面向5G移动通信系统的极化码构造研究

发布时间：2018-07-30 07:28

【摘要】：在信道编码技术领域,追求香农极限容量一直都是通信学者们梦寐以求的目标。早期的Turbo码和Low-density Parity-check(LDPC)码技术已经十分接近这一目标,但是没能得到严格的数学理论证明,目前唯一能够达到香农极限的信道编码技术也只有极化码技术.极化码是一种基于信道极化现象而产生的一种信道专属编码,关于极化码的理论研究可以被划分为极化码的构造方法以及极化码译码算法的研究。研究极化码构造方法的主要目的是在信道产生极化之后精准的挑选出信道容量趋近于“1”的比特信道来传送有用的信息位,而研究极化码译码算法的目的则是在信息接收端以最精确的方式还原发送端所传送过来的信息。本文首先对极化码的理论进行了深入的分析和研究,详细分析和介绍了信道极化现象及性质,对信道合并与拆分过程及原理进行了详细的分解与说明。紧接着对极化码的编码与译码进行的了仔细的分析和研究,目前针对译码算法的研究已经相当的成熟和普遍,最具代表性的三种译码算法分别是Successive cancellation decoder(SCD),SCLD和CRC-SCLD译码算法,而关于极化码构造方法的研究则相对稀少,并且通过改进编码构造方法可以有效的提升极化码的性能。接着对Monte-Carlo方法,Density evolution方法和高斯近似估计方法这三种经典的编码构造方法进行了详细的比较与研究,Monte-Carlo算法的精确度过度依赖重复计算的次数,因此在实际应用时其复杂度十分的高。密度演进算法的精确度虽然非常高,但是该方法在计算的过程中涉及到卷积运算,这对硬件的要求十分的苛刻。高斯近似估计方法存在极化速率过低的问题。最后本文回归到极化码编码构造的本质,通过设计不同的巴哈塔切亚参数来进行极化码的编码构造,然后对这些参数构造的极化码进行性能评估,从中挑选出性能最好的参数,实验结果表明这种通过设计参数的构造方法在性能上十分的接近于高斯近似估计方法,其不存在极化速度慢的问题,而且在所有构造方法中具有最低的计算复杂度。
[Abstract]:In the field of channel coding, the pursuit of Shannon limit capacity has always been the goal of communication scholars. The early Turbo code and Low-density Parity-check (LDPC) code technology have been very close to this goal, but the strict mathematical theory has not proved that the only channel coding technology that can reach Shannon limit is polarization code technology. Polarization code is a kind of channel exclusive coding based on channel polarization phenomenon. The theoretical research on polarization code can be divided into the construction method of polarization code and the study of decoding algorithm of polarization code. The main purpose of the study of polarization code construction is to select exactly the bit channel whose channel capacity approaches "1" to transmit useful information bits after the channel is polarized. The purpose of studying the polarization code decoding algorithm is to restore the information transmitted by the sender in the most accurate way. In this paper, the theory of polarization codes is analyzed and studied in detail, the polarization phenomena and properties of channel are analyzed and introduced in detail, and the process and principle of channel merging and splitting are analyzed and explained in detail. Then, the coding and decoding of polarimetric codes are analyzed and studied carefully. At present, the research on decoding algorithms has been quite mature and universal. The three most representative decoding algorithms are Successive cancellation decoder (SCD) and CRC-SCLD decoding algorithms, respectively. However, the research on the construction method of polarization code is relatively rare, and the performance of polarization code can be improved effectively by improving the coding method. Then, three classical coding methods, Monte-Carlo method, Density evolution method, and Gao Si approximate estimation method, are compared in detail and the accuracy of Monte-Carlo algorithm is studied. Therefore, its complexity is very high in practical application. Although the accuracy of density evolution algorithm is very high, it involves convolution operation in the process of calculation, which is very demanding for hardware. The Gao Si approximate estimation method has the problem of low polarization rate. Finally, this paper returns to the essence of the encoding construction of polarization code, designs different parameters of Bahatatchea to construct the code of polarization code, and then evaluates the performance of the polarimetric code constructed by these parameters, and selects the best parameters. The experimental results show that the proposed method based on the design parameters is very close to the Gao Si approximate estimation method in performance and has the lowest computational complexity among all the construction methods without the problem of slow polarization speed.
【学位授予单位】：深圳大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TN911.22;TN929.5

【相似文献】