報告題目：How to Decode High-Density Parity Check Code?

報告時間：2023年7月18日上午10:00-11:00

報告地點：kaiyun開云官方網(wǎng)站犀浦校區(qū)X7510

報告人：劉子龍

摘要： ABP decoding is a soft-decoding approach for high-density parity check codes such as BCH codes and Reed-Solomon (RS) codes. Nowadays, these codes play a very important role for supporting ultra-reliable low-latency communications (URLLC) in 5G/6G systems due to their short block lengths and low/no error floors.

     The main idea of conventional ABP decoder is to adaptively sparsify columns of parity-check matrix corresponding to the (n-k) least-reliable bits using Gaussian elimination (GE) followed by BP decoding in each iteration, where n and k denotes codeword length and message length, respectively.

      In this talk, I will introduce an improved ABP decoding method in which a small number of columns (say, p) corresponding to randomly selected reliable bits are selected in GE, in addition to columns associated to n-k-p least-reliable bits. Mutual information analysis shows that this new decoding strategy is able to extract larger amount of extrinsic information and therefore leads to significantly improved decoding performance. As an application of this improved ABP decoding, I will present a series of numerical results for a BCH-coded continuous phase modulation (CPM) turbo receiver in which improved ABP brings in fast convergence and lower error floor in turbo decoding.

報告人簡介：劉子龍，新加坡南洋理工大學(xué)博士，英國埃塞克斯大學(xué)教授。從2018年1月到2019年11月，他是薩里大學(xué)5G創(chuàng)新中心（5GIC）所在地通信系統(tǒng)研究所（ICS）的高級研究員。目前他是 IEEE高級會員，同時擔(dān)任國際知名學(xué)術(shù)期刊 IEEE Transactions on Vehicular Technology, IEEE Transactions on Neural Networks and Learning Systems, IEEE Wireless Communications Letters, IEEE Access的副編輯。目前已發(fā)表95篇期刊論文，其中50余篇發(fā)表在TIT, TSP, TCOM, TWC, TVT, JSAC, JSTSP等期刊上。