【作者】 陈庆春；

【导师】 范平志；

【作者基本信息】 西南交通大学 ， 交通信息工程及控制， 2004， 博士

【摘要】 高数据速率、高可靠性的分组数据传输将成为未来无线通信的主流。在保障移动环境下数据的可靠传输，需要采用有效的差错控制技术，如前向纠错(FEC)编译码、混合自动重发请求(HARQ)等；为了提高系统接收性能，还需要充分利用移动多径环境下分集接收所带来的空间分集增益和基于重传数据码合并处理所带来的时间分集增益等；此外，各类差错控制技术的实施，都离不开可靠的信道估计信息。本文以无线分组数据传输混合差错控制技术为中心，重点研究了码合并时间分集对HARQ性能的影响，等增益合并(EGC)分集接收的系统性能，高速Turbo并行编译码技术，以及基于重复信息的SNR估计算法等。 本文首先讨论了AWGN信道和平坦Rayleigh衰落信道条件下带码合并的Ⅰ类HARQ性能。通过将Pursley在异步CDMA通信系统下的信干比分析方法推广到干扰环境下的HARQ，分析了码合并对合并码字信干比的影响。基于拓展信于比的分析结果表明，异步CDMA通信环境下码合并对HARQ性能的改进源于码合并能增加合并码字符号的有效信干比，且增加的系数等于合并码字数目(码合并的累积效应)。提出了Ⅰ类HARQ的准解析分析方法，在对应前向纠错FEC性能的基础上，该方法提供了一种干扰环境下估计带码合并Ⅰ类HARQ性能和干扰环境下Ⅰ类HARQ系统的设计方法。接着本文通过分析独立同分布Rayleigh衰落变量乘积的不等约束关系与近似关系，提出了一种基于统计特性的近似变量替换分析方法。该方法可将独立同分布Raylei曲衰落变量的乘积项替换为两个具有类似统计特性的衰落变量平方之和，并在近似变量替换分析的基础上，推导出了理想交织平坦Rayleigh衰落信道条件下Ⅰ类HARQ性能的上、下界。研究表明，理想交织平坦Rayleigh衰落信道条件下码合并对HARQ性能的改进源于码合并能有效增加系统的分集阶数，且增加的系数大致等于合并码字数目。 然后，本文将HARQ中利用基于重传数据码合并处理所带来的时间分集性能分析推广到采用速率匹配穿孔卷积(RCPC)编码的Ⅲ类HARQ。研究表明，合并穿孔距阵可以清楚地描述多个RCPC码字合并后的码字特性。对Viterbi的成对错误概率分析进行推广，推导出了AWGN信道条件下的广义成对错误概率的闭式解。在广义成对错误概率的基础上，结合合并穿孔距阵、AWGN信道条件下码合并的累积效应，以及合并RCPC码字特性的变化，推导出了AWGN第n页西南交通大学博士研究生学位论文信道条件下111类HARQ性能的上、下界，给出了HARQ性能的近似估计方法。在推广Viterbi的成对错误概率分析的基础上，推导并得出了平坦Rayteigh衰落信道条件下的广义成对错误概率的闭式解;在广义成对错误概率的基础上，根据合并穿孔距阵，结合理想交织平坦Rayleigh衰落信道条件下分集合并的分集阶数累积效应以及合并RCPC码字特性的变化，推导出了理想交织平坦Rayleigh衰落信道条件下nI类HARQ性能上、下界，给出了衰落信道条件下HARQ性能的近似估计方法。 接着，本文探讨了频率选择性Rayleigh衰落信道条件下等增益合并EGC分集接收的系统性能。基于统计特性的近似变量替换的分析思路，将独立同分布Rayleigh衰落变量的乘积项替换为两个具有类似统计特性的衰落变量平方之和，本文推导得出了频率选择性Ravleigh衰落信道条件下等增益合并EGc分集接收系统性能的下界，提出了EGC分集接收系统性能的近似估计方法。EGC分集接收系统性能的下界和近似估计算法不仅为分析EGC分集接收系统提供一种简捷、有效的方法，一而且通过下界和近似估计还可大致估计出EGC分集接收与最大比合并MRC分集接收间的性能差异。采用近似变量替换的分析思路，不仅EGC分集接收性能分析计算简单，计算结果与理论值或仿真结果十分吻合，而且适用范围广，不受多径数目限制，适用于密集多径/多路独立同分布衰落信道条件下(例如MIMO系统、 OFDM系统)分集接收系统的性能分析。 随后，本文在传统Turb。分块并行译码算法研究的基础上，创造性地提出了一种基于分块归零编码和Turb。分块并行译码结构的高速Tuth。并行编译码算法。该并行编译码算法通过将一个长的信息帧分隔成若干个互不重叠的信息元分块，在逐块编码结束时执行Trellis归零处理，各个分块的编码输出合并为一个完整的Tul-bo码字。由于信息元分块编码后的分段肠ellis归零处理，Turb。码字各个分块的初始和结束状态为已知零状态。在译码端，对应地实施基于分块的Ttlrbo码并行迭代译码处理。仿真研究表明，在大幅度降低译码时延的同时，新的Turb。码编译码结构在高码率时的可靠性能接近乃至优于传统Turb。码译码性能，特别适用于高数据速率、高可靠性要求的高速Tuth。编译码，具有重要的应用价值。 最后，在现有SNR估计算法的基础上，本文提出了一种基于重传数据或重复控制信息(如导频信号)的SNR估计算法。新的估计算法通过充分利用系统所提供的冗余信息，并在冗余信息的基础上构造新的观测序列(累加有扰观测信号序列AONSS或差分有扰观测信号序列DONSS)。在新的观测序列基础上，提出了基于AONss的最大似然估计算法和基于DoNss的盲最大似然估计算法，推导得出了基?更多还原

【Abstract】 High speed packet data transmission with high reliability is the most promising direction for the future wireless communications. In order to gurantee reliable transmission, various error contol schemes including the forward error correction (FEC) and the hybrid ARQ mechanism, are essential. To improve receiver performance, both space diversity in the multipath environment and time diversity, such as code combining in the hybrid ARQ, should normally be employed. Besides, the implementation of a hybrid error correction control scheme is heavily dependent on the proper channel estimation algorithm. Therefore, in this thesis, focusing on various hybrid error control schemes for wireless packet-oriented communication system, the following issues are addressed, hybrid ARQ scheme and its performance enhancement owing to the code combining employed, the performance analysis of equal-gain-combining (EGC) diversity reception, the high-speed parallel Turbo coding scheme, and SNR estimation algorithm based on redundant repetition information.At first, the performance evaluation of hybrid ARQ with code combining in packet-oriented CDMA system is investigated. By extending Pursley’s signal-to-interference plus noise ratio (SINR) analysis to hybrid ARQ with code combining in packet-oriented CDMA system, an extended SINR formula is derived, which describes explicitly the SINR variation of the code symbol involved in code combining. Moreover, it is shown that the effective SINR of the combined symbol is increased by a coefficient which is proportional to the number of repeated replicas involved in the code combining. Based on the extended SINR formula and the resultant SINR variation, a quasi-analytical approximation method is proposed for the performance evaluation of hybrid ARQ with code combining. It is validated that the quasi-analytical approximation offers a simplified routine to’ estimate the performance of hybrid ARQ with code combining, particularly for the applications whose reliability performance of the FEC counterpart could be numerically calculated or evaluated through simulation.Thereafter, the thesis focuses on the performance analysis of hybrid ARQ with diversity combining on the ideally interleaved flat Rayleigh fading channel, which is known as a frustrated problem due to the difficulty in solving the multiple-foldintegration involved. A novel analysis method, i.e., the approximate variable is proposed to simplify the analysis, wherein the involved cross-product of two random variables is substituted with the approximate sum of the power of two random variables. Both the lower and upper bounds are derived, approximate analysis method is also proposed for evaluating the performance of hybrid ARQ with diversity combining on the ideally interleaved flat Rayleigh fading channel. It is disclosed that the diversity combining contributes to an increase in the diversity order on the ideally interleaved flat Rayleigh fading channel, which is approximately proportional to the number of repeated copies involved.Next, the thesis moves its attention to the type III hybrid ARQ with code combining. The code combining in type III hybrid ARQ becomes more complicated due to the variation of the combined codeword. And the combined puncturing pattern is proposed to illuminate the involved composite code combining. By extending the analysis to the convolutional codes whose symbols exhibit different effective signal-to-noise ratio owing to the composite code combining, a closed form of the residual pairwise error probability of the combined RCPC codes is derived for the AWGN channel. Based on the combined puncturing patterns and the accumulate SINR effect, upper and lower bounds on the error performance are derived, together with an approximation method which is shown to be quite satisfactory by numerical and simulation results. By extending the analysis to the convolutional codes whose symbols experience fading path with different fading power gain due to the composite code combining, a closed form of the residual pairwise更多还原