Low Complexity High Throughput Coding Scheme for 6G Immersive Communication
Xia An, Chao Zhang, Kewu Peng, Zhitong He, Jian Song · 2024
The immersive communication scenario in 6G, stemming from the enhanced mobile broadband (eMBB) scenario in 5G, demands low-complexity high-throughput channel coding scheme due to its huge data traffic. The current data channel coding scheme of eMBB, i.e., 5G new radio (5G-NR) low-density parity-check (LDPC) codes, shows limitation on further improvements of high throughput due to their highly irregular base graph structure. In this paper, we first introduce a rate compatible quasi-cyclic (QC-) spatially-coupled (SC-) LDPC code structure, and then apply an adaptive quantized and normalized min-sum algorithm (AQNMSA) with very low message passing bit-width of 4-bit to the QC-SC-LDPC codes. The asymptotic decoding thresholds provided by multi-edge-type density evolution tool and the simulation results both show that QC-SC-LDPC codes with 4-bit AQNMSA have better decoding performance than conventional non-adaptive normalized min-sum algorithm as well as approaching decoding performance compared to 5G-NR LDPC codes with optimal belief propagation decoding. In general, QC-SC-LDPC codes with AQNMSA, row-parallel architecture, and layered-scheduling sliding window decoding could achieve superior decoding performance than current coding scheme, including at least 30 times of throughput gain and better decoding performance with lower decoding complexity.