| 摘要: | 隨著第五代行動網路 (5G) 廣泛部署，基地台 (Base Station，BS) 已成為行動網路能耗的主要來源，尤其在物聯網 (IoT) 與智慧終端迅速增長的驅動下，網路能源消耗問題愈發嚴峻。為了因應5G-Advanced (5G-A) 與第六代行動網路 (6G) 對於綠色低碳的需求，行動網路節能 (Network Energy Saving，NES) 已成為重要研究議題。
本研究聚焦於行動網路基地台節能機制之設計與優化，基於3GPP Release 19所提出的網路節能功能 (NES Function，NESF) 框架，結合用戶設備 (User Equipment，UE) 輔助資訊 (UE-Assisted Information，UAI) 與流量預測技術，提出一套具備分散式與集中式運算架構之基站節能控制策略。本文針對基站運行與節能狀態轉換的決策過程，分別提出三種優化方案：基站導向節能 (BS-oriented NES，B-NES)、用戶導向節能 (UE-oriented NES，U-NES) 與混合式節能 (Hybrid NES，H-NES)。
於B-NES與U-NES策略，分別以基站流量負載與用戶設備回報之接收信號參考功率 (Reference Signal Received Power，RSRP) 為依據，輔以長短期記憶網路 (Long Short-Term Memory，LSTM) 進行流量預測，協調基站之間的節能狀態與服務範圍。而H-NES則進一步結合兩者資訊，並採用哈里斯鷹優化演算法 (Harris Hawks Optimization，HHO) 與鯨魚優化演算法 (Whale Optimization Algorithm，WOA)，透過多目標啟發式優化方法，在滿足用戶服務品質 (Quality of Service，QoS) 需求下，達成整體網路能效最佳化。
為降低核心網路 (Core Network，CN) 之運算負載，本文亦設計分散式流量預測與用戶資訊回報機制，使網路管理功能AMF得以快速辨識適合進入節能狀態或需恢復運行之基地台，並有效因應動態流量變化與潛在的換手問題，確保服務穩定性。本研究於MATLAB環境下建立行動網路節能模擬平台，進行多種場景之模擬與性能評估。結果顯示，所提出之H-NES策略在節能效果上表現最佳，惟計算複雜度亦最高；U-NES策略則在低計算資源需求下，仍能達成優於B-NES之節能成效，展現良好效能與實務應用潛力。
綜上所述，本論文所提出之用戶協助與流量預測整合之基站節能控制機制，能有效降低行動網路能耗，兼顧節能效率與服務品質，為5G-A與6G行動網路節能發展提供具體可行之解決方案，並具備未來於O-RAN開源平台實作驗證之可行性與擴展性。
;With the widespread deployment of fifth-generation (5G) mobile networks, base stations (BSs) have become the primary source of energy consumption in cellular systems. Driven by the rapid proliferation of the Internet of Things (IoT) and smart devices, the issue of network energy consumption has become increasingly critical. To meet the green and low-carbon requirements of 5G-Advanced (5G-A) and sixth-generation (6G) networks, network energy saving (NES) has emerged as a key research focus.
This study concentrates on the design and optimization of BS energy-saving schemes based on the Network Energy Saving Function (NESF) framework defined in 3GPP Release 19. By integrating User Equipment-Assisted Information (UAI) and traffic prediction technologies, a hybrid distributed and centralized architecture is proposed to dynamically manage the transition of BSs between active and energy-saving states. Three NES schemes are developed, namely the BS-oriented NES (B-NES), the UE-oriented NES (U-NES), and the Hybrid NES (H-NES), each leveraging different sources of network information to enhance energy efficiency.
In the B-NES and U-NES schemes, BS traffic load measurements and UE-reported Reference Signal Received Power (RSRP) values are utilized, respectively. These are combined with Long Short-Term Memory (LSTM) models for traffic prediction, enabling coordinated energy-saving decisions across BSs and adaptive adjustment of service coverage. The H-NES scheme further integrates both traffic and UE information and employs heuristic multi-objective optimization algorithms, including Harris Hawks Optimization (HHO) and Whale Optimization Algorithm (WOA), to achieve optimal network energy efficiency while maintaining user Quality of Service (QoS).
To alleviate the computational burden on the core network (CN), this study also designs a distributed traffic prediction and UE information reporting mechanism. This approach enables the AMF or NESF to promptly identify BSs suitable for transitioning into energy-saving or active states, facilitating dynamic responses to traffic fluctuations and potential handover issues, thereby ensuring service continuity and stability.
A simulation platform was developed in MATLAB to evaluate the proposed schemes under various network scenarios. Simulation results indicate that the H-NES scheme achieves the highest energy-saving efficiency but at the cost of increased computational complexity. In contrast, the U-NES scheme demonstrates superior energy-saving performance compared to B-NES while maintaining lower computational overhead, showcasing its practical applicability in real-world network environments.
In summary, this thesis proposes a comprehensive BS energy-saving control framework that integrates UE-assisted information and traffic prediction to effectively reduce network energy consumption while balancing energy efficiency and QoS. The proposed schemes align with 3GPP standards and offer feasible solutions for energy-efficient operations in 5G-A and 6G networks, with potential for future implementation and validation on the O-RAN platform. |
