Performance of Self-Corrected Min-Sum Decoding Algorithm for Decoders with Quantized Input

In this paper we explore the Min-Sum decoding algorithm for LDPC codes with self-correction (SC-MS) modification and how it affects the error correction performance for quantized input data. Three levels of quantization are observed: 3-, 2- and 1-bit. It was shown that the self-corrected modification significantly improves the performance of Min-Sum decoding algorithm with quantized input in comparison with basic Min-Sum decoding algorithm. Convergence of the algorithm is also better than for basic Min-Sum algorithm. Results of error correction performance and convergence are presented for two codes: CCSDS (8176, 7156) and IEEE 802.3an (2048, 1723) LDPC. The results are also presented for the SC APP-based and basic Min-Sum decoding algorithms.