低軌衛星（LEO）星座已成為次世代通訊網路的重要基礎，具備全球覆蓋、低延遲與高頻寬等特性。然而，由於 LEO 系統本身具高度動態性，使得在進行衛星與地面閘道間的回傳鏈路（feeder link）切換時，維持穩定連線變得格外困難。此切換過程受限於硬體與頻譜資源，需依序對各波束進行切換。若切換順序設計不良，可能導致使用者設備（UE）接連遭遇服務中斷，特別是當 UE 進入尚未完成切換的波束時。
本研究提出一種波束層級的 feeder link 切換策略，整合使用者移動趨勢與服務品質（QoS）需求，以降低服務中斷的機率。該策略採用雙階段排序方法：第一階段根據衛星軌道預測與 UE 的波束路徑建立初始順序，第二階段則根據切換時段內每個波束所涵蓋的 UE 數量進行優化。在此基礎上，我們使用拓撲排序框架，確保對使用者移動性影響較大的波束被優先切換。我們透過 MATLAB 與 STK 模擬進行驗證，以 Iridium星座為案例。模擬結果顯示，所提出的方法相較於現有策略顯著減少了使用者中斷事件，並提升整體網路的服務連續性與效能。;Low Earth Orbit (LEO) satellite constellations have emerged as a key enabler of next-generation communication networks, offering global coverage, low latency, and high bandwidth. However, the dynamic nature of LEO systems introduces challenges in maintaining stable connectivity, particularly during feeder link switching between satellites and ground gateways. This switching process is constrained by hardware and spectrum limitations, requiring beams to transition sequentially. Poorly designed switching orders can result in multiple consecutive service interruptions, especially when user equipment (UE) enters beams that have not yet completed the switching process.
This paper proposes a beam-level feeder link switching strategy that integrates user mobility trends and service quality (QoS) requirements to minimize service disruptions. The strategy employs a two-stage sorting approach: an initial sequence determined by satellite orbital predictions and UE beam paths, followed by refinement using the number of UEs under each beam during the switching window. A topological sorting framework is applied to ensure that beams critical to user mobility are prioritized. We validate the proposed approach through MATLAB and STK simulations, using the Iridium constellation as a case study. Evaluation results demonstrate that our method significantly reduces user interruption events compared to baseline strategies, improving service continuity and overall network performance.
