Design of Rate-Compatible QC-IRA-d Codes Based on Gaussian Approximation Analysis and Graph Extension

In this paper, we propose a novel extending algorithm to design rate compatible (RC) QC-IRA-d codes through Tanner Graph extension. The structure of QC-IRA-d codes provides high efficiency and low complexity for graph extension. Thus, in the proposed method, there are three features: (1) close relation between the extended parts and original parity check matrix; (2) larger length of introduced cycles; (3) balanced check node degree. All of them guarantee good performance of the designed RC family codes. Simulation results show that the graph extension method outperforms existed algorithms from 0.05 to 0.47 dB for different rate codes.