LDPC Decoding Over Nonbinary Queue-Based Burst Noise Channels

摘要: Iterative decoding based on the sum-product algorithm (SPA) is examined for sending low-density parity check (LDPC) codes over a discrete nonbinary queue-based Markovian burst noise channel. This channel model adopted due to its analytical tractability and recently demonstrated capability in accurately representing correlated flat Rayleigh fading channels under antipodal signaling either hard or soft output quantization. SPA equations are derived closed form this terms of parameters. It then numerically observed that potentially large coding gains can be realized with respect Shannon limit by exploiting memory as opposed ignoring it via interleaving. Finally, LDPC performance both matched mismatched regimes evaluated. shown provides noticeable interleaving effectively capture underlying behavior when codes.