论文标题:（2，1，4）卷积码的逻辑代数译码方法研究
The Research of the Logic Algebra Decoding Method of (2, 1, 4) Convolutional Code
论文作者
论文导师 杨万全,论文学位 硕士,论文专业 通信与信息系统
论文单位 四川大学,点击次数 158,论文页数 82页File Size2359K
2006-04-25论文网 http://www.lw23.com/lunwen_1237502/
Convolutional code; Logic Algebra Decoding; Viterbi algorithm; Error pattern; The syndrome
随着数字通信、数据处理和计算机通信网的飞速发展，用户对信息传输的可靠性和有效性，提出了更高的要求。卷积码作为一种重要的信道纠错码，由于性能优异，在移动通信、卫星通信和空间通信等领域发挥着重要作用。卷积码的传统译码方法有两大类：一类是门限译码，另一类是概率译码，概率译码又分为序列译码和维特比译码。寻求新的算法简单而性能良好的卷积码译码方法仍然具有实用意义。 本文利用逻辑代数中模2运算的可逆性，构造出了一种卷积码新型译码方法——逻辑代数译码，并以(2，1，4)卷积码为基础，对译码方法进行了深入地研究与探讨。文章首先分析了卷积码逻辑代数译码的基本原理，接着讨论了(2，1，4)卷积码的各种误码类型，确定以一个约束长度内10位码元错1位和错2位为研究对象，共有19种误码类型。之后研究了误码判定的规则，证明了伴随式只与错误图样有关，而与编码器输入的信息序列无关，通过分析错误图样，得出了19种误码类型所对应的伴随式。在19个伴随式中，8个伴随式与误码类型之间存在模糊现象，在增加观测时刻的条件下，模糊现象得到解决。这样，19个伴随式与误码类型之间确立了一一对应关系，从而也就得到了(2，1，4)卷积码的逻辑代数译码规则。 通过系统仿真，本文设计的译码器的性能虽然略差于Viterbi译码，但它具有分析方法和译码算法简单易懂、译码速度快和运算量小、不需要寄存大量数据、实现简单等诸多优点。若在此基础上对其进行进一步改进和完善，该译码方法不失为一种性能优良、具有很好应用前景的卷积码译码方法，文章最后对卷积码逻辑代数译码思想的推广应用也进行了探讨。
With the rapid development of digital communication, data processing and computer communication networks, the more reliability and efficiency of information transmission is demanded. As an important channel error-correcting code, convolutional code has excellent performance, and plays an important role in many fields , such as mobile communication , satellite communication and space communication. The traditional decoding methods of convolutional code can be classified into two types: threshold decoding and probability decoding. The latter can be divided into sequential decoding and Viterbi algorithm. Searching a new decoding method with simple algorithm and good performance for convolutional code has important practicality.In this paper, a new decoding method of convolutional codes named Logic Algebra Decoding is constructed using the reversibility of Mode 2 operation in Logic Algebra, and the decoding method of (2, 1, 4) convolutional code is deep researched and discussed. The fundamental of Logic Algebra Decoding is analyzed at first. Then the bit error types of (2, 1, 4) convolutional code are discussed, and 19 bit error types in a constraint length including 10 bits are found, if one or two bit errors are considered. Furthermore, the bit error decision rules are researched, and it is proved that the syndrome is a function only of the error pattern and is independent of the information sequence from encoder. The corresponding syndrome for each of the 19 bit error types is obtained through the error patterns analysis. Among the 19 syndromes, 8 syndromes and their bit error types exist ambiguity which can be solved by increasing the observational time. Thus, the one-to-one relationship between the 19 syndromes and the bit error types is established, and the Logic Algebra Decoding algorithm of (2, 1,4) convolutional
|
