Decoding Wave Velocity Analysis for SC-LDPC Ensembles on BMS Channels using Interpolation

This paper is concerned with decoding waves for spatially-coupled low-density parity-check (SC-LDPC) ensembles on binary memoryless symmetric (BMS) channels. We propose an approach termed as interpolated density evolution (IDE) to predict the velocities of decoding waves. In this approach, we approximate the densities of decoding waves via interpolation and establish two transfer functions for the update of interpolation functions. Comparison results indicate that the proposed IDE approach achieves a good trade-off between computational cost and accuracy as compared with the standard coupled density evolution (CDE) approach and the extrinsic information transfer (EXIT) approach based on Gaussian approximation (GA). This highlights the attractiveness of the proposed approach in the analysis of SC-LDPC ensembles on general BMS channels.