喷泉码中不定帧长LT码及其生成矩阵的优化  

Variable frame length LT code in fountain code and its generation matrix optimization

作　　者：赵旦峰[1] 钱晋希 吴宇平[1] 

机构地区：[1]哈尔滨工程大学信息与通信工程学院,黑龙江哈尔滨150001

出　　处：《哈尔滨工程大学学报》2011年第7期953-957,共5页Journal of Harbin Engineering University

基　　金：黑龙江省自然科学基金资助项目(F200810)

摘　　要：针对喷泉码中LT(Luby transform)码编码效率不高、需传送的数据量大、时延长且译码概率低等弊端,将数学极限分析方法引入LT码的度数分布中,提出了一种不定帧长LT码方案,并且在不定帧长LT码的生成矩阵中引入改进的熵编码压缩算法,得到了以熵编码为基础的不定帧长LT码系统模型.其核心思想是先将原始数据进行适当的帧长设计,通过不定帧长的方法对原始数据进行拆分编译码,以满足不同的性能要求,然后对生成矩阵按列进行熵编码优化,使得在不破坏喷泉码特性的前提下大大减少所需传送的数据量.理论分析和仿真结果表明:与已提出的LT码相比,该系统能够有效地提高其编码效率和译码概率,降低LT码的编译码数据量和运算量.Because Luby transform（LT） code in fountain code has the drawbacks of low coding efficiency,a large amount of transfer data,long decoding time,and low decoding probability,a mathematical method of limiting analysis was introduced to the degree distribution of LT code.Moreover,a system program of variable frame length LT（VFLLT） code was proposed,and an improved entropy coding algorithm was also introduced to the generation matrix of VFLLT,deriving a VFLLT system model based on the entropy coding.Its core idea is first to make an appropriate design of the frame length of the source data,split the source data,and encode the data pieces separately to meet different performance requirements.Next,the goal is to optimize the generation matrix by columns through entropy coding in order to decrease the amount of data without breaking the characteristics of the fountain code.Theoretical analysis and simulation results show that compared with the existing LT codes,the proposed system can effectively improve the coding efficiency and decoding probability and reduce the amount of encoding and decoding as well as the number of operations.

关 键 词：喷泉码 LT码 不定帧长 极限分析方法 熵编码压缩 

分 类 号：TN911.22[电子电信—通信与信息系统]