Method and device for realizing signal processing based on large-point Fourier transform

Abstract

The object of the present invention is to provide a method and a device for processing signals based on the large-point Fourier transform. Specifically, the large-point Fourier transform corresponding to the signal to be processed is converted into a two-dimensional small-point Fourier transform; according to the second sequence length, the product of the first Fourier transform and the twiddle factor is repeatedly executed for the second Fourier transform to obtain the result signal obtained by large-point Fourier transform. Compared with the prior art, the present invention converts the second sequence length corresponding to the two-dimensional small-point Fourier transform from the large-point Fourier transform corresponding to the target sequence length to the two-dimensional small-point Fourier transform according to the signal to be processed. The product of a Fourier transform result and the twiddle factor repeatedly executes the second Fourier transform, realizing the Fourier transform of a large number of points only by using the Fourier transform of a small number of points plus a multiplication factor, It not only reduces the complexity of the algorithm, but also reduces the resource overhead of the hardware device.

Description

technical field [0001] The invention relates to the technical field of signal processing, in particular to a technology for realizing signal processing based on large-point Fourier transform. Background technique [0002] FFT / IFFT (Fast Fourier Transform / Inverse Transform, Fast Fourier Transform / Inverse FastFourier Transform) is a common technique in signal processing. However, existing implementations such as converting small-point frequency-domain signals into large-point-number time-domain signals, or converting small-point-number time-domain signals into large-point-number frequency-domain signals usually use large-point number (large size) FFT / IFFT, or , using upsampling / downsampling, but the former requires a lot of hardware resources, and the implementation accuracy of the latter is lower than that of the former. For example, for the Cooley-Tukey FFT algorithm, the length is N=N 1 N 2 The DFT (Discrete Fourier Transform, Discrete Fourier Transform) is decomposed in...

AI Technical Summary

Problems solved by technology

[0004] in, n=N 2 no 1 +n 2 , k=N 1 k 2 +k 1 , but this method still needs to calculate N 2 Times N 1 Point DFT and N 1 Times N 2 Point DFT, therefore, the hardware resources and algorithm complexity required by this method still need to be further improved

Images