华中科技大学博士学位论文MIMO-OFDM系统中的多用户自适应资源分配技术研究姓名：郭磊申请学位级别：博士专业：信息与通信工程指导教师：朱光喜20080428II摘摘 要要 移动通信在近 20 年获得飞速的发展，面向个人通信技术的研究成为当今的热点研究领域。在这些研究热点中，如何应付移动通信用户数的增长和多媒体业务需求增加所带来的频谱资源紧张问题成为大家关注的焦点。多天线（MIMO）和正交频分复用（OFDM）技术的结合已经被广泛认为是未来移动通信系统中的主要候选技术。本论文在国家“十五”863 计划重大课题“新一代蜂窝移动通信系统无线传输链路技术研究（2003AA12331005） ”和国家自然科学基金重大项目“未来移动通信系统基础理论与技术研究”子课题“基于 MIMO-OFDM 系统的空中接口自适应技术研究（60496310） ”以及“十一五”国家 863 项目“具有公平性与 Qos 保障的高效MIMO-OFDM 传输技术研究（2006AA01Z277） ”的资助下，展开了对多用户MIMO-OFDM 系统中自适应资源分配问题的研究，并希望能以更优化的解决方案实现 MIMO-OFDM 系统的实用性。 首先，论文分析了无线环境中信道传输的特性以及 MIMO-OFDM 系统模型。以提高系统性能为出发点，将系统所需误码率和各个用户传输速率作为前提条件，建立了基于用户个体反馈信道信息的自适应子载波组分配模型。在发射端准确获取各个用户反馈的全部信道信息下，提出了以最小化系统总体发射功率为目标的自适应子载波组分配算法、比特加载算法。利用各个用户的连续子载波相关性，将子载波分配的最小尺度变为子载波组，大大降低子载分配算法的运算量；推导了基于统计波束成型的最优子载波组分配准则，在降低运算量的前提下保证了系统整体性能无明显衰退，从而使该算法更具有实用性。 接着，论文又研究了结合了天线维自由度的自适应子载波组分配算法。子载波组分配算法利用频率维自由度所带来的多用户分集增益将不同的资源分配给用户从而获得系统整体性能的提升，但是在 MIMO 环境中随着天线数的增加，运算的复杂度会有大幅提升。针对这个问题，我们提出了结合天线维自由度自适应选择的子载波组分配思想，通过一种新型的基于容量最优的天线选择策略，提高了多天线的使用效率，减小了各个用户的多天线数量，有效降低运算复杂度。推导了基于天线选择III策略的子载波组分配准则，其使用最优天线子集的策略在不明显降低系统性能的前提下简化了算法复杂度，显著减小了运算量。 然后研究了基于公平性目标的多用户自适应子载波组分配算法。 以最大化系统总体传输速率为目标，提出了结合公平性算法的自适应子载波组分配策略。借鉴于纳什契约公平性思想，推导了基于纳什契约原则的两级资源公平性分配准则，通过在随机用户组内与用户组间的迭代分配，保证系统整体传输速率最大的前提下实现用户间资源的公平分配。分级的资源分配算法可避开传统算法中非线性合并优化带来的技术复杂度大的缺陷，降低了求解的复杂度，提高了算法的实用性；连续相关子载波组分配的思想进一步降低了算法的计算量，并不明显影响系统的整体性能。 最后，论文研究了随机波束成型技术在子载波组分配算法中的应用。在最优资源分配算法的求解过程中，通常要求发射端获取全部的信道信息，而反馈信息所带来的大量额外带宽需求会给实际系统带来很大的压力。针对这个问题，我们提出了一种利用受限反馈信息进行子载波组分配的算法，算法利用随机波束成型技术低反馈带宽消耗特点，提出了一种改进的随机波束成型算法，推导了在该随机波束成型方式下的受限反馈信息子载波组分配准则。改进的随机波束成型算法利用波束先验知识和随机选择的特点，使发射端仅需获取信噪比匹配度等少量信息即可完成自适应资源分配，从而大大节约了反馈信道带宽，同时获得与全反馈信息子载波组分配算法相近的性能，并具有较好的公平性特征。 关键词：关键词：MIMO OFDM 自适应子载波分配 多用户分集 天线选择 公平原则，随机波束成型，信道反馈 IVAbstract With rapid development of mobile communication theory in recent decade, the research on personal wireless communication techniques become more popular. Among these works, how to cope with the shortage of frequency resource coursed by growing number of mobile users and demands of diverse multimedia services is a focal problem. Fortunately, the combination of MIMO(Multiple-Input Multiple-Output) and OFDM(orthogonal Frequency Division Multiplexing) has been comprehensively considered as a major candidate techniques to deal with these problems for future radio communication systems. Supported by the National High Technology Research and Development Program of China under Grant (No. 2003AA12331005 2006AA01Z277) and National Science Foundation of China under Grant No. 60496310, we focus our eyes on the advanced adaptive resource allocation schemes in multi-user environments and expect that our proposed optimal allocation could be helpful to improve the practicability of MIMO-OFDM systems. First, the wireless channel characteristic and the MIMO-OFDM system model are discussed. Conditioned by preplanned BER(Bit Error Rate) and required rate of users diversity feedback information, the adaptive subcarrier group allocation strategy for improving entire system performance is structured. Assuming exact CSI(Channel State Information) is obtained by transmitter, the adaptive subcarrier group allocation and bit loading algorithm for the purpose of transmit power minimization was presented. Benefiting from relativity of coherent subcarriers, the allocation is transformed to subcarrier group assignment, which significantly simplifies the computation. The optimal subcarrier group allocation criteria based on statistical beamforming was deduced and proved that low computational complexity and close system performance can be obtained to assure the practicability. Secondly, a combinational adaptive subcarrier group allocation algorithm considering antenna-dimension is analyzed. The multi-user diversity gain employed by subcarrier group allocation improves entire system performance by assigning diverse resource to corresponding user in frequency-dimension, but it is clear that computational complexity increases with growing number of antennas in MIMO. A new subcarrier group allocation combined with antenna adaptive selection is proposed, which utilizes optimal capacity based antenna selection algorithm to reduce antenna number and achieve more multi- antenna efficiency. Subcarrier group allocation based on antenna selection policy was deduced and it is shown that optimal antenna subset employment simplifies complexity Vwithout obvious performance decline. Thirdly, adaptive subcarrier group allocation based on fairness consideration is investigated. Targeting maximal entire transmit rate, we propose an adaptive fairly subcarrier group allocation. Using nash bargaining fair idea for reference, we deduce a two-level resource assignment scheme based on nash bargaining criteria, which realizes fairly resource allocation with assuring entire transmit rate maximized by way of iterative calculation among and within random user groups. Leveled fair arrangement avoids high complexity caused by traditional none-linear combination optimiz