基于解调软判决的信道编码参数识别技术研究

发布时间：2019-05-25 03:22

【摘要】：在编码参数未知的情况下,接收方常常利用信道编码分析技术对接收到的编码序列进行处理,根据识别出的编码参数对接收序列进行译码,最终完成对信息序列的提取。该技术在信号截获、智能通信等领域具有重要的应用价值。传统编码分析方法的研究对象通常是解调硬判决序列,硬判决序列所能提供的信息量十分有限,随着新型通信技术和编码体制的广泛应用,这些方法容错性差、适用面窄的缺点日益凸显。同硬判决序列相比,软判决序列往往可以为接收方提供更加丰富的先验信息,因此,为克服硬判决方法所遇到的困难,本文研究解调软判决条件下的编码分析技术。通过求解由接收序列构造的含错方程组来寻找编码约束关系,是一种常用的编码分析思路。基于解调硬判决的Walsh-Hadamard变换法在求解含错方程组时存在容错性能不足的问题。为此,本文提出将Chase2译码算法中构造候选码字集的思想与Walsh-Hadamard变换相结合,通过对系数向量依概率加权,定义了软判决条件下Walsh-Hadamard变换谱系数的含义;同时,推导了适用于分组码监督向量的判决门限。仿真结果表明,在不明显增加运算量的前提下,本文算法的容错性能优于传统硬判决算法,且信噪比越低,优势越明显。已有方法在处理中短码长LDPC码开集识别问题时,通常采用矩阵初等变换的方法来搜索监督向量,在有误码的条件下,复杂的矩阵变换过程将引起严重的错误传播问题,进而导致算法失效。为解决这一问题,本文利用软判决序列所提供的概率信息,给出了一种基于可靠监督关系的预判决方法,并对相关参数的选取进行了推导,使得向量搜索空间在构造过程中具有更强的容错能力;在此基础之上,利用广义对数似然比检验对搜索空间中的向量进行判决,得到监督向量;最后,提出一种基于多组数据和软输入软输出译码的迭代识别方案。仿真结果表明,低信噪比条件下,本文算法可以有效应对中短码长LDPC码的开集识别问题。现有软判决算法在对高码率LDPC码进行闭集识别时容错性能有限。为此,本文首先对不同码率下监督关系对数似然比的分布特征进行了分析,指出了已有算法性能不足的原因;在此分析之上,提出了一种可以全面体现随机变量分布特征的新判决参数——偏差比,进而给出一种利用最大偏差比判决器的LDPC码识别算法。仿真结果表明,本文算法可以有效地完成低信噪比环境下的LDPC码闭集识别问题,特别地,对于高码率LDPC码,新算法具有较为明显的优势。
[Abstract]:In the case of unknown coding parameters, the receiver often uses channel coding analysis technology to process the received coding sequence, decode the received sequence according to the identified coding parameters, and finally complete the extraction of the information sequence. This technology has important application value in signal interception, intelligent communication and other fields. The research object of traditional coding analysis method is usually demodulation hard decision sequence, and the information provided by hard decision sequence is very limited. With the wide application of new communication technology and coding system, these methods have poor fault tolerance. The shortcomings of narrow applicable surface are becoming more and more prominent. Compared with the hard decision sequence, the soft decision sequence can often provide the receiver with more prior information. Therefore, in order to overcome the difficulties encountered by the hard decision method, this paper studies the coding analysis technology under the condition of demodulation soft decision. It is a common coding analysis idea to find the coding constraint relation by solving the error-containing equations constructed by the receiving sequence. The Walsh-Hadamard transform method based on demodulation hard decision has the problem of insufficient fault-tolerant performance in solving error-containing equations. In this paper, the idea of constructing candidate code word set in Chase2 decoding algorithm is combined with Walsh-Hadamard transform, and the meaning of Walsh-Hadamard transform spectral coefficient under soft decision condition is defined by weighted coefficient vector according to probability. At the same time, the decision threshold suitable for block code supervision vector is derived. The simulation results show that the fault-tolerant performance of the proposed algorithm is better than that of the traditional hard decision algorithm without obviously increasing the amount of computation, and the lower the signal-to-noise ratio (SNR), the more obvious the advantage. When dealing with the problem of open set recognition of LDPC codes with medium and short code length, the existing methods usually use the method of matrix elementary transformation to search the supervised vector. Under the condition of error code, the complex matrix transformation process will cause serious error propagation problem. This leads to the failure of the algorithm. In order to solve this problem, this paper presents a pre-decision method based on reliable supervision relationship by using the probability information provided by soft decision sequence, and deduces the selection of related parameters. The vector search space has stronger fault-tolerant ability in the construction process. On this basis, the generalized logarithmic likelihood ratio test is used to determine the vectors in the search space, and the supervised vectors are obtained. finally, an iterative recognition scheme based on multiple sets of data and soft input and soft output decoding is proposed. The simulation results show that the proposed algorithm can effectively solve the problem of open set recognition of LDPC codes with medium and short code length under the condition of low SNR. The fault-tolerant performance of the existing soft decision algorithms for closed set recognition of high bit rate LDPC codes is limited. Therefore, this paper first analyzes the distribution characteristics of logarithmic likelihood ratio of supervisory relationship at different bit rates, and points out the reasons for the lack of performance of the existing algorithms. On the basis of this analysis, a new decision parameter, deviation ratio, which can fully reflect the distribution characteristics of random variables, is proposed, and then a LDPC code recognition algorithm based on the maximum deviation ratio discriminator is proposed. The simulation results show that the proposed algorithm can effectively complete the problem of LDPC code closed set recognition in low SNR environment, especially for high bit rate LDPC codes, the new algorithm has obvious advantages.
【学位授予单位】：解放军信息工程大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TN911.22

【参考文献】