【作者】 孙胜贤；

【导师】 龚耀寰；

【作者基本信息】 电子科技大学 ， 信号与信息处理， 2004， 博士

【摘要】 随着无线通信需求的快速增长,当前容量有限的单天线收发无线通信系统正面临着严峻的考验,必须进一步提高当前无线通信系统的频谱利用效率和通信的质量。MIMO无线通信系统(后简称为MIMO系统,并特指单用户无记忆MIMO系统)为拓宽这个瓶颈、解决这个挑战提供了新的思路和方法。从提出到现在的近10年中,MIMO系统得到了迅猛的发展。当前,单用户无记忆环境下的MIMO技术日渐成熟。在实际无线通信系统中,通信系统主要有害因素不但包括热噪声,而且包括多用户所引起的共道干扰(CCI)和信道记忆所引起的符号间干扰(ISI),因此,记忆MIMO系统、多用户MIMO系统和多用户记忆MIMO系统与单用户MIMO系统相比有本质上的区别。在上述三种有干扰MIMO系统中,若不进行对干扰的预处理,干扰将对目标信号的检测与估计产生很大的影响,这将导致系统性能急剧下降。为了尽可能地减小系统性能受干扰的影响,采用抗干扰技术是很有必要的。以前,大量的研究是基于单用户MIMO系统或单用户记忆MIMO系统的。这些系统要么没有考虑ISI,要么没有考虑CCI或者两者的联合影响,如著名的BLAST系统,此系统能否在多用户环境中应用仍是一个值得研究的问题,甚至就是最复杂的空时频编码,虽然充分考虑了系统中存在的ISI,对单用户而言是优化的,但系统却完全忽略了CCI的影响。本文首先考虑记忆MIMO系统与多用户记忆MIMO系统的互信息容量,然后考虑在前述干扰系统中(部分)利用ISI和抗CCI的技术——(空时)波束形成。由于MIMO系统与传统阵列系统的差异,MIMO系统波束形成概念也与传统波束形成的概念存在很大的不同,MIMO系统波束形成不再是传统意义的实际空间上的波束形成,而是抽象的矩阵空间波束形成。由于波束形成能在干扰MIMO环境保持非常强的抗干扰能力,甚至在某些特殊环境能够同时实现优化的分集与复用,因此在未来无线通信中具有十分广泛的前景。 <WP=8>本文的主要创新工作在于:针对分块瑞利衰落记忆MIMO系统,在不同记忆成分具有独立的高斯统计特性时,利用矩阵理论和多元统计知识,获得系统互信息容量的随机表达式。然后利用容量的随机表达式,得到在信息帧长远大于信道记忆阶数时系统的信道极限容量近似正比于信道记忆阶数的结论,为实际应用提供了一个理论上的参考。针对分块瑞利衰落多用户记忆MIMO系统,在不同用户间的不同记忆成分有独立的高斯统计特性时,利用优化理论、多元统计理论、矩阵理论和群论,获得整个系统在不同信干噪比矩阵下全系统中每个用户所需要选择的优化(对全系统优化,并非对个体本身优化)发射方式的判决策略,并详细分析了几种极端情况下容量的表达式,为未来商用多用户记忆MIMO系统提供一个理论上的容量指标。针对记忆MIMO系统、多用户MIMO系统和多用户记忆MIMO系统不同的干扰特性,引入了具有很强抗干扰能力的联合收发最大比波束形成,得到系统的联合收发最大比波束形成的闭式解或数值解,并详细地分析了系统在各种环境下的联合收发最大比波束形成的抗干扰性能,为MIMO系统点对点直接通信(不需要任何中继)的商用提供了具体的算法。针对多用户MIMO系统考虑了对算法复杂度要求较高但不需要估计信道衰落幅度的联合收发等增益波束形成。由于系统受等增益约束,联合收发等增益波束形成不可能得到闭式或解析解,本文利用数值迭代优化方法求得联合收发等增益波束形成的数值优化解,并详细地比较了算法的抗干扰性能,为MIMO系统的小型化或特殊的应用提供了具体的算法。考虑到MIMO系统与现行的蜂窝无线通信系统的兼容性,详细分析了收发两端约束不对称的多用户MIMO系统的联合波束形成算法——基站与移动终端的上行链路的联合EGT/MRC和下行链路的联合MRT/EGC算法。虽然上行链路和下行链路干扰环境<WP=9>和实际约束的差别导致算法的复杂度在不同链路间存在很大的差别,但系统却能保持良好的抗干扰性能。这一部分可以为未来MIMO系统的实用和商用提供与当前蜂窝通信系统兼容的算法。更多还原

【Abstract】 The Rapidly increasing requirement of wireless communication demand and better quality of service defy current single-antenna transmission- receiving communication system for the limited capacity and low spectra efficiency. In order to increase the wireless channel capacity and spectra efficiency, MIMO wireless communication system is resorted to break the bottleneck of channel capacity and enhance spectra efficiency.During the past 10 years, from the born to current, the MIMO wireless communication system was developed very rapidly. The research in Rayleigh fading channel without memory is near the climax now. In realistic wireless communication channels, the main impairment is not thermal noise, but the co-channel interferences and inter-symbol interference. The MIMO system with memory, multi-user MIMO system and multi-user MIMO system with memory are much different from single-user MIMO system without memory. Since the target signal detection and estimation are affected by the interferences, it is important to use anti-interference technology in MIMO system so as not to degrade the system performance dramatically.Most of researches are assumed that the channel is quasi-static fading or with single-user, all of them ignored the effect of co-channel interference or inter-symbol interference or the both. It is still a problem that some systems, such as BLAST, can be used or not in multi-user case. Even the most complex space-time-frequency codes make use of the ISI and it is optimum for single-user MIMO system with memory, they ignore the effect of CCI in multi-user case. This dissertation will first consider the mutual capacity of MIMO system with memory and multi-user MIMO system with memory. Then (space-time) beamforming will be used to (partly) make use of the ISI and mitigate the CCI. The definition of beamforming in MIMO system is much different from the <WP=11>classic because of the difference between the MIMO wireless communication system and the classic arrays. The beamforming in MIMO system is not the beamforming in (classic) realistic space but the matrix space. As the beamforming in MIMO system can keep good performance in interference cases, even it can get both optimum diversity and multiplexing in some special case, it will have wide application in the future.The main contributions of this dissertation are:In Rayleigh block-fading MIMO system with memory, the matrix theory and multivariate statistical theory are used to derive a stochastic formula of channel mutual capacity with assumption of independent Gaussian statistic character between different memory components. We conclude that the limiting mutual capacity is scale with the order of channel memory if the data frame is much longer than the order of memory. It can be a useful reference for application in future.Using majorization theory, multivariate statistics analysis, matrix theory and group theory, we derive the optimum (for the whole system, rather for the individual) transmission scheme to be selected for each user in the whole system according to the special signal to interference and noise ratio (SINR) matrix in Rayleigh block-fading multi-user MIMO system with memory with assumption of independent Gaussian statistic character between different memory components and different user. Some extreme cases are analyzed in detail. It can be a useful reference capacity for MIMO commercial application in future.Due to the different characteristic of CCI and ISI, i.e., any CCI is deconstructive while ISI can be (partly) available in some schemes devised properly, we apply the joint transmission and receiving maximal ratio (MR) beamforming to (partly) make use of ISI and mitigate CCI in MIMO system with memory, multi-user MIMO system with or without memory. The closed-form optimal weight of joint transmission and <WP=12>receiving adaptive maximal ratio beamforming is formulated in multi-user MIMO system. As the closed-form optimum weights of joint transmission and receiving MR beamforming do not exist in MIMO system with memory a更多还原