Optimization of Bit Mapping and Quantized Decoding for Off-the-Shelf Protograph LDPC Codes with Application to IEEE 802.3ca

TL;DR

This paper enhances the performance of off-the-shelf protograph LDPC codes for optical networks by optimizing bit mapping and quantized decoding, achieving notable gains in error rate performance.

Contribution

It introduces an optimized bit mapping and quantized decoding approach specifically tailored for standard LDPC codes used in IEEE 802.3ca, improving their efficiency.

Findings

Achieved 0.4 dB and 0.3 dB SNR gains at BER 1e-6 for shaped and uniform signaling.

Optimized the clipping value for quantized sum-product decoders using discretized density evolution.

Demonstrated improved performance of standard LDPC codes in optical network applications.

Abstract

Protograph-based, off-the-shelf low-density parity-check (LDPC) codes are optimized for higher-order modulation and quantized sum-product decoders. As an example, for the recently proposed LDPC code from the upcoming IEEE 802.3ca standard for passive optical networks (PONs), an optimized mapping of the bit channels originating from bit-metric decoding to the protograph variable nodes gains 0.4 dB and 0.3 dB at a bit-error rate of 1e-6 for shaped and uniform signaling, respectively. Furthermore, the clipping value for a quantized sum-product LDPC decoder is optimized via discretized density evolution.