该文给出一种直观的信噪比估计非理想时的先验信息建模,使用外信息转移图较为全面地分析了非理想信噪比估计对带迭代译码的比特交织编码调制系统性能的影响,分析结果表明非理想信噪比估计对典型的星座类型和标号在高斯和瑞利信道下的影响相似;通过对不同帧长的系统性能比较可知,长帧下信噪比估计偏小在-3dB以下时性能会大大下降,而估计偏大则没有影响;而对于中短帧而言,在相同的偏差下,偏小的影响要大于偏大的影响;此外,还给出由拟合得到的8PSK高斯信道和16QAM高斯与瑞利信道下的信噪比估计算法,最后通过系统仿真验证了该算法的有效性。
This paper proposes an intuitive modeling of a prior information with Signal-to-Noise-Ratio (SNR) mismatch for Bit-Interleaved Coded Modulation with Iterative Decoding (BICM-ID) systems. The extrinsic information transfer chart is applied to analyze the impact on BICM-ID caused by non-ideal SNR estimation for different constellations, labeling methods and channel types. It is found that the results are semblable for these cases. The performance of BICM-ID systems with different frame length is compared with SNR mismatch varied. The system performance decreases acutely when SNR mismatch is smaller than -3dB and maintains when it is bigger than 0dB for long frame, while it is more sensitive for SNR underestimation than overestimation for moderate and short frames. Besides these, SNR estimation methods are presented for 8PSK in Gaussian channel and 16QAM in Gaussian and Rayleigh channels. Finally, simulation results show these methods are effective.