【摘要】 本文提出两个新的功率谱估值方法——双最大熵迭代法及模拟噪声迭代法。它们均出发于比最大熵估值法更一般的信号模式:自回归——滑动平均模式,扩展了估值适应的范围,提高了估值准确度。它们都具有比最大熵估值法更好的估值特性,并只要求低得多的估值阶数。这两方法也继承了最大熵估值法的优点,能很好地处理短输入数据。它们适用于要求准确估测功率谱分布的场合。本文叙述了推导过程,并列举了部分计算机模拟试验结果。更多还原

【Abstract】 Two new methods of spectral estimation, the Double-Maximum-Entropy It-erative Method and the Pseudo-Noise Iterative Method, are reported in this paper.Both arebased on a generalized model-the autoregressive-movingaverage model, where the Maximum Entropy Method (MEM) can not work properly. They provide betterperformance in spectral estimation than MEM,and can be used with lower orders.Like MEM these approaches can be used to treat input data of shorter length effectively.The algorithms are useful inall areas where exact spectral estimation is required. Both the derivation and the numerical experiments are described in this paper.更多还原