基於干擾對齊下叢集基地台與排程使用者之研究

Abstract

第五代行動無線通訊網路是未來的趨勢，使用者的人數增長非常的快速，干擾的影響是現代通訊中不可或缺的考量因素。為了降低干擾對通訊系統的影響，許多人提出干擾抑制(Interference Suppression)的方案，主要可以在實體層(Physical layer)或媒體存取控制層(Media Access Control layer)以訊號處理或用資源分配的方式解決。 在實體層中，干擾對齊(Interference alignment)是一個有前途的干擾抑制方案。選擇部分發射端來合作，同時設計傳輸的編碼器和解碼器，在接收端接收訊號時能將來自其他發射端的干擾對齊至同一個子空間，而訊號就能在一個無干擾的子空間中傳輸，廣義來說，這也是一種波束成型(Beamforming)的方法。 這個技術在理論中已被證實在消除干擾的效能十分顯著，雖然干擾對齊的效能很好，但也存在著兩個大問題，一是每個傳送端都需要知道需要被消除的干擾源的通道狀況資訊，這對骨幹網路來說是個大的挑戰，另一個是合作人數上的限制，因為有限的天線數只能消除有限的干擾源，因此存在了另一個問題就是我們如何決定該消除系統中的哪些干擾。 目前適合的解決方法是先將系統中的所有發射端藉由叢集演算法分類，包括基因演算法及粒子群演算法，每個叢集內部再分別使用干擾對齊，至於過多的使用者，我們使用排程來管理這樣的狀況。 而這篇論文的目的是發展一套干擾抑制的方案，藉由結合叢集演算法和和排序使用者，最後經由叢集內的功率分配來降低干擾以此增進效能，並利用理論分析和數據模擬來說明這樣的干擾管理效用和限制這樣的干擾管理方案能夠減少多用戶網路的干擾。

5G mobile wireless communication network is future trend. The number of users grow rapidly and the impact of interference cannot be ignored. In order to reduce the impact of interference to communication systems, many interference suppression schemes have been proposed. It is mainly solved by signal processing in the physical layer or by resources allocation in the media access control layer (MAC layer). In the physical layer, interference alignment (IA) is a promising solution of interference suppression scheme. A selection of transmitters will cooperate, and encoders and decoders will be designed accordingly. The signals from interfering transmitters are aligned to the same sub-space at the receiver, and the desired signal can be transmitted in the interference-free signal subspace. Generally speaking, it can be viewed as a beamforming technique. The technique has been proven effective in eliminate interference in theory. Although the performance gain of IA is significant, there are two serious issues. One is that every transmitter needs to know the channel status information (CSI) of each interference channels; it is a big challenge to backhaul networks. The other is limitation of the number of cooperative transmitters. The limited number of antennas can only align limited interferences, so another problem is how to decide which interference should be aligned. The current feasible solution is to divide all transmitters into many isolated and appropriate groups by a clustering algorithm, such as the genetic algorithm and the particle swarm optimization algorithm. Then apply interference alignment to each cluster. As for the excessive number of users, scheduling is deployed to manage the situation. The goal of this thesis is to develop an integrated interference suppression scheme by incorporating the clustering algorithm and scheduling algorithm for IA, and reduce power allocation is conducted to improve performance further more. Theoretical analysis and numerical simulations are provided to illustrate the effectiveness and limitation of this interference management scheme in reducing the multiuser interference in interference networks.