微动特征提取递归分解快速傅里叶变换算法的研究及应用

Research and Application on a Recursively Fast Fourier Transformation Decomposion Algorithm for Micro-Motion Features Extraction

摘要: 针对雷达目标微动特征提取提出一种递归分解快速傅里叶变换算法，并使用C语言编程实现。算法可对任意非素数点数序列实现快速离散傅里叶变换。以复数乘法和加法次数为指标对算法进行性能分析，结果表明对序列点数为2的整次幂，算法效率与传统基–2快速傅里叶变换算法相同；对其它序列点数为非素数情况，算法效率优于直接计算离散傅里叶变换且简化了对输入序列重排过程。目前该算法已应用于雷达信号处理的微动目标特征提取中。

Abstract: Aiming at micro-motion features extraction on radar targets, a recursively fast Fourier transfor-mation (FFT) decomposition algorithm is proposed, which is programmed and implemented in the C language. In this algorithm the fast discrete Fourier transformation is carried out on an arbitrary selected sequence whose elements are at a non-prime number. Performance analysis is addressed on the specification of complex multiplication and addition rounds. It demonstrates that if the number of sequence elements equals 2 raised to an integer power, the algorithm efficiency is the same as the traditional 2-based FFT algorithm, whereas if the number doesn’t but it is still a non-prime number, this efficiency is superior to the direct discrete Fourier transformation com-puting. Furthermore, this algorithm simplifies the process of rearranging the input sequences. Currently this algorithm is utilized in the micro-motion target features extraction in radar signal processings.