加性高斯白噪声
解码方法
极地的
信仰传播
信噪比(成像)
算法
计算机科学
灵敏度(控制系统)
噪音(视频)
语音识别
作者
Siyu Zhang,Behnam Shahrrava
标识
DOI:10.1109/iemcon53756.2021.9623241
摘要
Intuitively, for polar decoding, an accurate signal-to-noise ratio (SNR) estimation is necessary when using belief propagation (BP) polar decoders. It has been claimed in literature that the BP decoding is sensitive to SNR underestimation. However, there are two kinds of BP decoders, the conventional BP and “min-sum (MS)” based BP, adopted in practice. In this paper, an investigation toward SNR sensitivity is implemented based on original and MS based BP decoders by utilizing mathematical proof. It is shown by that the conventional BP decoder is sensitive to SNR mismatch, while the MS based BP decoder is not. However, simulation results indicate that in practical cases where the SNR mismatch in not significant, both conventional BP and MS based BP decoders are less affected by SNR mismatch for additive white Gaussian noise (AWGN) channels.
科研通智能强力驱动
Strongly Powered by AbleSci AI