現行 5G New Radio (NR) 以網路端集中控制為主，能提供穩定的接入、排程、移動性管理與可達性。然而，許多短時間尺度的決策，其關鍵資訊往往最先出現在使用者設備 (User Equipment，UE) 端，例如下行通道品質的變化、上行資料的急迫性，或近期較為穩定的移動範圍。當這些資訊必須先回報給基地台或核心網路才能觸發後續程序時，可能造成接入延遲、排程不精準、無線鏈路中斷，以及追蹤區域過於粗略所帶來的過度呼叫傳送。 本論文提出以使用者為中心的無線接入網路 (UE-Centric RAN) 設計原則：由最先掌握即時且相關資訊的實體率先做出初始反應，而授權、全域協調、公平性與備援 (fallback) 仍由網路端掌握。 依據此原則，本論文將六個主要 NR 程序家族修改為以使用者為中心之運作，涵蓋 UE 發起之多基站量測、以該量測結果的時間戳輔助初始接入、UE 輔助的換手與雙連結準備、Leader UE 輔助的上行排程、側鏈群組的移動性與復原、使用者為中心追蹤區域 (UE-Centric Tracking Area，UCTA) 的呼叫與註冊更新。 其中五項以程序層級 (procedure-level) 的 ns-3 模擬進行量化評估：UE 發起之多基站量測、初始接入、UE 輔助換手、Leader UE 輔助上行排程、UCTA 呼叫；換手與雙連結家族透過換手案例部分量化，雙連結準備與側鏈復原則保留為設計層級的延伸。在模擬假設下，這些程序層級研究顯示具體效益：例如分階段 UCTA 呼叫使呼叫細胞傳送次數減少約 79%，UE-Centric 初始接入設計使廣播傳輸（週期性同步訊號區塊(SSB)及其按需回應）減少約 90%。 本論文的核心並非移除網路控制，而是重新安排第一反應權限。透過此架構，系統可望降低因網路端收集延遲、推論延遲或失真所導致的效能損失，同時維持朝 NR 標準相容設計所需的授權邊界與回復至原始規範程序的能力。;Current 5G New Radio (NR) operation is dominated by network-side centralized control， which provides robust access， flexible scheduling， mobility management， and reachability. Many short-timescale decisions， however， depend on information that the user equipment (UE) observes before the gNB， the NR base station， or the core network can infer it: a weakening downlink， an urgent uplink burst， or a stable recent mobility footprint. When such information must first be reported to the network before a procedure can be triggered， the system can incur access delay， imprecise scheduling， radio link failure， and the paging overhead of an overly coarse tracking area. This thesis investigates a UE-Centric Radio Access Network (UE-Centric RAN) design principle: the first response to timely and locally relevant information is initiated by the entity that observes it first， while the network retains authorization， global coordination， fairness， and fallback control. Following this principle， this thesis proposes UE-Centric operation for six major NR procedures: 1) UE-initiated probing; 2) network-assisted initial access; 3) UE-assisted handover and dual connectivity preparation; 4) Leader UE (L-UE) assisted uplink scheduling; 5) mobility and sidelink recovery for UE group; and 6) UE-Centric Tracking Area (UCTA) paging and update. Five of these procedures are quantitatively evaluated with ns-3 simulation. The UE-assisted handover and dual-connectivity preparation procedure is quantified through the handover cases， while the mobility and sidelink recovery procedure remains design-level extensions. Under the modeled assumptions， these procedure-level studies show concrete advantages. For example， staged UCTA paging reduces paging-cell transmissions by about 79% in the modeled scenario， and the UE-Centric initial access reduces periodic synchronization signal block (SSB) broadcast transmissions by about 90%. Notably， the core of this work is not to remove network control but to rearrange first-response authority. Through this architecture， the system is expected to reduce the performance loss caused by delayed or distorted network-side inference， while preserving a compatibility-oriented authorization and fallback capability for NR.
