This study presents an extensive analysis of electromagnetic field (EMF) exposure in Greek urban and suburban areas during the ongoing deployment of the 5G technology. A total of 400 in situ measurements of electric field strength were conducted at ground level across four distinct regions. The first part of the study provides a statistical analysis of the overall EMF exposure, highlighting the contributions of different cellular generations (2G–5G) to the total field. The results show that 4G networks are the dominant contributors, while the impact of 5G remains limited. The second part investigates the correlation between electric field levels in the 3.5 GHz band and the distance from 5G base stations (BS). Using geospatial analysis, we computed distances to the nearest BS and examined their relationship with measured EMF levels. The findings indicate a general inverse correlation between EMF levels and BS distance, although variations occur due to factors such as line-of-sight conditions and urban built environment characteristics. These results confirm that distance remains a key predictive feature for AI-based EMF exposure models, even in the case of 5G networks.

This study explores the application of Artificial Intelligence (AI) in predicting electromagnetic field (EMF) exposure levels in urban environments. Machine Learning (ML) and Deep Learning (DL) models were developed to estimate the electric field (E, in V/m) at specific locations without requiring direct measurements. Measurements were conducted in Thessaloniki, Greece, covering a diverse urban landscape, including commercial and residential areas. The methodology involved collecting EMF data using high-precision equipment, validating and refining publicly available datasets, and extracting key features using Geographic Information System (GIS) tools. Essential input features included distances to the nearest base stations, their installation heights, classification of base stations regarding their emitted power, number of obstructing buildings, built area density, and line-of-sight (LOS) conditions. The AI models demonstrated strong predictive performance, with mean absolute errors (MAE) slightly above 0.3 V/m. The study highlights the importance of proper data preprocessing, feature selection, and integration of real-world measurements into AI-based prediction models. The results suggest that AI can serve as a reliable alternative for EMF exposure assessment, potentially reducing the need for costly and time-consuming field measurements while ensuring compliance with safety regulations.

This study introduces a low-cost alternative method for mapping the electric field strength from 4G LTE base stations and identifies areas where this mapping is more accurate. A drive test campaign was conducted in the urban environment of Thessaloniki, Greece, using data obtained from three identical smartphones, each connected to a different mobile operator and an exposimeter. The smartphones used a mobile application to record Reference Signal Received Power (RSRP) values, while the exposimeter measured the electric field strength in selected frequency bands. In the first part, the variability of the received power over different periods within certain areas was studied, and the reasons for this variability were identified. In the second part, a linear factor was calculated to convert RSRP values into electric field strength using data from both the application and the exposimeter. The converted RSRP values were subsequently compared with the exposimeter data for validation. The results indicate that in areas where the variability of the received power is lower, the linear relationship between smartphone and exposimeter data is statistically stronger resulting in calculated electric field strength values are closer to the measured.

This study is based on a drive test measurement campaign in the urban environment of Thessaloniki, Greece, utilizing three identical smartphones connected to different providers and a portable exposimeter to collect data on electromagnetic radiation from mobile networks. The investigation specifically addresses the impact of neighbor cells in the process of assessing exposure to electromagnetic fields (EMF). We study the correlation between RSRP values obtained from a mobile app and power density flux from the exposimeter. The results demonstrate that considering only the main and the first three neighbor - alternative cells from each provider can allow for an improved estimation of the spatial variation of exposure to EMF. Notably, the study underscores the practicality of using smartphones with low-cost applications as an effective alternative to expensive instruments for such assessments.

Radiofrequency (RF) exposure assessment in indoor environments has attracted considerable interest due to the widespread use of wireless communication technologies and the increasing amount of time people spend indoors. In the framework of the SEAWave project, extensive frequency-selective measurements were performed in Greece to characterize environmental RF exposure across diverse indoor environments. A harmonized dataset comprising more than 26,000 measurement samples collected at 40 indoor locations was statistically analyzed, covering residential, commercial, educational, office/workplace, public/cultural, event/market, and transport environments. Transport environments exhibited the highest exposure levels (RMS = 0.856 V/m), whereas residential (0.122 V/m) and office/workplace (0.136 V/m) environments showed the lowest values. Cellular networks accounted for more than 80% of the total RF exposure in all environments, reaching 94.4% in transport environments, while Wi-Fi contributed between approximately 5.6% and 17% depending on the environment. Repeated measurements revealed environment-dependent differences in RF exposure, with the largest variations between measurement campaigns observed in public transport, commercial, and public environments, whereas no statistically significant relationship between floor level and RF exposure was observed. RF exposure varied across indoor environment categories, while an exploratory geographical comparison within Commercial/Retail environments did not reveal statistically significant differences among urban, suburban, and rural groups.