线性分组码参数的盲识别算法研究

发布时间：2018-03-14 10:07

本文选题：信道编码技术　切入点：线性分组码　出处：《河北大学》2017年硕士论文　论文类型：学位论文

【摘要】：在数字通信系统中,为了抗击信息在传输过程中受到的影响和干扰,提高信息传输的可靠性,信道编码技术得到迅速发展。由于信道编码技术应用范围日益增大,信道编码盲识别技术应运而生。其目的是在仅有少量甚至没有任何先验信息的前提下识别出截获码流的编码体制和相关参数,从而恢复出原始信息序列。它在信息对抗、通信侦测以及智能通信等领域具有重要的研究价值。本文重点研究了二进制线性分组码参数的全盲识别问题,论文的主要工作如下:(1)介绍了线性分组码的理论知识、盲识别的数学模型和识别参数,为后面章节线性分组码的全盲识别算法研究奠定了基础,并指出线性分组码是本文的研究对象。(2)针对线性分组码参数盲识别容错性能差、复杂度高和半盲识别的问题,提出了一种基于特征融合的线性分组码全盲识别算法。首先根据实际序列与随机序列码重分布概率间较大的差异性,研究了一种运用码重标准差率差值、码重信息熵分别同时识别码长和起始点的算法;然后比较这两种特征参数识别效果又进一步改进,提出了一种新的融合特征参数来同时识别码长和起始点的算法;最后通过建立矩阵进行模二化简求解生成矩阵,完成线性分组码参数的全盲识别。通过理论分析和仿真验证,该算法简单易行且复杂度低,在误码率为0.025的条件下对中短码的全盲识别率高达90%,误码率为0.005的条件下对中长码的全盲识别率达到80%。(3)针对高误码率条件下高码率循环码参数的全盲识别问题,提出了一种基于最大公因式阶数相异度的循环码全盲识别算法。首先根据循环移位前后码字的最大公约式阶数,利用实际序列与随机序列阶数分布概率间的差异性,提出了一种基于数据挖掘中相异性度量函数同时识别起始点和码长的算法;然后根据循环码特性,计算阶数分布差值来识别生成多项式,实现了循环码参数的全盲识别。该算法简单易行且容错性较强,在误码率为0.013的条件下对中长码的全盲识别效果较好。
[Abstract]:In digital communication system, in order to resist the influence and interference of information transmission and improve the reliability of information transmission, channel coding technology has been developed rapidly. The blind identification technique of channel coding arises at the historic moment. Its purpose is to recognize the coding system and related parameters of the intercepted bitstream without even any prior information, so as to recover the original information sequence. Communication detection and intelligent communication have important research value. This paper focuses on the whole blind identification of binary linear block code parameters. The main work of this paper is as follows: 1) the theoretical knowledge of linear block code is introduced. The mathematical model and identification parameters of blind block code are the foundation of the whole blind recognition algorithm of linear block code in the following chapters. It is pointed out that linear block code is the research object of this paper. (2) for the parameter blind recognition of linear block code, the fault-tolerant performance of linear block code is poor. In order to solve the problem of high complexity and semi-blind recognition, this paper proposes a full blind recognition algorithm for linear block codes based on feature fusion. Firstly, according to the large differences between the distribution probability of code weight distribution between real sequences and random sequences, In this paper, an algorithm using the difference of code weight standard deviation rate and the code weight information entropy to identify the code length and the starting point respectively is studied, and the recognition effect of these two characteristic parameters is further improved. This paper presents a new algorithm to identify the length of code and the starting point at the same time by combining the characteristic parameters. Finally, by establishing the matrix to solve the generation matrix by module reduction, the full blind identification of the parameters of the linear block codes is accomplished, which is verified by theoretical analysis and simulation. The algorithm is simple and easy to implement and has low complexity. Under the condition that the BER is 0.025, the full blind recognition rate of medium and short codes is as high as 90 and that of medium and long codes is up to 80 and 80 under the condition of BER 0.005.) for the problem of full blind recognition of cyclic code parameters at high bit error rate, In this paper, a full blind recognition algorithm for cyclic codes based on the order dissimilarity of the largest common factor is proposed. Firstly, according to the maximum convention order of the code word before and after cyclic shift, the difference between the distribution probability of the order of the real sequence and the random sequence is used. In this paper, an algorithm based on the heterogeneity measure function in data mining is proposed to identify both the starting point and code length, and then, according to the characteristics of cyclic code, the order distribution difference is calculated to identify the generating polynomial. The algorithm is simple and fault-tolerant, and has a good effect on the full blind recognition of medium and long code under the condition of error rate of 0.013.
【学位授予单位】：河北大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN911.22

【相似文献】