Towards Outdoor Electromagnetic Field Exposure Mapping Generation Using Conditional GANs

Abstract

Overview

With the increasing rollout of the fifth-generation cellular network (5G), understanding and managing electromagnetic field exposure has become considerably important. Accurately mapping these fields in geographic locations poses challenges due to limited data.

Methods

This study introduces a novel approach using conditional generative adversarial networks (GANs) aimed at reconstructing precise electromagnetic exposure maps. The focus is predominantly on outdoor urban settings where data is collected from a few sensors.

Key objectives:

• To accurately regenerate electromagnetic field exposure maps based on environmental topology.
• Training models to discern and predict electromagnetic field propagation.

Findings

The conditional GAN-based method demonstrates higher accuracy in generating exposure maps compared to traditional methods like simple kriging.

Conclusion

This technique shows significant promise as a robust solution for creating detailed and accurate maps of electromagnetic field exposure, which is crucial for public health planning and intervention in the era of increasing 5G implementation.