to the Serbian EMF RATEL system, which has continuously monitored RF-EMF since 2017, by adding drive test functionality to improve spatial coverage. The authors report preliminary quantitative drive test measurements and validation of the upgraded approach. The work emphasizes that characterizing spatial and temporal RF-EMF patterns can support exposure assessment relevant to public health risk evaluation.

This paper describes measurements and simulations to support terahertz (THz) sensing of European honey bees for environmental monitoring. It reports dielectric characterization of bee body parts across 1–500 GHz, scattering-based validation of 3D-printed bee mockups, and THz imaging demonstrating detailed anatomical visualization. The work includes dosimetric simulations at 300 GHz to evaluate feasibility of non-invasive, continuous monitoring and notes potential relevance to assessing high-frequency EMF impacts on insect health and habitat safety.

This exposure assessment measured environmental and auto-induced RF-EMF across more than 800 microenvironments in ten European countries, with a focus on 5G-related bands. Non-user environmental exposure was reported to be below international guideline values and similar to prior European research, while induced traffic substantially increased measured exposure, especially in uplink scenarios. The study also reports systematic differences by setting (cities vs villages) and by national precautionary limit policies.

This exposure-assessment study analyzed two years of continuous frequency-selective monitoring from 13 sensors in the five largest cities in Greece, focusing on the 3.6 GHz 5G band. It reports a gradual increase in 3.6 GHz environmental EMF levels over time and greater variability than legacy cellular bands, with 30-minute averaging reducing observed fluctuations. Despite the upward trend and variability, all measurements were reported to remain well below Greek and ICNIRP reference levels, and the authors emphasize the value of continuous monitoring as 5G deployment expands.

This paper proposes a time-averaging approach for analyzing long-term EMF exposure using time-series data from three sensors in a regulatory monitoring network. Sensors were installed at two kindergartens and one elementary school, and analyses reported daily/weekly patterns, differences between weekdays and weekends, and site-specific annual increases/decreases. The work emphasizes the value of continuous monitoring in sensitive areas, while not directly assessing health outcomes.

This engineering/methods study proposes an evolutionary computation approach (PCRO-SL) to optimize the selection of measurement points for constructing RF-EMF exposure maps. Tested on real measurements in Meco (Madrid, Spain), it reports good agreement with a reference exposure map while reducing required sampling density. The authors provide practical point-selection rules (e.g., line-of-sight within 250 m and directional sampling within 500 m) intended to maintain representativeness and avoid interpolation artifacts.