| 摘要: | 在5G網路環境下，擴增實境（XR）、自動駕駛與即時遠端控制等新興應用迅速發展，這些服務皆具備對極低延遲與高可靠性的高度需求。傳統的切換機制（Handover, HO）主要由基地台（gNB）集中管理，用戶端設備（User Equipment, UE）需持續進行訊號測量並回報，作為基地台決策依據，導致顯著的訊號負載與傳輸延遲，難以滿足上述應用對即時性的要求。為了解決此問題，本論文提出一種以用戶端為中心之主動觸發切換機制（UE-centric Handover, UHO），引入地理圍欄（Geofence）作為觸發依據
。地理圍欄的設計係根據歷史切換成功與失敗位置資料，結合距離門檻過濾掉鄰近失敗區域的風險點，再以凸包（Convex Hull）方式產生邊界區域，並針對每段邊界分配對應的切換門檻與目標基地台。UE 僅在接近圍欄邊界時才啟動測量程序，並依據區域規則決定是否執行切換，大幅降低無謂測量與上行訊號回報。藉由此機制，UE 得以在接近風險區域時提前評估與執行切換，降低因延遲或錯失時機而導致的服務中斷風險，並提供更佳的即時通訊體驗。透過 NS-3 模擬平台進行效能驗證，結果顯示相較於傳統切換機制與條件式切換（Conditional Handover, CHO），所提 UHO 方法能有效降低測量頻率與訊號負載，並縮短切換延遲，提升低延遲應用於5G環境的穩定性與擴展性。
;Emerging low-latency services in 5G networks, such as Extended Reality (XR), autonomous driving, and real-time remote control, demand ultra-low latency and high reliability. Traditional handover (HO) mechanisms are predominantly managed by the serving gNB and require continuous signal measurement and reporting by the user equipment (UE), resulting in considerable signaling overhead and transmission latency that hinder their suitability for latency-sensitive applications. To address this challenge, this thesis proposes a UE-centric handover (UHO) mechanism based on geofence awareness. The proposed approach allows the gNB to construct geofences using historical handover success and radio link failure (RLF) data, applying a guard distance to filter out regions near failure-prone areas. A convex hull algorithm is then used to define the geofence boundary, and each edge is annotated with localized hysteresis thresholds and target cell IDs. This configuration is delivered to the UE via the RRC connection reconfiguration message using a customized IE structure. The UE monitors its own location and triggers measurements only when approaching the geofence boundary. Based on local signal evaluation and pre-configured rules, the UE autonomously determines whether a handover is required and notifies the gNB accordingly. This design effectively reduces the frequency of measurement gaps and uplink signaling while enabling timely and context-aware handover decisions. Performance evaluations conducted with the NS-3 simulation platform demonstrate that, compared to legacy HO and conditional handover (CHO) mechanisms, the proposed UHO scheme significantly reduces signaling load, measurement frequency, and handover latency. As a result, it improves handover reliability, scalability, and overall service continuity for diverse low-latency applications in 5G networks. |
