Scholarlink

<p>The rapid development of sixth generation (6G) cellular networks demands intelligent, adaptive and autonomous mechanisms for efficient mobility management. Traditional cell selection approaches based on Reference Signal Received Power (RSRP) and Signal-to-Interference-plus-Noise Ratio (SINR) thresholds are insufficient for the ultra-dense, heterogeneous and high-frequency deployments envisioned for 6G. This paper proposes a Dueling Double Deep Q-Network with Prioritized Experience Replay (D3QN-PER) algorithm for 6G heterogeneous network environments. A custom network simulator incorporating terahertz (THz) channel models, ultra-massive MIMO, Reconfigurable Intelligent Surface (RIS)-assisted links, and dense small-cell deployments is developed to evaluate the proposed algorithm under realistic 6G mobility scenarios. Simulation results show that D3QN-PER achieves up to 66.8% reduction in handover failure rate, 34.6% improvement in average throughput, and 21.0% reduction in energy consumption compared to conventional RSRP-based, Q-learning, and standard DQN baselines. The results are validated across 10 independent simulation runs (mean ± std). These results position D3QN-PER as a strong candidate for deployment as a near-RT RIC xApp within the O-RAN architecture, advancing the vision of AI-native mobility management for 6G.</p>

错误：搜索内容不能为空，请输入英文关键词
错误：关键词超出字数限制，请精简
高级检索

A novel deep reinforcement learning-based cell selection algorithm for efficient mobility management in 6G networks

  • 期刊论文

  • 发表日期：

    2026年7月28日

查看全文
  • Kalpesh Popat,
  • Divyakant Meva

摘要

The rapid development of sixth generation (6G) cellular networks demands intelligent, adaptive and autonomous mechanisms for efficient mobility management. Traditional cell selection approaches based on Reference Signal Received Power (RSRP) and Signal-to-Interference-plus-Noise Ratio (SINR) thresholds are insufficient for the ultra-dense, heterogeneous and high-frequency deployments envisioned for 6G. This paper proposes a Dueling Double Deep Q-Network with Prioritized Experience Replay (D3QN-PER) algorithm for 6G heterogeneous network environments. A custom network simulator incorporating terahertz (THz) channel models, ultra-massive MIMO, Reconfigurable Intelligent Surface (RIS)-assisted links, and dense small-cell deployments is developed to evaluate the proposed algorithm under realistic 6G mobility scenarios. Simulation results show that D3QN-PER achieves up to 66.8% reduction in handover failure rate, 34.6% improvement in average throughput, and 21.0% reduction in energy consumption compared to conventional RSRP-based, Q-learning, and standard DQN baselines. The results are validated across 10 independent simulation runs (mean ± std). These results position D3QN-PER as a strong candidate for deployment as a near-RT RIC xApp within the O-RAN architecture, advancing the vision of AI-native mobility management for 6G.