低复杂度大规模MIMO检测算法的研究

发布时间：2018-06-07 18:44

本文选题：大规模MIMO + 检测算法　；参考：《西安电子科技大学》2014年硕士论文

【摘要】：在发送端和接收端都有多根天线的MIMO系统由于其在发射分集和数据速率等方面的优势,变得越来越受欢迎,从而发展到今天已经相当成熟了。而为了进一步提高频谱效率和信道容量,提出了以大型阵列天线代替原来的几根天线的大规模MIMO系统。目前对大规模MIMO系统的研究尚处于起步阶段,而接收端的检测技术作为MIMO系统的一个重要组成部分,对其检测算法的研究将会极大地推动对大规模MIMO研究的发展。传统MIMO系统的检测算法包括ZF、MF、MMSE和ZF-SIC等算法。其中球形检测算法在以高复杂度为代价的基础上可以获得近似ML检测的性能。而本文研究的大规模MIMO系统的发射天线和接收天线数量可以达到数十乃至数百根,若采用传统的MIMO检测算法,其检测复杂度将会变得难以忍受,因此寻找适用于大规模MIMO系统的低复杂度的检测算法成为关键。本文首先介绍了大规模MIMO技术的研究背景和研究现状,指明了全文的研究方向。接着对大规模MIMO系统进行了简单的介绍。主要简单地分析了大规模MIMO系统的总速率上界问题,探讨了大规模MIMO系统的关键技术,主要包括信道估计技术、下行预编码和上行检测技术,从技术特点和研究现状两个方面进行了介绍。同时还讨论了大规模MIMO系统未来有待进一步研究的难点和热点问题。本文重点对低复杂度的大规模MIMO的检测算法进行了研究。详细介绍了基于本地邻域搜索的似然上升搜索(LAS)检测算法的基本原理和实现过程。接着从三个方面分析了LAS算法的复杂度,并在不同SNR或者不同天线数的条件下对LAS算法的检测性能进行了仿真验证。然后介绍了一种基于LAS的MOS-LAS算法,该算法通过生成多个不同的局部最小点并从中找出最优值的方式达到了改善的目的。根据生成局部最小点的方式不同又细分为MIV-LAS算法和MSCS-LAS算法。本文对MIV-LAS算法和MSCS-LAS算法的思想进行了详细地介绍,并通过仿真验证了算法的有效性。最后还介绍了另外用一种低复杂度的主动禁忌搜索(RTS)检测算法。本文详细阐述了RTS算法的算法原理和实现过程,并从算法原理和复杂度等方面将RTS算法和LAS算法进行了全面地比较,并通过仿真验证了RTS算法的检测性能,得出了RTS算法的检测性能优于LAS算法的结论。
[Abstract]:MIMO systems with multiple antennas at both transmitter and receiver have become more and more popular due to their advantages in transmit diversity and data rate. In order to further improve the spectral efficiency and channel capacity, a large scale MIMO system with large array antennas instead of several original antennas is proposed. At present, the research of large-scale MIMO system is still in its infancy. As an important part of MIMO system, the detection technology of receiver will greatly promote the development of large-scale MIMO research. The detection algorithms of traditional MIMO systems include the algorithms of MMSE and ZF-SIC. The spherical detection algorithm can obtain approximate ML detection performance at the cost of high complexity. However, the number of transmit and receive antennas in the large scale MIMO systems studied in this paper can reach tens or even hundreds. If the traditional MIMO detection algorithm is used, the detection complexity will become more and more intolerable. Therefore, it is the key to find low complexity detection algorithms for large scale MIMO systems. Firstly, this paper introduces the research background and research status of large-scale MIMO technology, and points out the research direction of this paper. Then the large-scale MIMO system is introduced briefly. In this paper, the total rate upper bound of large scale MIMO systems is briefly analyzed, and the key technologies of large scale MIMO systems are discussed, including channel estimation, downlink precoding and uplink detection. The technical characteristics and research status are introduced. At the same time, the difficulties and hot issues that need further research in the future of large scale MIMO systems are also discussed. In this paper, the detection algorithms of large scale MIMO with low complexity are studied. This paper introduces the basic principle and implementation process of the LAS-based detection algorithm based on local neighborhood search. Then the complexity of Las algorithm is analyzed from three aspects, and the detection performance of Las algorithm is simulated under different SNR or different antenna number. Then an algorithm based on Las for MOS-LAS is introduced. The algorithm improves the algorithm by generating several different local minimum points and finding out the optimal value. The MIV-LAS algorithm and MSCS-LAS algorithm are subdivided into MIV-LAS algorithm and MSCS-LAS algorithm according to the different ways of generating local minimum points. The idea of MIV-LAS algorithm and MSCS-LAS algorithm is introduced in detail, and the validity of the algorithm is verified by simulation. Finally, a low complexity active Tabu search (RTS) detection algorithm is also introduced. In this paper, the algorithm principle and realization process of RTS algorithm are described in detail, and the RTS algorithm and Las algorithm are compared in terms of algorithm principle and complexity, and the detection performance of RTS algorithm is verified by simulation. It is concluded that the detection performance of RTS algorithm is better than that of Las algorithm.
【学位授予单位】：西安电子科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN919.3

【相似文献】