Survey on Path Loss Estimation of 5G mmWave using Deep Learning

This paper reviews various current path loss prediction models for mmWave-based 5G networks in dense urban areas, with a particular focus on deep learning techniques such as convolutional neural networks (CNN). This study explores how deep learning approaches, especially those utilizing satellite imagery, aim to improve the accuracy of path loss predictions. The results of literature analysis show that integrating distance information into CNN models can significantly improve prediction accuracy compared to models that only use image satelite. This paper highlights the strengths and weaknesses of various approaches, and proposes a simple CNN-based model that combines satellite imagery with distance information, which is able to provide more accurate and efficient path loss predictions in the context of 5G mmWave networks. This conclusion shows that the addition of distance information can improve prediction models that were previously only based on satellite imagery, improving prediction trendlines that are more in line with real conditions. These results provide valuable insights for further research and practical applications in 5G network planning and optimization.