【作者】 陈亮；

【导师】 李建东；

【作者基本信息】 西安电子科技大学 ， 通信与信息系统， 2009， 博士

【摘要】 未来移动通信对于高速数据的传输业务的需求更加紧迫。为此B3G乃至4G中多输入多输出(MIMO),正交频分复用(OFDM),信道编码及迭代检测等技术成为解决一系列相关问题的关键技术。MIMO技术可以在不占用额外带宽的情况下提高系统的频谱效率；迭代检测技术结合了检测与译码,通过迭代在降低复杂度的同时有效保证了系统性能。本文主要针对MIMO系统中的迭代检测问题进行了研究。本文的主要内容和成果如下：1.提出了一种应用于MIMO系统的基于QR分解算法的新型低复杂度迭代接收机。接收端首先通过对信道矩阵QR分解对接收信号进行变换,根据QR分解变换后的上三角形式采用分层检测,对不同层分组并进行部分并行干扰抵消可以同时抵消掉多层信号,组内采用MAP联合检测输出软信息。通过交织解交织器将软入软出检测器与Turbo迭代译码器级联起来,经过迭代有效的消除了多天线间的干扰。与传统基于QR分解的串行干扰抵消迭代接收机相比性能有了极大提高,并且受信噪比估计误差的影响较小。2.基于QR分解的最优排序无法直接获得,需要通过改变排序顺序搜索获得。针对最优排序复杂的问题提出了一种低复杂度的基于QR分解的BLAST险测算法。该算法避免了传统分层空时检测算法中对信道矩阵求逆的过程,减小了运算复杂度,同时克服了排序QR分解检测复杂度高的问题。在采用次优排序的基础上通过串行干扰抵消逐层去除最小错误概率的信号,减小了抵消过程中误差传播的概率,提高了系统的误比特性能。该算法性能优于传统的分层空时检测算法以及次优排序QR分解检测算法。将该算法应用到迭代系统可以有效提高检测输出比特外信息的互信息。3.通过结合HBLAST与VBLAST系统的特点以及Turbo编码的结构提出了一种新的应用于MIMO系统的BLAST收发系统TLBLAST,实现了高速数据传输,在保证分集增益的同时每根天线可以采用不同的编码方式对数据实行不同等级的保护。接收端将软输入软输出检测器与迭代译码级联起来,经过迭代有效的提高了系统性能,并且与已有的VBLAST系统迭代接收结构复杂程度相当。此外针对基于串行抵消结构的迭代接收算法的特点提出了非规则调制的MIMO系统,在保证总的传输速率一致的前提下对每根天线上分配不同的调制方案,通过仿真表明采用非规则编码调制使系统误码性能得到了极大的提高。4.研究了MIMO系统在接收端所知信道信息非理想情况下的性能并提出了改进方案。针对非理想信道信息提出了一种应用于多输入多输出(MIMO)系统的新的迫零接收机。在接收端已知信道信息估计值的情况下,通过引入信道误差的统计信息,修正了传统迫零接收的滤波系数。当接收端无法获取正确的信道信息,该接收机受信道信息误差的影响较传统的迫零接收机小,具有较好的鲁棒性。5.对基于线性滤波结构的迭代系统和带干扰抵消的迭代系统在收端非理想信道信息下性能进行了分析。为了获得最佳的迭代性能,通常采用迭代系统采用特殊的符号映射方式如反格雷映射,然而此类映射方式只有在迭代通道打开的前提下才有效。当信道信息存在误差时,采用此类映射很容易失效。通过分析得出采用格雷映射的迭代系统更加健壮,当估计的信道信息的偏差较大时,采用格雷映射的迭代系统反而性能更好。基于此提出了实际情况下根据信道估计质量改变映射的自适应系统,通过切换映射方式或采用混合映射使系统迭代过程中输出的外信息更可靠,从而非理想信道信息下更加健壮,有效提升系统性能。此外改进了非理想信道信息下SIC-MMSE迭代算法,通过修正信号估值的方差的计算,使得输出的外信息更为可靠,仿真表明采用修正的算法系统性能得到改进,优于传统算法。更多还原

【Abstract】 There is an urgent need for the service of high data rate transmission in the future mobile communication systems. For solving relative problem,Multiple Input Multiple Output (MIMO),Orthogonal Frequency Division Multiplexing(OFDM), channel coding and iterative detection technologies are taken as the key technologies in B3G and 4G. Wireless communication systems with MIMO have the potential of vastly improving spectral efficiency without expanding the bandwidth. The detection and decoding are considered in iterative detection technology. By iteration, the performance of system is improved with low complexity. In this thesis, the iterative detection in MIMO is studied. The main research results of this dissertation are as follows:1. A novel low-complexity iterative receiver based on QR decomposition in Multiple Input Multiple Output (MIMO) system is presented. In the receiver, the channel matrix is decomposed and the received signal is processed. According to the upper triangular matrix after QR decomposition, the signal is detected layer by layer. Partial parallel interference cancellation based on QR decomposition is used, in which different layers are grouped and jointly detected by MAP. The interference between different antennas is cancelled effectively through serially concatenating a soft input soft output (SISO) detector and Turbo decoder by iterative process. Three dimension extrinsic information transfer characteristics (3D EXIT charts) are used to analysis the processes of iterative. Compared with traditional QR based serial interference cancellation iterative receiver, the performance of the proposed receiver is improved greatly. And it is not sensitivity to the mismatched SNR.2. To find the optimal order of QR-decomposition, the columns of channel matrix should be permutated and it can only be obtained by search method. An improved QR based detection algorithm is proposed for use in the Bell Labs Layered Space-Time (BLAST) systems. It doesn’t require pseudo inverse of matrix, then has less computational load in contrast to the standard detection algorithm with inverse of matrix. It has less computational complexity than ordered QR-based detection .The probability of error propagation is reduced through serial interference canceling; The signal of minimum error probability is removed layer by layer. Simulation shows that it outperforms the classic V-BLAST and ordered, suboptimal ordered QR based detection algorithms. Combined with the iterative decoding, the extrinsic information delivered by detector will be improved further.3. A novel BLAST transceiver named Turbo-Like BLAST (TLBLAST) for MIMO communications is proposed, which combines the characteristics of HBLAST and VBLAST with the structure of Turbo encoder. The high data rate transmission can be implemented and in each transmitted antenna, different encode schemes can be used to supply different protection levels. The system performance is improved effectively through serially concatenating a soft input soft output (SISO) detector and decoder by iterative process with comparable complexity of VBLAST. Simulation results show that the performance of TLBLAST is better than HBLAST and VBLAST in Rayleigh flat fading channels. Aimed at the iterative receiver based on serial cancellation, an irregular modulation scheme in MIMO is proposed. In each antenna, different modulation is assigned with the same transmission rate. Simulation shows that the bit error rate is improved greatly with the irregular modulation scheme.4 The performance of MIMO system under imperfect channel information is analyzed and the improved scheme is proposed. A robust zero-forcing receiver in Multiple Input Multiple Output (MIMO) channel with channel estimation error is presented. By considering the statistic of channel estimation error, the coefficient of zero-forcing filter is modified with the estimated channel information. The presented receiver is more robust than traditional ZF receiver in the presence of inaccurate channel information in the receiver.5. The performance of iterative receivers based on linear filter and interference cancellation in the present of imperfect channel information are analyzed. To achieve the best performance by iteration, the special mapping is used, such as anti-gray map-ping. It can only take effect when the tunnel of iteration is open, and it will be useless easily when the channel state information is mismatched. By analyzing, iterative sys-tem with gray mapping is more robust under imperfect channel information, and the better performance than anti-gray mapping will be achieved when the channel esti-mation error is large. An adaptive transmission system is presented, whose mapping scheme is switched based on channel estimation quality. It is shown that the output extrinsic information is more reliable by switching the mapping scheme or adopting mixed mapping scheme and the performance of system is improved. It is more robust with the imperfect channel information. Additionally, the improved sic-mmse iterative algorithm is proposed, the more reliable extrinsic information will be output with the modified variance of estimated signal. It is shown that the modified algorithm is better than traditional algorithm.更多还原