An decoding parameter optimization method for LDPC codes in 5G NR based on 3D EXIT chart

Abstract

Low-density parity-check (LDPC) codes are channel code of fifth-generation(5G) New Radio(NR), and it is important to explore low-complexity decoding algorithms that are easy to implement on hardware. In order to realize 5G ultra-reliable low latency communication (URLLC), low-complexity decoding of short packets is necessary. For quasi-cyclic LDPC (QC-LDPC) codes in 5G NR, the layer offset min-sum (LOMS) algorithm and the layer normalized min-sum (LNMS) algorithm are two efficient algorithms with low complexity. The performance of these two simplified decoding algorithms is significantly affected by the values of the above two decoding parameters. In this paper, the 3D extrinsic information transfer (EXIT) chart is introduced as an analysis tool to indicate the performance of different values for the offset and normalization parameters. The simulation results show that with different values of the above two factors, the changing of the output information of EXIT chart is consistent with that of the block error rate (BLER) in decoding simulations. In addition, it is shown that this analysis method can be extended to the decoding with fixed-point inputs.