在次世代幽靈基地台架構下降低換手失敗率

Abstract

近幾年由於行動裝置的數量日增月益，以及使用者對於網路的資料需求量和服務品質與日俱增，未來若只藉由大型基地台的服務，難以滿足使用者多樣性的需求和全球數據流量大幅成長的相關問題。因此在長期演進技術(Long Term Evolution)發展趨勢中，小型基地台是目前公認極具潛力的發展方向之一，利用大型基地台和小型基地台的混合佈建的網路架構方式，來提升整體網路的容量，稱為異質性網路。而在異質性網路中，隨著小型基地台佈建密度逐漸上升，仍有重要的議題需去克服，像是換手(Handover)程序頻繁的發生、網路系統能源龐大的消耗和鄰近基地台間的干擾問題，這些問題有很大的原因是現行LTE中的無線接網路(Radio access network)中，控制存取點和資料存取點的緊密偶合，讓無限資源的彈性度多了相當程度上之限制。因此，一個新穎的幽靈基地台架構(Phantom cell architecture)可以克服異質性網路所帶來的問題，概念上是把無線接取網路中的控制平面(Control Plane)和資料平面(Data Plane)，分離成由大、小基地台去傳輸，讓大型基地台負責大範圍的使用者控制連線，而小型基地台專門傳輸資料，這讓分離式異質性架構在使用者移動管理、能源效率和干擾管理上有了更佳的優勢。 然而，因為幽靈式有著獨特的架構，在跨兩個不同的大型基地台情況下有著較高的換手失敗率，使用者不希望在換手過程中有失敗的情況發生，造成進行中的服務被中斷，導致較差的體驗品質。因此本文提出一套獨立換手觸發程序並且設計換手流程，進而降低換手失敗率過高的現象，並且在資源排程上提出貪婪式啟發演算法，使基地台可以服務越多的使用者。最後，我們的模擬結果驗證所提出的換手觸發程序與資源排程有著顯著的低換手失敗率。

As the global mobile data traffic increases, a paradigm shift from conventional macro base stations with high-power to heterogeneous network like small cell with low-power. However, the heterogeneous network would cause several problems, such as high interference, excessive unnecessary handover and energy consumption. The major reason of the above problems is the tight coupling between network access point (control plane) and data access (data plane) points. Hence, the new architecture, called phantom cell is a novel concept by means of the separation of the control plane and data plane, could probably overcome these issues. Then, the macro cell is responsible for the control plane that provides ubiquitous coverage and the small cell is responsible for the data plane that provides high data rate service. However, due to the special architecture of phantom cell, it would suffer from high handover failures between inter-macro cells. Thus, this phenomenon would let mobile user lose connection and degrade the quality of experience. In this thesis, we propose a independent handover trigger decision and design the handover procedure to reduce the handover problem. Moreover, we also consider the admission control of resource scheduling and propose a greedy heuristic algorithm that let base station can serve more users. Finally, our proposed scheme and resource scheduling method can significantly reduce the inter-macro cell handover failure rate from simulation results.