Summary
International Symposium on Nonlinear Theory and Its Applications
2016
Session Number:B2L-E
Session:
Number:B2L-E-2
Robust Scale-Free Luby Transform Code and its Performance
Yuli Zhao, Francis C.M. Lau, Zhiliang Zhu, Hai Yu, Li Guo,
pp.-
Publication Date:2016/11/27
Online ISSN:2188-5079
Summary:
Compared with Luby Transform (LT) codes having an ideal/robust soliton degree distribution, LT codes with encoded-symbol degree following a modified power-law distribution (scale-free LT codes) have been shown to possess a higher probability of successful decoding and a lower encoding/decoding complexity when the information symbol length ranges from 512 to 2048. In an attempt to reduce the size of the initial ripple set of scale-free LT (SF-LT) codes so as to prevent the ripple set from becoming empty, a new class of LT codes, namely robust SF-LT (RSF-LT) codes, is proposed in this work. The performance and characteristics of the proposed RSF-LT code are compared with those of other LT codes including robust LT code, SF-LT code, and ``LT code with decreasing ripple set''. Results show that the proposed RSF-LT code outperforms the other codes with respect to average overhead, encoding/decoding efficiency and probability of successful decoding.