異質網路中結合模糊理論及效益函數之Wi-Fi分流選取機制

Abstract

隨著智慧型手機日益普及，愈來愈多人使用手機上網，大量耗損現今3G基地台的數據頻寬，而未來4G行動通訊也將會面臨相同的問題。基地台業者開始推行比較平價低廉的小型基地台，好降低電信商擴建的財務負擔，而業界也提出Wi-Fi分流 (Wi-Fi offloading) 方案，因Wi-Fi能以廉價佈建來滿足外出的定點數據服務需求，在現階段為一項解決大型基地台負荷不錯的考量。但Wi-Fi資源分配為競爭的方式，會容易讓分流到Wi-Fi的使用者得不到在4G系統的服務品質(4G系統會依照使用者通道狀況去分配適當的資源)，因此如何使用Wi-Fi做適當的分流，不只達到減輕系統負荷的效果，還同時能保證使用者一定的服務品質，成為分流延伸出來的另一項重要議題。 而在本篇論文中，在考慮實務面的可行性下，我們提出了在同時有4G基地台及電信業者佈建之Wi-Fi覆蓋範圍(trusted-AP)的環境下，能選擇出通訊狀況較好的AP並做適當分流的機制。此機制使用一中心機台 (offload-server) 以模糊理論 (fuzzy logic theory) 考慮多種系統參數選擇出適當AP，將資料回傳至使用者終端 (user equipment)的應用程式，再由效益函數 (utility function) 決定不同資料流要以4G基地台或者Wi-Fi AP作為資料處理的服務者，用這樣的方式 (FUOD) 做分流，不僅能減少業者基地台負荷，更同時能讓使用者體驗到好的服務品質。

Mobile Internet access is getting ever-increasingly popular and today provides various services and applications. To cope with this explosive growth in traffic demands with the limited current network capacity, Wi-Fi offloading seems to be the most viable solution at the moment. However, the contention-based characteristic of Wi-Fi AP will easily cause bad transmission experience to users. Therefore, how to soothe the loading of 4G’s base station (eNB) and satisfy users transmission requirements is another issue. In this thesis, we propose a fuzzy logic-based AP selection and utility offloading decision scheme (FUOD). We use trusted Wi-Fi to be the offload candidates for its high compatibility for today’s wireless network environment and it can also increase revenue and capacity for operators. The offload-server can use fuzzy logic to mimic the system’s behavior with some system parameters. With this, the offload-server can choose a proper AP for UE to offload. Furthermore, the utility function is used in the final step to help the user equipment (UE) choose a service node (Wi-Fi AP or 4G eNB) for every traffic flow. With FUOD scheme, the loading of eNB can be reduced and the user can have nice quality of service as well.