【作者】 张碧军；

【导师】 朱光喜；

【作者基本信息】 华中科技大学 ， 信息与通信工程， 2006， 博士

【摘要】 近年来,无线移动通信已成为通信领域的研究热点。目前,已有多项技术被人们看好,将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OFDM (Orthogonal Frequency Division Multiplexing)技术,基于空分复用或空间分集,能显著提高系统容量或有效改善系统性能的MIMO(Multiple Input Multiple Output)技术等。正是在这种应用和研究背景下诞生的空时编码(Space-Time Code, STC)技术受到广泛重视。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(2001AA123014)”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)”的资助下,针对恶劣的无线移动通信环境,如信道空间相关、时频选择性衰落等,对多用户环境下的空时编解码传输方案、预编码技术等进行研究,力图通过稳健的解码器设计或者预编码技术的运用改善系统对抗衰落的性能。论文首先基于已有空时相关MIMO信道模型,分析给出了联合空时互相关函数解析表达式,并推导给出了空时分组编码传输方案理论成对错误概率表达式。通过评估收发两端天线阵列间距、倾斜角度及其散射体角度扩展、不同信道时变等因素对空时分组编码传输系统性能影响,进一步验证了该空时相关MIMO信道模型的正确性。同时,已有的空时相关信道模型,或者仅考虑信道空间相关,或者仅考虑信道时变。我们提出了一种混合模型对信道空时相关联合建模。最后,基于上述两种空时相关信道模型,评估给出:当信道时变较小时,大的信道空间相关带来系统性能损失更为严重;而当信道时变增大时,无论信道何种程度空间相关,大的信道时变则主导系统性能损失。由此可见,信道特性,诸如空间相关、时频选择性衰落等,特别是多用户环境下,对空时编解码传输方案设计有着重要影响。在以往的空时分组编码多用户上行系统中,ZF和MMSE解码器设计方案均假定信道准静态衰落,信道一旦时变,则接收端在解码某用户信号时即引入符号间干扰,且这种干扰随信道时变增大而增加。由此提出了两类新解码器设计方案。方案一,基于时变信道下接收信号表达式联合解码,从而去除解码期望用户信号时符号间干扰的影响。方案二,基于ZF思想,构造一变换矩阵,作用该矩阵于接收信号矢量使之对角化,从而达到去除符号间干扰影响之目的。上述两类解码器设计方案仅适用于时变平坦衰落信道、2用户上行系统。接着,将上述适用于2用户系统的新解码器设计方案推广至已有空时分组编码多用户系统框架,替代迭代软入软出多用户检测方案,从而降低计算复杂度。然后,结合OFDM技术,基于每个平坦衰落子载波信道运用上述新解码器设计方案,将其推广至时频选择性衰落信道、多用户上行系统。在各种系统配置下,相对于已有传统解码器,新解码器获得显著性能增益,同时不存在误码平层现象。对于传统的组分层空时结构解码器设计方案,首先利用组检测技术获得期望分组的发射信号,再空时解码获得该分组内的发射符号。Dai提出了一种解码器设计方案,但其本质思想并未改变,即首先运用组检测技术抑制组间干扰,再最大似然解码获得期望分组内的发射符号。已有解码器设计方案均假定信道准静态衰落,信道一旦时变,则不再适用。作为改进,我们提出了一种适用于时变信道、组分层空时结构的新解码器设计方案。对于该方案,首先构造“正交对”矩阵去除所有分组内符号之间的干扰,再运用递推降维矩阵求逆算法,并行检测获得所有分组的发射符号。我们将证明,Zhuang提出的适用于时变信道、4发天线系统的解码器即为本文的特例。与传统的解码器相比较,新设计的解码器计算复杂度得到大大降低。论文最后针对多用户下行空分复用系统,提出了一种预编码器设计方案。在基站发射端,根据最小平均矢量符号错误概率准则,预编码器构造为所有用户信道状态信息的函数。由于多用户干扰在基站端被预去除,所以终端用户仅利用简单的线性解码即可获得发射符号。该多用户下行空分复用系统预编码器设计方案获得了与单用户预编码系统几乎一致的性能。接着,针对频率选择性衰落信道,提出了一种适用于多用户下行MIMO-OFDM系统的预处理传输方案。在基站发射端,通过所有终端用户在每个子载波信道上发射预处理矢量的良好设计,将多用户下行MIMO系统转换为并行独立的单用户MIMO系统,并确保每终端用户的多用户干扰被良好去除。在用户接收端,利用简单的线性解码即可获得最大解码后符号信噪比,从而最小该用户误码率。更多还原

【Abstract】 Recently, wireless mobile communication has been the hot topic in the domain of communication. Nowadays, many techniques have been comprehensively considered as the major candidates of the future radio communication architecture. Among which includes OFDM and MIMO, the former can effectively combats multipath fading and improves transmission rate of data, the latter can effectively enhances system performance or notably improves system capacities using spatial diversity or spatial multiplexing techniques. Space-time code appears at the background of applications and immediately attracts much more attentions. Supported by the National High Technology Research and Development Program of China under Grant No. 2001AA123014 and National Science Foundation of China under Grant No. 60496310, our research work concentrates on the severe communication circumstances, for example, spatially correlative and time-frequency selective fading channel, and focuses on the design of codec for space-time architecture and the utilizing of linear precoding technique especially on the multiuser scenario trying to improve the system performance to combate various fading.Firstly, we provide the analytical expression of the joint space-time cross-correlation function and deduce the theoretical pairwise error probability of space-time block coded transmission scheme based on the existed spatially and temporally correlated MIMO channel model. Also, we investigate the impacts of antennas array spacing, angles, the spread of scatterings at the both ends of the link upon the system performance and further testify the validity of the general MIMO channel model. At the same time, the all existed MIMO channel models considered spatial correlation or temporal variation respectively, we present a hybrid model to model spatial-temporal correlation jointly. Finally, we arrive at a userful conclusion: high spatial correlation dominates the performance loss when the temporal variation is small, whereas large temporal variation contributes to the performance loss whatever degree spatial correlation is based on the two above channel models. Thus it can be seen various channel conditions, such as spatial correlation, time-frequency selective fading etc, have important influences on the space-time transmission scheme especially on the multiuser scenario.For the space-time block coded multiuser uplink, the design of conventional ZF and MMSE decoder assumed channels are quasi-static. When the channels are time-selective fading, these conventional decoders induce to the Inter-Symbol Interference (ISI) in order to obtain one user’s signals, furthermore, the ISI becomes larger resulted from the increase of the channels time variations. So, we propose two kinds of new decoders for the space-time block coded multiuser uplink over time-selective fading channels. For the first kind of decoder, ISI is removed using jointly decoding based on the new received signal expressions over time-selective fading channels. For the second kind of decoder, we construct a transformed matrix using the ZF idea and act this matrix on the received signal vector to result in a diagonal matrix thereby eliminating the influence of ISI. The two above decoders are appropriate to two users system and the channels are flat time-selective fading. Next, we extend the above decoder to the existed space-time block coded multiuser architecture and substitute it for the iterative soft input soft output (SISO) multiuser dectection thus reducing the calculational complexities. Then, the above decoder is used at each flat fading subcarrier channel and is extended to the frequency-selective fading channels combining with OFDM technique. Compared with the conventional decoders, the new decoder obtains notable performance gains and does not exhibit error floors for the various system configurations.For the existed group layered space-time architecture, the conventional group detection technique is applied firstly to suppress the interferences from other groups and space-time decoding is then performed to obtain the desired group’s signals. Dai et al proposed a decoding scheme but the essential ideas do not change, also, exploiting firstly group detection to suppress the other groups’interferences and ML is then used to get the desired group’s signals. The all existed decoders assumed the channels are quasi-static. They are not appropriate once the channels are time-selective fading. So, we present a new decoder appropriate to group layered space-time architecture over time-selective fading channels. For the new decoder, the orthogonal pairs of matrixes are constructed firstly and employed to remove the inter-interference of all groups then use the low complexity matrix inverse algorithm to obtain the all groups’transmitted signals by parallel detecting. It is also proved that the old decoding scheme proposed by Zhuang et al is a special case of our decoder. Compared with the conventional decoders, the new decoder reduces greatly the calculational complexities.Finally, a precoding scheme is proposed for the multiuser downlink spatial multiplexing system. At the transmitter, the precoder is constructed as the function of all users’channel state information according to the performance criterion of minimizing average probability of vector symbol error. Since the multiuser interferecnce (MUI) is precancelled at the transmitter, each terminal user just utilizes the simple linear decoding to obtain the transmitted symbols. It is verified that our proposed precoding scheme arrives at almost the same performance as the single user precoding system. Next, we present a transmission scheme for the multiuser downlink over frequency-selective fading channels. At the transmitter, the multiuser MIMO system is transformed to parallel independent single user MIMO system through the perfect design of transmit preprocessing vectors at the subcarrier channel thus assuring the good elimination of MUI at each terminal user. At the receiver, use just simple linear decoder to achieve the maximum signal-to-noise ration of the decoded symbol thus minimizing the bit error rate of each user.更多还原