This page organizes papers related to artificial intelligence published around the world. This page is summarized using Google Gemini and is operated on a non-profit basis. The copyright of the paper belongs to the author and the relevant institution. When sharing, simply cite the source.
Communications to Circulations: 3D Wind Field Retrieval and Real-Time Prediction Using 5G GNSS Signals and Deep Learning
Created by
Haebom
Author
Yuchen Ye, Chaoxia Yuan, Mingyu Li, Aoqi Zhou, Hong Liang, Chunqing Shang, Kezuan Wang, Yifeng Zheng, Cong Chen
Outline
We present G-WindCast, a novel deep learning framework that retrieves and predicts 3D atmospheric wind speed by leveraging signal strength variations in 5G GNSS signals. This framework utilizes feedforward neural networks (FNNs) and Transformer networks to capture the complex spatiotemporal relationships between GNSS-derived features and wind dynamics. Initial results demonstrate promising accuracy for both wind speed retrieval and short-term wind speed forecasting (up to 30 minutes), with skill scores comparable to high-resolution NWP output in certain scenarios.
Takeaways, Limitations
•
Presenting the possibility of retrieving and predicting 3D atmospheric wind speed using GNSS data.
•
It shows similar performance to high-resolution NWP in short-term wind speed forecasting.
•
Provides forecasts that better match observations than ERA5 reanalysis data.
•
Demonstrates cost-effectiveness and scalability by maintaining excellent performance even with a reduced number of GNSS stations.
•
Limited prediction time (up to 30 minutes)
•
Performance limitations in certain scenarios.
•
The bias of the NWP model may not be completely overcome.
