The proliferation of wireless devices that utilize the 5G mmWave technology continues to stir several design requirements for optimizing performance. However, mitigating the underlying constraints posed by path loss remains a challenging task. The use of empirical methods and heuristic techniques have only provided one-shot solution. We adopt a reusable learning-based method to alleviate the complexity in path loss prediction. Our method involves the design of a deep classifier and a regression model which learn relevant network parameters from data and uses such knowledge to differentiate different classes of path loss in a practical 5G radio environment. We show that deep learning can alleviate the path loss modelling complexity in the optimization of 5G networks.
Paper
Full text
Deep Learning-Based Path Loss Prediction Model for 5G mmWave
Semantic Scholar · Computer Science · 2022
Abstract
The proliferation of wireless devices that utilize the 5G mmWave technology continues to stir several design requirements for optimizing performance. However, mitigating the underlying constraints posed by path loss remains a challenging task. The use of empirical methods and heuristic techniques have only provided one-shot solution. We adopt a reusable learning-based method to alleviate the complexity in path loss prediction. Our method involves the design of a deep classifier and a regression model which learn relevant network parameters from data and uses such knowledge to differentiate different classes of path loss in a practical 5G radio environment. We show that deep learning can alleviate the path loss modelling complexity in the optimization of 5G networks.