Performance Analysis for Finite Length LT Codes via Classical Probability Evaluation
Longlong Suo, Gengxin Zhang, Jing Lv, Xiang Tian · IEEE Communications Letters · 2017
Luby Transform (LT) codes play an important role in both data transmission and storage, but it is still a challenging problem that how to analyze their performance due to the degree distributions. In this letter, a novel methodology to analyze the performance of LT codes is proposed. This method establishes a probability transfer model between two adjacent steps of the belief propagation decoding algorithm, and provides an intuitive and unifying tool to analyze finite length LT codes. We first present the comprehensive description and theoretical identification of the proposed methodology, and then give some simulation results to validate its effectiveness.