Page 368 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

Use Case 16: AI-Driven Predictive Beamforming for Climate-

Resilient 5G/6G Networks

Country: Ethiopia

Organization Name: Addis Ababa Science and Technology University

Contact: Bereket Teklew Kibret, Bereketteklew@ gmail .com

1 Use Case Summary Table

Category 5G
Problem Addressed Millimeter-wave (mmWave) 5G/6G communication systems are suscepti-
ble to significant signal attenuation due to adverse weather conditions,
such as heavy rain and fog, resulting in reduced signal attenuation.
The use of predictive AI algorithms for weather data analysis, dynamic
Key Aspects of Solu- beamforming adjustments, and reinforcement learning-based real-time
tion adaptation enables proactive decision-making and improves signal
strength

Technology 5G, 6G, AI, Predictive Beamforming, Reinforcement Learning, Weather
Keywords Attenuation

Data Availability The Weather Dataset from Kaggle is used for weather type classification
tasks [1]

Time-series weather data (rain rate, humidity, wind speed, temperature);
Metadata (Type of
Data) numerical beamforming parameters (RSSI, SINR); geospatial data (lati-
tude, longitude, elevation)
Model Training and Recurrent Neural Networks (RNN) for weather forecasting, ensemble
Fine-Tuning models for anomaly detection, and Reinforcement Learning (RL) for adap-
tive beamforming. Training on fused multi-modal weather and signal
data, with fine-tuning via transfer learning and policy optimization.
Testbeds or Pilot The use case is part of a larger research project
Deployments
Code Repositories Not Available

332