A deep learning based dynamic network management system in heterogeneous networks
The deep learning based dynamic network management system optimizes network resources in heterogeneous 5G networks by analyzing traffic patterns and user behaviors, ensuring uninterrupted connectivity and efficient resource allocation through real-time anomaly detection and dynamic resource management.
Patent Information
- Authority / Receiving Office
- WO · WO
- Patent Type
- Applications
- Current Assignee / Owner
- TURKCELL TEKNOLOJI ARASTIRMA & GELISTIRME AS
- Filing Date
- 2024-12-23
- Publication Date
- 2026-06-18
AI Technical Summary
Current heterogeneous 5G networks face challenges in integrating heterogeneous network nodes and connections, leading to incompatibilities that adversely affect network performance and quality of service, particularly in integrating high-altitude platforms (HAPs) with low-altitude platforms (LAPs, which are often overlooked in existing studies.
A deep learning based dynamic network management system that utilizes high-altitude platform stations to analyze traffic patterns and user behaviors, perform real-time anomaly detection, and optimize resource allocation using multi-agent deep reinforcement learning (MA-DRL) to ensure uninterrupted connectivity and efficient resource use.
The system provides uninterrupted connectivity and efficient resource allocation in unexpected situations like natural disasters or security threats by dynamically managing network resources and detecting anomalies, enhancing network performance and quality of service.
Smart Images
Figure 1 Figure 2