华中科技大学 博士学位论文 新一代无线通信系统中的传输方案及资源分配研究 姓名：肖啸 申请学位级别：博士 专业：信息与通信工程 指导教师：朱光喜 20061029 I 摘摘 要要 近年来，无线移动通信已成为通信领域的研究热点。当前，已有多项技术被人们 看好，很有希望成为未来移动通信系统架构中的关键技术。这其中包括能有效克服 多径效应，提高系统数据传输速率的正交频分复用(OFDM, Orthogonal Frequency Division Multiplexing)技术；以及能显著提高系统容量或有效改善系统性能的多输入 天线、多输出天线(MIMO, Multiple Input Multiple Output)技术等。本文在国家“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 “基于 MIMO-OFDM 系统的空中 接口自适应技术研究(60496315)”的资助下，针对恶劣的无线移动通信中，多天线、 多用户环境以及多业务服务质量（QoS, Quality of Service）要求下的空间发送分集技 术、MIMO 传输方案以及自适应资源分配技术等进行了创新研究。 本文首先概述了当前 MIMO 信道中的空间分集技术，特别是发送分集技术。接 着针对发送端已知信道状态信息时的最大比发射(MRT，Maximum Ratio Transmitting) 技术，提出一种在基站端采用最优选择最大比发射(OMRT, Optimum -Selection Maximum Ratio Transmitting)的发送分集技术，从而在满足性能要求的同时，降低基 站的硬件复杂度、功耗等。在此基础上，我们利用随机矩阵(Random Matrix)和排序 统计(Order Statistics)的理论知识，对该技术进行了性能分析，得到瑞利衰落信道下 中断概率(Outage Probability)、误码率(BER, Bit Error Rate)等性能指标的表达式。 其次针对 MIMO 通信中，由于尺寸大小、功率能量以及生产成本等方面的限制 而造成当前大多数手持设备不太可能配置多的天线数（或射频电路数）这一现实特 点，通过收、发端的联合设计，提出两种新的 MIMO 传输方案，从而达到提高系统 性能， 降低收发端的硬件复杂度、功耗等目的。首先， 我们提出一种在基站按照 MRT 技术发射信号，而在移动台端基于信噪比(SNR, Signal-to-Noise Ratio)最大的原则仅 选择一根接收天线来接收、处理信号的传输方案(即 MRT/RAS, Maximum Ratio Transmitting/ Receiver Antenna Selection Combining)。接着，我们又提出在基站采用 选择发射联合移动台端采用最少选择最大比合并接收的传输方案(即 TAS/MMRC, II Transmitting Antenna Selection /Receiver Minimum-Selection Maximum Ratio Combining)。分析和仿真结果均表明 MRT/RAS 方案可以取得很好的阵列增益(Array Gain)及满分集增益(Full Diversity Gain)，在相同频谱条件下，性能超过某些复杂的 空时编码系统。同样我们对 TAS/MMRC 方案也进行了数学分析。这些结论既表明 我们所提出的传输方案非常满足当前对 MIMO 传输技术的要求，同时也为它们的实 际应用提供了理论指导。 我们还研究了基于特征波束成形传输策略的 MIMO-OFDM 系统中的资源分配问 题。首先在保证用户误码率的条件下，以总的发射功率最小化为目标，提出多用户 MIMO-OFDM 系统中的自适应资源分配准则，并对此时的系统性能和取得的多用户 分集(Multiuser Diversity)增益进行了定量分析。在此基础上，还将考虑用户的传输速 率要求，考虑到算法的实用性和复杂度，给出了一种低复杂度的次优分配算法，使 得在用户 QoS 要求得到保证的基础上系统总的发射功率最小。此外，我们还研究了 相关衰落信道下 MIMO-OFDM 系统中的自适应资源分配策略。 最后，在不同信噪比条件下，我们讨论了使多用户 MIMO-OFDM 系统（包括实 际应用的 V-BLAST(Vertical Bell Laboratory Layered Space-Time)/OFDM 系统）的瞬 时容量最大的资源分配准则。提出一种简单的资源迭代分配算法，从而在考虑用户 QoS 要求的同时，最大化 V-BLAST/OFDM 系统的容量。最后为了改善系统在低信 噪比下的频谱效率，还将联合功率再分配算法来进行自适应的资源分配。 关键词关键词： MIMO OFDM 空分复用 分集 服务质量 自适应资源分配 III Abstract Current MIMO spatial diversity techniques, especially the spatial transmit diversity techniques. For the maximum ratio transmitting (MRT), which the transmitter knows the CSI(Channel State Information), we then propose a new transmit diversity scheme, namely optimal-selection maximum ratio transmit (the OMRT scheme), to meet the system performance and reduce the hardware complexity, power consume further. Based on the theory about random matrix and order statistics, we carry out a though analysis for the OMRT. The closed-form outage probability and the BER (Bit Error Rate) of the OMRT system, for Rayleigh fading scenario, are presented. The analytical results are verified by simulation. Second, because the conventional MIMO systems require the number of radio frequency circuits should be equal to the number of antennas. Obviously, that will lead the system hardware complexity, cost and power consumption increase deeply, especially for the mobile receiver. So with the transmitter and the receiver joint design, two new MIMO schemes are proposed to improve the system performance and reduce the hardware complexity, cost and so on. We investigate a downlink MIMO scheme combining maximum ratio transmission and receiver antenna selection combining, in which a single receive antenna, maximizing the receive SNR, is selected for demodulation. We refer to it as the MRT/RAS scheme. Then we investigate another MIMO scheme, in which base station selects a transmit antenna with the largest SNR and mobile set adopts the MMRC (Minimum-Selection Maximum Ratio Combining) technology, namely the TAS/MMRC scheme. The analysis and simulation demonstrate that the MRT/RAS scheme can achieve array gain and full diversity order gain, as if all the receive antennas were used. It is shown that the MRT/RAS scheme outperforms some more complex space-time codes of the same spectral efficiency. We also carry out a though analysis for the TAS/MMRC. Extensive analysis and simulation results are presented to validate the analysis and shown that the two schemes are suitable for the downlink communications in cellular radio systems and provide theory guidance for their practical applications. Moreover, we study the adaptive resource allocation for the multiuser MIMO-OFDM IV system based on eigen-beamforming. First, only considering the users BER requirements, we propose an adaptive resource allocation criteria to minimize the total system transmit power.We quantitatively analyze the system performance, including the obtained multiuser diversity gain. Then we propose a hypo-optimal resource allocation algorithm while ensuring the fulfillment of users QoS requirements and considering the schemes practicability and computational complexity. Furthermore, we study the adaptive resource allocation for the correlated MIMO-OFDM systems via a new channel model, called virtual channel representation, which con