| 摘要: |
| 为提升基于原模图构造的空间耦合低密度奇偶校验(Spatially-Coupled Low-Density Parity-Check，SC-LDPC)码滑窗译码(Sliding Window Decoding，SWD)算法译码性能，提出了残差滑窗译码(Residual SWD，RSWD)算法，通过动态选择可靠度最低（残差值最大）的边信息优先传输，降低边信息无效更新次数，提高了译码性能。RSWD译码在窗口内易出现贪婪组和静默节点现象，导致译码误码率(Bit Error Rate，BER)恶化。为改善这一问题，提出了基于节点的残差滑窗译码(Node-wise RSWD，NW-RSWD)算法和消除静默节点残差滑窗译码(Eliminating Silent Node RSWD，ESN-RSWD)算法。NW-RSWD算法在译码过程中以变量节点为单位，动态更新窗口内最大残差所在边的变量节点。ESN-RSWD算法在译码过程中根据残差值大小，遍历更新窗口内每一个变量节点，使更多的信息参与到窗口译码，避免滑窗译码陷入局部区域更新。仿真结果表明，信噪比处于3~3.5 dB区间时，相较于SWD算法，NW-RSWD算法复杂度增加约15%，ESN-RSWD算法复杂度增加约25%。在窗口大小为8时，为了达到10-6误码率，相较于SWD算法，NW-RSWD算法提升约0.7 dB性能，ESN-RSWD算法提升约0.85 dB。在10-3误码率时，SWD算法、NW-RSWD算法和ESN-RSWD算法分别需要约50次、10次和8次迭代才能达到相同的误码性能。所提算法以增加较少计算复杂度为代价，降低了译码误码率，减少了译码平均迭代次数。 |
| 关键词: 空间耦合低密度奇偶校验码 滑窗译码 信息传递 节点残差算法 消除静默节点残差算法 |
| DOI：10.20079/j.issn.1001-893x.240225001 |
|
| 基金项目: |
|
| Node-wise Dynamic Sliding Window Decoding of SC-LDPC Codes |
| ZHOU Hua,b,XU Chenchen |
| (a.School of Electronic and Information Engineering;b.Jiangsu Collaborative Innovation Center of Atmospheric Environment and Equipment Technology,Nanjing University of Information Science and Technology,Nanjing 210044,China) |
| Abstract: |
| To improve the decoding performance of sliding window decoding(SWD) algorithm for spatially-coupled low-density parity-check(SC-LDPC) codes，the residual sliding window decoding(RSWD) is proposed,in which the edge information of the lowest reliability(or equivalently the largest residual) is dynamically updated as a first priority.This approach effectively reduces the probability of invalid edge updating and leads to noticeable improvements in the decoding performance.However,RSWD is likely to be affected by the phenomenon of greedy and silent nodes,resulting in high bit error rate(BER).To address this issue,two algorithms are proposed,i.e.,the node-wise residual sliding window decoding(NW-RSWD) and the eliminating silent node residual sliding window decoding(ESN-RSWD).The NW-RSWD algorithm dynamically updates the variable node where the edge with the maximum residual is involved,while the ESN-RSWD algorithm traverses and updates each variable node within the window based on the node residual.These methods significantly increase the amount of edge information participating in the iterative decoding process and prevent the decoding from being trapped in a local area of the window.Simulation results show that compared with the SWD algorithm,when the signal-to-noise ratio(SNR) is in 3~3.5 dB,the complexity of the NW-RSWD algorithm increases by about 15% and the complexity of the ESN-RSWD algorithm increases by about 25%.When the window size is 8,in order to achieve bit error rate(BER)=10-6,the NW-RSWD algorithm is increased by about 0.7 dB and the ESN-RSWD algorithm by about 0.85 dB compared with the SWD algorithm.When BER=10-3,the SWD algorithm,the NW-RSWD algorithm and the ESN-RSWD algorithm need respectively about 50,10 and 8 iterations to achieve the same bit error performance.The proposed algorithms improve the decoding performance and reduce the average number of decoding iterations with acceptable increase in computational complexity． |
| Key words: spatially-coupled low-density parity-check(SC-LDPC) code sliding window decoding message passing node-wise residual sliding window decoding(NW-RSWD) eliminating silent node residual sliding window decoding(ESN-RSWD) |
《电讯技术》编辑部 