This project implements an innovative integration of Large Language Models (LLMs) into Open Radio Access Network (O-RAN) architectures. It leverages a specialized LLM engine and mediation layer to provide seamless functionality across various O-RAN components, including the RAN Intelligent Controller (RIC) via agent xApp and E2 interfaces.
- LLM-powered dynamic spectrum allocation
- Real-time network data analysis
- Adaptive resource management
- Integration with O-RAN components (RIC, xApp)
- Continuous learning mechanism
The system consists of several key components:
- LLM Engine: Utilizes Hugging Face models for network data analysis
- Mediation Layer: Orchestrates interactions between components
- Data Management: Handles data preprocessing and InfluxDB interactions
- API Management: Manages communications with xApp and RIC
- Training Server: Implements continuous learning mechanisms
- Additional Services: Content serving, caching optimization, safety management, and system monitoring
- Python 3.8+
- InfluxDB
- Access to OAI 5G stack - O-RAN components (RIC, xApp)
- Clone the repository: