复杂线性分组码参数盲估计

发布时间：2018-08-20 12:53

【摘要】：信道编码参数盲估计一直是非合作通信领域的热门话题,在通信对抗、电子侦察和信息获取等军事领域有着广泛的应用。在非合作的电子对抗领域,从对截获信号研究转入信息底层解码的第一步是从获取到的信息码流中成功提取有效信息,并正确识别编码类型和估计编码参数。线性分组码是信道编码中非常重要的一类。本文重点研究复杂线性分组码,围绕复杂线性分组码的参数盲估计方法开展工作,以下是主要内容:1.介绍了信道编码中需要用到的抽象代数基础知识和重要性质,为后续线性分组码参数盲估计算法的探讨与实现做了铺垫。2.完成了BCH-乘积码参数盲估计算法。利用分组码元之间的相关性,通过求数据矩阵归一化秩完成对码组偏差和分组码长的估计。基于译码匹配的估计方法,通过统计码字正确率,完成对生成多项式中含本原多项式部分的估计,再估计出最小多项式中的非本原多项式,两者乘积即为估计的生成多项式。3.提出了基于含本原元因式的RS码参数盲估计算法。通过统计RS码中含有因式(x(10)?)的码字个数,完成对分组码长和本原多项式。利用RS码生成多项式是由含本原元的连续幂根构成这一特性,筛除错误码字后,通过验证连续幂根可完成对生成多项式的估计。此算法复杂度较低,只需要码元序列中有少量正确码字就能完成估计,因此,即使在码长较长的情况下,本文算法也具有较强的抗误码性,仿真表明,在误码率为0.002时,识别概率仍达到90%。4.提出了抽取RS码部分码字进行参数盲估计的算法。该算法核心思想是利用RS码等效准循环码的性质,因为抽取的部分码字间存在关联性,通过对不同位数的码元抽取,求得其空间维数,以此完成对RS码的码长估计。该算法的优点在于只需要用到部分码字就能完成参数估计,节省了算法运行的空间复杂度,与使用完整码字在伽罗华扩域求归一化秩相比,运行速度快了一倍以上。5.提出了删除RS码的参数盲估计算法。在传统欧几里得算法基础上进行改进,利用两个相异删除RS码码字求最大公因式仍然等于原生成多项式这一特性,假设不同参数下相异符号码字之间求得最大公因式是否满足条件,完成参数的盲估计。此算法抗误码性较好,在码长较长且误码率低于0.001的情况下,仍能达到90%以上的识别概率。
[Abstract]:Blind estimation of channel coding parameters has been a hot topic in the field of non-cooperative communication, and has been widely used in military fields such as communication countermeasures, electronic reconnaissance and information acquisition. In the field of non-cooperative electronic countermeasures, the first step from the research of intercepting signals to the low-level decoding of information is to extract the effective information from the obtained information stream successfully, and to correctly identify the encoding types and estimate the coding parameters. Linear block code is a very important class of channel coding. This paper focuses on the study of complex linear block codes, focusing on the blind parameter estimation method of complex linear block codes. The following is the main content: 1. This paper introduces the basic knowledge and important properties of abstract algebra used in channel coding, and paves the way for the discussion and implementation of blind parameter estimation algorithm of linear block codes. BCH-product code parameter blind estimation algorithm is completed. Based on the correlation between block symbols, the normalized rank of the data matrix is used to estimate the block deviation and block length. Based on the estimation method of decoding matching, the estimation of the part of the generating polynomial containing the primitive polynomial is completed by statistical codeword accuracy, and the non-primitive polynomial in the minimum polynomial is estimated. The product of the two polynomials is the estimated generating polynomial. 3. A blind estimation algorithm for RS code parameters based on primitive elements is proposed. By counting the factor (x (10 in the RS code?) The number of codewords, complete block code length and primitive polynomial. The RS code generating polynomial is composed of a continuous power root with primitive elements. After the error codeword is eliminated, the estimation of the generating polynomial can be completed by verifying the continuous power root. The complexity of this algorithm is low, only a small number of correct codewords in the code sequence are needed to complete the estimation. Therefore, even if the code length is longer, the algorithm also has a strong error resistance. The simulation results show that the error rate is 0.002. The probability of recognition is still 90. 4. A blind parameter estimation algorithm for extracting partial codewords from RS codes is proposed. The key idea of the algorithm is to use the property of the RS code equivalent quasi-cyclic code, because there is correlation between the extracted part of the code word, by extracting different bits of symbols, the spatial dimension is obtained, and the code length estimation of the RS code is completed. The advantage of the algorithm is that only some codewords are needed to complete the parameter estimation, and the space complexity of the algorithm is saved. Compared with solving the normalized rank of the complete codeword in the Galois domain, the speed of the algorithm is more than twice as fast as .5. A blind parameter estimation algorithm for deleting RS codes is proposed. On the basis of the traditional Euclidean algorithm, using two different deleting RS code words to find the largest common factor is still equal to the original generating polynomial. Assuming that the maximum common factor can satisfy the condition between the words of different symbol numbers under different parameters, the blind estimation of the parameters is completed. The algorithm has good error tolerance, and the recognition probability is more than 90% when the code length is longer and the bit error rate is less than 0.001.
【学位授予单位】：电子科技大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN911.22

【参考文献】