基于BLUE的多窗口谱估计

Multi-Taper Spectral Estimation Based on BLUE

文章提出了一种基于BLUE的多窗口功率谱估计算法。该算法是在无偏估计的约束条件下,使其方差最小化而得到的特征谱的最佳线性组合。通过仿真实例分析得到,与自适应的MTM相比,该算法的计算量大大减小了。由于正弦窗具有解析的表达式,不需要进行特征值分解,并且有良好的统计性能,因此,使用正弦窗的MTM可以进一步降低计算量并改善估计性能。

A new arithmetic for MTM based on BLUE is proposed. The best linear combination of eigen-spectrums is obtained by minimizing the variance of estimated spectrum on the constraint of unbiased estimation. An example and the analysis of it are given. It is showed that the computational load of this method is greatly reduced compared with adaptive MTM. Since the sinusoidal tapers have analytic expressions, it is not necessary to conduct numerical eigenvalue decomposition. Moreover, they have good statistical performances. Hence, MTM using the sinusoidal tapers has less computational load and better estimation performances.