Density Evolution and Thresholds for Accumulate Repeat Tree Codes in Mobile Communication Systems
Maofan Yang, Hua Ren Zhou, Xin Zhang, Dacheng Yang · 2009
Channel coding theories are widely used in computer science and communication systems. This paper proposes a novel channel coding scheme called accumulate repeat tree (ART) codes for improving the channel coding performance in mobile communication systems. This class of codes can be viewed as turbo-like codes combining many advantages of turbo codes and low-density parity-check (LDPC) codes. ART codes can be represented by the Bayesian network and Tanner graph, which allows for high-speed iterative decoding implementation using belief-propagation (BP) algorithm. The density evolution method is presented, and the practicable Gaussian approximation algorithm for ART codes to analyze the thresholds and decoding performance is discussed. ART codes have low coding complexity, and they show good performance improvement by simulation.