Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Online Learning-based Adaptive Beam Switching for 6G Networks: Enhancing Efficiency and Resilience

Seyed Bagher Hashemi Natanzi, Zhicong Zhu, Bo Tang

TL;DR

The paper tackles the challenge of adaptive beam switching in dense 6G networks by introducing an online, stability-aware DRL framework. It enhances the state with blockage history and uses a reward that penalizes excessive switching and SNR fluctuations, aiming for robust, long-term link quality. In large-scale Sionna simulations with 100 UEs, the approach achieves throughput comparable to a reactive MAB baseline while significantly improving operational stability (about 43% better than a vanilla DRL). The results indicate that stability-oriented DRL can deliver efficient, reliable, real-time beam management suitable for mission-critical 6G scenarios, with feasible inference latency and scalable pathways explored for ultra-dense networks.

Abstract

Adaptive beam switching is essential for mission-critical military and commercial 6G networks but faces major challenges from high carrier frequencies, user mobility, and frequent blockages. While existing machine learning (ML) solutions often focus on maximizing instantaneous throughput, this can lead to unstable policies with high signaling overhead. This paper presents an online Deep Reinforcement Learning (DRL) framework designed to learn an operationally stable policy. By equipping the DRL agent with an enhanced state representation that includes blockage history, and a stability-centric reward function, we enable it to prioritize long-term link quality over transient gains. Validated in a challenging 100-user scenario using the Sionna library, our agent achieves throughput comparable to a reactive Multi-Armed Bandit (MAB) baseline. Specifically, our proposed framework improves link stability by approximately 43% compared to a vanilla DRL approach, achieving operational reliability competitive with MAB while maintaining high data rates. This work demonstrates that by reframing the optimization goal towards operational stability, DRL can deliver efficient, reliable, and real-time beam management solutions for next-generation mission-critical networks.

Online Learning-based Adaptive Beam Switching for 6G Networks: Enhancing Efficiency and Resilience

TL;DR

The paper tackles the challenge of adaptive beam switching in dense 6G networks by introducing an online, stability-aware DRL framework. It enhances the state with blockage history and uses a reward that penalizes excessive switching and SNR fluctuations, aiming for robust, long-term link quality. In large-scale Sionna simulations with 100 UEs, the approach achieves throughput comparable to a reactive MAB baseline while significantly improving operational stability (about 43% better than a vanilla DRL). The results indicate that stability-oriented DRL can deliver efficient, reliable, real-time beam management suitable for mission-critical 6G scenarios, with feasible inference latency and scalable pathways explored for ultra-dense networks.

Abstract

Adaptive beam switching is essential for mission-critical military and commercial 6G networks but faces major challenges from high carrier frequencies, user mobility, and frequent blockages. While existing machine learning (ML) solutions often focus on maximizing instantaneous throughput, this can lead to unstable policies with high signaling overhead. This paper presents an online Deep Reinforcement Learning (DRL) framework designed to learn an operationally stable policy. By equipping the DRL agent with an enhanced state representation that includes blockage history, and a stability-centric reward function, we enable it to prioritize long-term link quality over transient gains. Validated in a challenging 100-user scenario using the Sionna library, our agent achieves throughput comparable to a reactive Multi-Armed Bandit (MAB) baseline. Specifically, our proposed framework improves link stability by approximately 43% compared to a vanilla DRL approach, achieving operational reliability competitive with MAB while maintaining high data rates. This work demonstrates that by reframing the optimization goal towards operational stability, DRL can deliver efficient, reliable, and real-time beam management solutions for next-generation mission-critical networks.
Paper Structure (15 sections, 1 equation, 3 figures, 1 table, 1 algorithm)

This paper contains 15 sections, 1 equation, 3 figures, 1 table, 1 algorithm.

Table of Contents

  1. Introduction
  2. Related Work
  3. System Model and Problem Formulation
  4. System Setup
  5. Blockage Model
  6. Signal Model and Performance Metrics
  7. Problem Formulation
  8. Proposed Approach: Online Learning for Adaptive Beam Switching
  9. Simulation Setup and Results
  10. Simulation Setup
  11. Results and Analysis
  12. Complexity and Scalability Analysis
  13. Real-Time Feasibility
  14. Scalability
  15. Conclusion

Figures (3)

  • Figure 1: System Model Visualization. A single-cell 6G scenario where a Base Station serves mobile users moving along a trajectory, subject to dynamic blockage from the environment.
  • Figure 2: Holistic Performance Comparison. The radar chart illustrates that the Proposed Framework (Blue area) achieves the best overall balance among conflicting objectives. It matches the high SNR and Coverage of the Greedy baseline while achieving Stability and Service Continuity comparable to or better than the robust MAB baseline. Notably, the Vanilla DRL (Orange dotted) collapses in Stability and Continuity, confirming the necessity of the proposed stability-aware reward design.
  • Figure 3: Operational Stability Comparison. The proposed stability-aware DRL framework (Blue) achieves a Stability Score of $\approx 0.75$ in converged runs (lower is better), significantly outperforming the Vanilla DRL baseline (1.33) and surpassing the robustness of the reactive MAB benchmark (0.86). This confirms that the integration of the penalized reward function and temporal state features successfully suppresses erratic beam switching.