5_B5G6G网络智能数据采析（英）_28页_3mb

B5G/6G Network Intelligence Data Acquisition and Analysis

Introduction

The B5G/6G research direction focuses on integrating AI, big data, and computing into networks to develop intelligent communication networks, enabling autonomous driving capabilities and personalized services. The core challenge lies in developing a novel network intelligence technology system featuring endogenous intelligence and compute/communication convergence. Data, as the foundation, determines the upper limit of network intelligence performance.

System Overview

A comprehensive data acquisition and analysis system consists of four functional modules:

1. Wireless Network Intelligent Open Platform: Provides real-time data acquisition, analysis, and visualization.
2. Data Acquisition: Collects diverse network data (wireless, core network, performance, alarm, configuration) through ETL processes.
3. Knowledge Graph-based Representation and Analysis: Transforms complex network data into structured knowledge using graph-based correlations.
4. Feature Data Sets Construction: Builds AI-ready datasets for network optimization, prediction, and decision-making.

Key Technologies

• Knowledge Graphs: Clarifies multi-source heterogeneous data correlations, enabling:

  • Entity extraction and relationship modeling.
  • Edge weight computation and node importance analysis.
  • Dynamic inference and intelligent feature engineering.

• Feature Data Sets:

  • Trains machine learning models for network performance assessment.
  • Supports applications like fault locating, radio network planning, and digital twin implementation.

Applications

1. Intelligent Network Optimization: Automatically identifies key parameters affecting performance.
2. Digital Twin Implementation: Creates virtual replicas of physical networks for real-time mapping and prediction.
3. Fault Locating: Accelerates root cause analysis in complex 6G networks with integrated space-air-ground coverage.

Conclusion

The proposed system establishes a standardized data acquisition flow, laying the foundation for advancing network intelligence technologies. The knowledge graph-based approach transforms "black-box" network operations into explainable "white-box" frameworks, supporting AI-driven intelligent decision-making.