基于部分干扰消除的广义干扰对齐方法研究

发布时间：2018-06-05 00:14

本文选题：基于部分干扰消除的广义干扰对齐方法 + 可行性条件　；参考：《电子科技大学》2014年硕士论文

【摘要】：近年来,人们在研究传统干扰对齐技术时发现了一个重要的问题,即这种方法需要足够的维度空间来实行对齐和消除干扰,这一问题导致传统的干扰对齐技术难以被用于规模较大的系统。因而,人们将研究的重点转向了广义干扰对齐技术,这种技术的核心思想是利用信道衰落的差异,只对其中影响较强的干扰进行对齐,从而使系统需对齐的干扰数目能够控制在一定的范围内,因而干扰对齐能够实行。依据这一思想衍生出多种方法,其中较为流行的是基于L-interfering模型的干扰对齐方法,但这一方法在适用范围上也有很大局限。针对这一局限,本文提出了基于部分干扰消除的广义干扰对齐方法,即每个用户根据自身的配置情况和需求及各干扰的影响情况,确定系统可行性条件,并根据该条件对其进行选择,将其中被选出的一部分干扰进行对齐和消除。相比于已有的基于L-interfering模型的方法,这一方法的优势在于:首先,在适用条件上,不必要求每个用户的发送流数完全相同,因而每个用户可对齐的干扰数也不必相同;其次,在空间资源的利用上,由于这种方法在设计之时,充分考虑到系统配置和发送流数等方面的差异性,因而它可以根据用户所拥有的空间资源进行灵活地调整,使资源的利用更加充分,因而无论从系统总体还是单个用户来说,都可以达到更优的性能;同时,由于该方法在资源调度方面的灵活性,该方法在实施过程中还可以与其他优化目标相结合,从而满足某些特殊性能的优化需要。在该方法中,可行性条件是基础,依据该条件实行的干扰选择策略是核心步骤。根据这两点,本文主要做了以下几项工作:(1)分析了现有的L-interfering模型在适用性方面的局限性,提出了基于部分干扰消除的广义干扰对齐方法,并从理论方面分析其在扩展适用范围和节省空间资源方面的作用。(2)利用图论的知识,以单流和多流两个场景为背景,推导了该方法实行的可行性条件,并通过仿真验证和与基于L-interfering模型的方法对比说明,以此可行性条件为基础的PIC-GIA方法在扩大适用范围和利用空间资源提升性能方面的重要作用。(3)以可行性条件为约束条件,以最优化系统传输速率为目标,利用整数规划的相关知识提出相应的干扰选择和优化策略。由于系统速率直接取决于接收机处理之后的信干噪比(Signal to Interference and Noise Ratio,SINR),因而,这里通过最大化处理后的SINR来确定选择策略,并近似推导出在以可行性条件为约束条件的情况下,通过选择部分干扰对齐和消除能够达到的最优系统速率。同时,本文还提出了两种简化选择策略,使其能够在逼近基于最大化SINR的性能的同时,降低复杂度。
[Abstract]:In recent years, people have found an important problem when studying the traditional interference alignment technology, that is, this method needs enough dimension space to align and eliminate interference. This problem makes it difficult for traditional interference alignment techniques to be used in large scale systems. Therefore, the focus of the research has been shifted to the generalized interference alignment technique. The core idea of this technique is to make use of the difference of channel fading and to align only the more influential interference. So that the number of interference needed to be aligned can be controlled within a certain range, so interference alignment can be carried out. According to this idea, many methods are derived, among which the more popular one is the interference alignment method based on L-interfering model, but this method is also limited in its application scope. Aiming at this limitation, this paper proposes a generalized interference alignment method based on partial interference cancellation, that is, each user determines the feasibility condition of the system according to the configuration and requirements of the user and the influence of each interference. According to the condition, the selected part of the interference is aligned and eliminated. Compared with the existing methods based on L-interfering model, the advantages of this method are: first, it is not necessary to require every user to have exactly the same number of streams sent by each user, so that the number of interference that can be aligned by each user does not need to be the same, and second, the number of interference that can be aligned by each user does not have to be the same. In the use of space resources, this method can be flexibly adjusted according to the space resources owned by the user, because it fully takes into account the differences in system configuration and the number of sending streams when it is designed. Make full use of resources, so that better performance can be achieved both from the system as a whole and from a single user; at the same time, because of the flexibility of the method in resource scheduling, The method can also be combined with other optimization objectives in the process of implementation, so as to meet the needs of some special performance optimization. In this method, the feasibility condition is the basis, and the interference selection strategy based on the condition is the core step. According to these two points, this paper mainly does the following work: 1) analyzes the limitations of the existing L-interfering model in applicability, and proposes a generalized interference alignment method based on partial interference cancellation. The paper also analyzes its function in expanding the scope of application and saving space resources theoretically. (2) using the knowledge of graph theory and taking two scenarios of single flow and multi-flow as the background, the feasibility conditions of this method are deduced. The simulation and comparison with the method based on L-interfering model show that the PIC-GIA method based on this feasibility condition plays an important role in expanding the scope of application and using space resources to improve its performance. Aiming at optimizing the transmission rate of the system, the corresponding interference selection and optimization strategy are proposed by using the relevant knowledge of integer programming. Since the system rate is directly dependent on the signal-to-noise ratio (signal-to-noise ratio) after processing by the receiver, the selection strategy is determined by maximizing the processed SINR, and the selection strategy is approximately deduced under the condition that the feasibility condition is taken as the constraint condition. By selecting partial interference alignment and eliminating the optimal system rate that can be achieved. At the same time, two simplified selection strategies are proposed to reduce the complexity while approximating the performance based on maximization of SINR.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN972

【参考文献】