【作者】 杨代明；

【导师】 张立明；

【作者基本信息】 复旦大学 ， 电路与系统， 2004， 博士

【摘要】 现代通信系统的发展非常迅速，在技术和市场上都获得了极大的成功。但是，高速的无线终端接入问题仍然没有解决，这使得人们通过无线技术传输文本，语音，图像等综合多媒体业务的梦想难以实现，为了提供高速的无线终端接入方法，在接收和发射端都采用多天线的MIMO无线技术就应运而生。 本文研究面向未来无线通信应用的MIMO系统的基带处理算法，其中包括MIMO均衡，空时编码，纠错编码三部分，MIMO系统的设计目标是能够提供高速率低延迟的传输能力。这也就是MIMO系统中均衡，空时编码，纠错编码等各个部分的设计目标。 MIMO系统的传输环境是移动的多径衰落环境，因此接收端必须消除接收信号的符号间干扰利通道间干扰。有多种均衡技术可以实现这个目的。目前大量采用的方法是训练方法，这种方法的缺点是要消耗大量的传输带宽。而其他的不需要训练的肓方法对信号结构或者噪声特性等有特殊的要求，应用场合有限，而且在误码性能，收敛性能和鲁棒性方面都各有缺点。 本文关于MIMO均衡的工作是提出了采用子空间跟踪的肓均衡方法和广义高斯近似的最小互信息的肓均衡方法。子空间跟踪的肓均衡方法用子空间跟踪技术将传统的的子空间方法改造为一个自适应方法，便于实时应用。另外，根据均衡器输出信号具有广义高斯近似分布的特点，本文提出了广义高斯近似的最小互信息的肓均衡方法。这种方法不需要对发射信号的结构进行限制，较之恒模盲均衡算法有更大的使用范围。 在MIMO系统空时编码方面，本文提出正交线性空时弥散码。正交线性空时弥散码运用正交矩阵的列矢量构造空时正交调制基矢量或者基矩阵。这种方法设计简单，便于采用相干检测的解调方法，调制以及解调都易于实现，而且误码性能优于一般空时弥散码。根据不同的传输需求，正交线性空时弥散码的形式还可以灵活变化。 纠错编码技术也是MIMO系统设计的一个关键问题。传统的无线通信系统采用的纠错技术是二进制卷积码(或者Turbo码)和多进制RS码的级联。这种与法的性能比较好，但是结构复杂，实现不了高速传输。本文建议了基于局部概率传播译码的低密度卷积码和RS码的级联方案。 低密度卷积码结合了低密度码性能良好和译码简单的特点和卷积码产生矩阵具有带状结构的特点，具有优于一般卷积码的性能。而其译码器更为简单，可以采用流水线结构，避免了低密度码的译码延迟大，卷积码和Turbo码译码复杂复!l_人学卜毕卜羊位论文}‘若」};l{题·计对高传输速率低延退!!勺设计{!标，训亢J‘采川八均枷方法，比交线性空}}」弥散码，‘氏密度卷积码和RS码级联红}错的MIMO系统的筷体性能。仿典实验表}叫，本文的M IMO系统万‘篇衬t一传输吞吐率和传输延迟性能方}闭优{一日前址有代夕仁}:的欧洲工5‘!项{川‘的讯朋系统方案。更多还原

【Abstract】 Modern communication system is now developing very fast and has get great success in both technology and commercial field. But there are still many technical problems to be solved for the dream of real time transmitting of multimedia service composed of texts, voice and video. One of these technical problems is high speed wireless access. Many wireless technologies has been proposed, the most attracting one is MIMO technology, which use array antenna at both the transmitting and receiving end.This dissertation focuses on algorithms for a base band processor of MIMO system. The base band processor includes three parts as MIMO equalization, MIMO modulation and de-modulation , and error correcting codes for MIMO system. The optimal design goal of the MIMO base band processor is the ability to provide high rate and low latent wireless transmitting .This global optimal goal rules the optimal design of the three parts of the base band processor.The transmitting environment of MIMO system is always a mobile and multi path environment, so the receiver must get rid of the inter symbol interference and the inter channel interference. There are many equalization method has the ability to cope with this problem. Most of them use training sequence, this leads to a large portion of band width waste. There are also some blind methods for MIMO equalization, but most of them has special requirement on signal structure or noise characteristics. This limits the application of these blind methods. This dissertation proposes MIMO blind equalization with subspace tracking and general Gaussian approximated minimum mutual information blind equalization method. MIMO blind equalization with subspace tracking use subspace tracking technology to modify the subspace method to be a adaptive one so as to be used real timely. Based on the observation that the output of MIMO equalizer has probability distribution very near to high order general Gaussian distribution, this dissertation proposes general Gaussian approximated minimum mutual information MIMO blind equalization. This method needs no restriction on the structure of the transmitting signal, so has wider application field than Constant Modulus algorithm.For the joint diversity and multiplexing optimal MIMO transmitting, this dissertation proposes orthogonal linear dispersive space time code. Orthogonal linear dispersive space time code constructs orthogonal base vector set of base matrix setwith orthogonal column vectors of a orthogonal matrix. This method is very simple in design, modulation and demodulation. The coherent detection of orthogonal linear dispersive space time code based on orthogonal base vectors also improves its BER performance. Orthogonal linear dispersive space time code can also be used as an adaptive one for different application requirement.Another challenge for MIMO base band processor design comes from error correcting codes. In many communication systems , the error correcting codes is concatenation of RS code and convolution code or Turbo code. But convolution code and Turbo code are very difficult to provide high speed transmitting. This dissertation proposes the approach of concatenation of RS code and low density convolution code.Low density convolution code combines the characteristics of convolution code with generating matrix in band form and characteristics of low density parity code with good performance and simple decoding. Low density convolution code has performance optimal to general convolution code and its decoding is very easy to be implemented in pipelined forms. This successfully avoids the decoding delay of low density parity check code.At last the dissertation investigates a MIMO base band processor solution with blind equalization, orthogonal linear dispersive space time code and low density convolution code for optimal design goals of high transmitting rate and low transmitting delay.Simulation result shows that the proposed MIMO system solution has transmitting through put and transmitting delay superior to that of the MIM更多还原