用于快速自回归参量谱估计的基4EFFT算法

The Base 4 EFFFTT Algorithm Used in Fast Estimation of Autoregressive Parametric Spectrum

<正> 在参量模型谱估计中,其主要运算量有两部分,即由信号采样序列估计模型参数及由模型参数计算谱.当信号样本数较少时,后半部分的运算量将明显大于前半部分.影响处理速度.本文讨论由 AR 模型参数计算谱值问题.

This paper introduces a base 4 expanded FFT(EFFT)algorithm in which discrete spectra are computed with AR coefficients.The good structure of base 4 FFT is employed in the algorithm. When the number of input points is less than that of transform points,zeros are not needed to add. So,the number of butterflies can be reduced.Because the number of AR coefficients is always much iess than the number of spectral samples in general,the application of the base 4 EFFT can decrease a great deal of computation and increase the velocity of signal processing.The algorithm is exactly derived and analysed in this paper.Finally,the computational velocity and precision of the method are compared with the FFT and the direct method.