摘要
信噪比(SNR)估计是信道估计的重要组成部分,很多先进通信系统和信号处理方法都将信噪比作为先验信息,因此对信噪比估计方法的研究尤为重要。基于多进制相移键控(MPSK)信号模型,对最大似然类、矩估计类和空间分解类算法进行了性能分析和仿真。在一定条件下,上述算法的估计偏差在[0,20]d B区间内均小于1 d B,其中最大似然类算法估计精确度最高,但易受解调误码率影响;矩估计类算法在低信噪比时性能较好,高信噪比时易受算法自噪声影响;空间分解类算法适应性最强,但实时性较差。通过对上述算法一致性和差异性分析,总结了信噪比估计的研究进展和主要问题,明确了复杂调制信号宽范围信噪比估计和空间分解方法的研究方向,为后续研究提供了解决思路和改进措施。
Signal to Noise Ratio(SNR) estimation is important for channel estimation. Many communication systems and signal processing algorithms need SNR as prior information. SNR estimation algorithms such as max likelihood, statistics and space breaking algorithms are simulated based on Multiple Phase Shift Keying(MPSK) signal models. The estimation bias of these algorithms can be less than 1 d B under some conditions in [0,20] d B. Max likelihood algorithm is the most exact one, but is liable to be influenced by demodulation errors. Statistics algorithm is better in low SNR than in high SNR because of the algorithms' noises. Space breaking algorithm is the most adaptable, but has worse real-time performance. Through analyzing the consistency and diversity, the research advancements and existing problems are summarized and the research direction such as complex modulation signals SNR estimation in wide area and space breaking methods are confirmed. In the end, some solutions and improvements are proposed to resolve the existing problems.
出处
《太赫兹科学与电子信息学报》
2016年第1期81-87,共7页
Journal of Terahertz Science and Electronic Information Technology
关键词
信噪比估计
最大似然准则
统计量
奇异值分解
数据拟合
SNR estimation
Max-Likelihood
statistics
Singular Value Decomposition(SVD)
data fitting
