Towards Outdoor Electromagnetic Field Exposure Mapping Generation Using Conditional GANs.

Area of Science:

• Electromagnetics
• Signal Propagation
• Computational Electromagnetics

Background:

• The widespread deployment of fifth-generation cellular networks (5G) raises concerns about electromagnetic field (EMF) exposure.
• Accurate reconstruction of EMF exposure maps in urban environments is challenging due to sparse measurements and complex propagation characteristics.

Purpose of the Study:

• To develop and evaluate a novel method for reconstructing detailed EMF exposure maps in outdoor urban settings.
• To accurately estimate EMF propagation based on environmental topology using limited sensor data.

Main Methods:

• A conditional generative adversarial network (cGAN) was proposed for EMF exposure map reconstruction.
• The cGAN model was trained to learn EMF propagation characteristics influenced by environmental topology.
• Performance was compared against the simple kriging interpolation method.

Main Results:

• The cGAN-based approach demonstrated accurate EMF exposure estimations.
• The proposed method effectively utilized environmental topology for improved mapping.
• Results indicate superior performance compared to simple kriging.

Conclusions:

• The conditional generative adversarial network presents a promising solution for accurate EMF exposure map reconstruction.
• This method can address the challenge of sparse measurements in complex urban environments.
• The findings support the use of advanced AI for environmental EMF monitoring.