论文部分内容阅读
LT码是一类前向纠错信道编码,用于纠正信道分组删除(Packet erasure)。这类编码具有广泛的用途,包括计算机科学、网络传输、媒体存储、大文件下载等等。在LT码的设计中,度分布是成功解码和快速运算的关键。这篇文章展示了一种新的LT码设计,它将弱分布用于前期解码,再将增强型分布用于后期解码。由于弱分布具有低的平均度数,它可以显著地增加编码及解码的速度。同时,增强型度分布具有高的平均度数,能够提高成功解码的概率。通过一系列的仿真,笔者观察到这种设计的编码冗余度和编解码所需的异或运算量比使用Robust soliton分布的参考方案降低大约50%。
LT codes are a type of forward error correction channel coding used to correct for packet erasure. This type of coding has a wide range of uses, including computer science, network transmission, media storage, large file downloads, and more. In the LT code design, the degree distribution is the key to successful decoding and fast computing. This article shows a new LT code design that uses weak distributions for early decoding and then enhanced distribution for later decoding. Because of the low average degree of weak distribution, it can significantly increase the speed of encoding and decoding. At the same time, the enhanced degree distribution has a high average degree, which increases the probability of successful decoding. Through a series of simulations, the authors observed that the XOR of this design’s coding redundancy and codec was reduced by about 50% over the reference scheme using the Robust soliton distribution.