论文部分内容阅读
针对目前现有的信噪比(SNR)估计方法不适用光通信中乘性噪声信道的问题,分析了乘性噪声信道下,SNR失配对低密度奇偶校验(LDPC)码性能的影响,并通过仿真表明乘性噪声信道下SNR精确估计的必要性。提出了一种基于量化的LDPC码判决反馈的SNR估计方法,首先对接收到的光信号进行量化,利用简化后的期望最大(EM)算法对量化后的SNR进行一次粗估计,接着利用基于LDPC迭代译码的判决反馈结果对SNR进行精估计。仿真结果表明,本方法能够在接收信号的均值和噪声方差等参数未知情况下,能有效完成LDPC码迭代译码的辅助工作;在误码率(BER)为10-5时,SNR估计后,LDPC码的译码性能距离理想情况下的译码性能,仅有约0.12dB以内损失。
In view of the fact that the current SNR estimation methods do not apply to multiplicative noise channels in optical communications, the effect of SNR mismatch on the performance of LDPC codes is analyzed. The simulation shows the necessity of accurate SNR estimation under the multiplicative noise channel. A method of SNR estimation based on quantized LDPC code decision feedback is proposed. First, the received optical signal is quantized, and a simplified rough estimate-maximum (EM) algorithm is used to estimate the quantized SNR. Then, The results of the iterative decoding decision feedback estimate the SNR. The simulation results show that the proposed method can effectively perform the iterative decoding of LDPC codes with unknown parameters such as mean value and noise variance of the received signal. When BER is 10-5, after SNR estimation, The decoding performance of LDPC codes is only about 0.12 dB less than the decoding performance under ideal conditions.