用於未來行動網路的協作技術之研究：從子載波、用戶到基地台之合作

Abstract

大部分的無線通訊研究，可以說都是為了克服有限頻譜資源與不斷增加的 用戶需求之間的衝突。過去二十年來對於多天線與正交子載波的探索，已經使頻譜效率有了長足的進展。然而，在某些情況下，如何彈性且動態的利用有限的資源，以適當的滿足用戶不斷變化的需求，其重要性甚至還大過於更高的頻譜效率。對於未來的移動網絡，已有一些可預見的趨勢與其伴隨之需求，包括：一、使用大量的天線，以及對分集增益與多工增益之間的動態取捨之需求; 二、網路中增加更多的用戶終端，以及對它們的高彈性排程之需求; 還有三、高密度的小蜂巢部署，以及對干擾管理之需求，尤其是頻率完全重用下產生的嚴重干擾。頻譜資源的限制，引發了對其他諸多領域的探索，如子載波、用戶和基地台（即蜂巢）等，而在這些維度的協調技術，即為本論文的研究重點。本論文內容主要分成三個方向，分別簡述如下： 首先，在第一部分，我們提出了一種新的混合循環延遲分集技術，用於多 輸入多重輸出正交分頻多工系統。與現有的雙空時發射分集與堆疊循環延遲分集等適合於整數多工率的技術相比，我們所提出的技術，可以實現任意非整數多工率。從我們的模擬結果，證明該技術在選擇所需的發送天線數量以及在調整分集和多工增益上，具有更高的彈性，可滿足用戶的各種需求。此技術在彈性多功率設定與低複雜度接收機之優勢，亦可透過應用在可擴展性視訊廣播中呈現。 在第二部分，我們首先指出儘管多天線系統可以提供空間多工增益並提高 頻譜效率，但卻會受限於手持式移動設備的天線數量遠不及於基地台的天線數量。為了解決這個問題，虛擬多天線技術可協調一群單天線移動用戶，成為一個虛擬多天線實體。但是，現有的虛擬多天線技術若非僅適用於小群用戶，就是計算複雜度太高。在本文中，我們提出了一個低複雜度的多用戶協調虛擬多天線技術，且具有彈性的群體大小。與現有的方法相較，我們提出的方法在6 × 6 虛擬多天線的情境下，可以減少多達 96% 的浮點運算，同時維持相同的吞吐量和公平性等性能。 在第三部分，我們首先提出兩個多蜂巢協作應用例：適用於資料導向網路的自適應換手啟動技術，以及宏分集天線合併技術。我們所提出的自適應換手啟動技術，適用於以資料為導向的網路，可提昇換手操作的資料率。此外，我們提出多種宏分集天線合併技術，並藉由模擬結果證實時頻編碼宏分集天線合併可帶來最高的頻譜效率。雖然多蜂巢協調技術在處理通信系統中的蜂巢間干擾發揮重要的作用，然而蜂巢間的信息交換效率，是主要的限制因素。藉由結合雲計算技術，集中式射頻接取網路可以用更低的資本支出與營運成本，達到更高的協作處理增益，被視為潛在的第五代射頻接取網路架構。本文對關鍵的協作技術，如換手、用戶與蜂巢關聯、排程和預編碼等，進行了全面的調查。其中多維最佳化雖然對通訊效能有所助益，但通常伴隨不切實際的高複雜度，這也使得低複雜度、僅有些微性能損失的啟發式技術，成為集中式射頻接取網路能否成功的關鍵。而為因應將來異質性前端網路、高密度布建和巨量天線帶來的挑戰，本文提出新的技術開發方向，如前端網路流量削減、彈性功能配置、以及射頻接取網路共享等。 總結來說，在本論文中，我們研究了子載波、用戶與蜂巢等維度之協作技 術，為現有的行動網路帶來如下的改進： • 使多天線分集與多工增益取捨以及多用戶容量與公平性取捨具有更高的彈性； • 對於天線、用戶終端與蜂巢數目具有更高的擴展性； • 更低複雜度的演算法，且幾乎不會使性能下降。

The contradiction between the limited spectrum resource and ever-increasing user requirement lays the foundation of most researches in wireless communications. The exploration in the use of multiple antennas and orthogonal subcarriers for the last two decades have extended the spectrum efficiency to a great extent. Besides the pursuit for higher spectrum efficiency, it is sometimes more critical to leverage the limited resource flexibly and dynamically to properly match the constantly changing user requirements. For future mobile networks, some trends and their accompanied requirements can be foreseen, including: a) the use of massive antennas and the requirement for dynamic tradeoff between diversity and multiplexing gain; b) the emergence of massive user terminals, and the requirement for flexible scheduling of them; and c) the deployment of dense small cells, and the requirement for managing interference among them due to full frequency reuse. The limitation in spectrum resource gives rise to the exploitation in many other domains, such as subcarriers, users, and base stations (i.e. cells), and the coordination technologies in these domains are the focus of this dissertation.

The dissertation includes three parts: • Multi-carrier coordination for diversity-multiplexing-tradeoff (DMT); • Multi-user coordination for virtual MIMO uplink; and • Multi-cell coordination for RAN enhancement.

In the first part, we propose a new hybrid cyclic delay diversity (HCDD) scheme for multiple input multiple output (MIMO) orthogonal frequency division multiplexing (OFDM) systems. Unlike the existing double space time transmit diversity (DSTTD) and stacked cyclic delay diversity (SCDD) schemes suitable for integer multiplexing rates, the proposed HCDD scheme can achieve arbitrary non-integer multiplexing rates. Our results demonstrate that the HCDD scheme can provide flexibility in selecting the required number of transmitting antennas and in adjusting diversity and multiplexing gains to match various users requirements. The advantages of the proposed HCDD, including the flexible rate assignment and less complex receivers, are presented in applications of scalable video broadcasting (SVB).

In the second part, we point out that although MIMO antenna systems can provide the spatial multiplexing gain to enhance spectrum efficiency, but they are hampered by the fact that the number of antennas in a hand-held mobile device is smaller than that of a base station. To overcome this issue, the virtual MIMO techniques coordinate a group of single-antenna mobile users to become a virtual multiple-antenna entity. However, existing virtual MIMO techniques are either applicable only to a small group or computationally too complex. In this part, we propose a low-complexity user scheduling for a virtual MIMO with a flexible group size. In comparison with existing approaches, our results show that the proposed user scheduling for virtual MIMO can reduce the floating operations by 96% in the case of a 6 × 6 virtual MIMO while maintaining the same throughput and fairness performance.

In the third part, we first provide two cases of multi-cell coordination: adaptive handover initiation for data-oriented networks and macrodiversity antenna combining techniques. The adaptive handover initiation technique is suitable for data-oriented networks and can improve data rates in handover operation. Besides, we propose macrodiversity antenna combining techniques, and use simulation results to show that space-frequency-block-coded macrodiversity antenna combing has the highest spectrum efficiency. Although multi-cell coordination technologies play an important role in addressing the inter-cell interference in communication systems with full frequency reuse, the efficiency in information exchange among cells, however, is the main limiting factor. By combining cloud computing technologies, centralized radio access network (C-RAN) can achieve higher coordination processing gain with lower CAPEX and OPEX than its predecessor and regarded as the potential 5G RAN architecture. The key coordination technologies, such as handover, user-cell association, scheduling and precoding, are comprehensively surveyed in this dissertation. Multidimensional optimization is desired but usually with impractical complexity, and low complexity heuristics with moderate performance loss become the key to the success of C-RAN. To cope with challenges brought by heterogeneous fronthaul, denser deployment, and scaled up antenna numbers, new technologies need to be developed such as fronthaul traffic reduction, flexible function allocation, and RAN sharing.

In summary, in this dissertation, we investigate coordination technologies in the subcarrier, user, and space (base station) domains to improve the performances of current mobile networks, including: • increased flexibility in multiple-antenna diversity-multiplexing-tradeoff and multiple-user capacity-fairness-tradeoff; • improved scalability in the number of antenna, users, as well as base stations; and • low-complexity algorithms with negligible performance degradation.