High-throughput FFT/IFFT-based FPGA signal processing method

Abstract

The invention discloses a high-throughput FFT / IFFT-based FPGA signal processing method and relates to an FPGA technology. The method comprises the following steps of 1) reading data of a low-frequency signal; 2) writing contents of the data of the low-frequency signal into a first storage region and other storage regions of an RAM in sequence; 3) under the control of a high-frequency read clock, reading the data contents written into the storage regions in sequence through a sliding window, and outputting the data contents in the sliding window in sequence to form data flow; and 4) from a starting moment of the data flow, enabling an Nth FPGA to enter a data processing stage after (N-1) waiting cycles to read the data flow, performing FFT calculation, and outputting a calculation result. Through the RAM, low-rate data is converted into high-rate data, so that subsequent modules can work at a relatively high frequency, the purpose of reducing degree of parallelism is achieved, and finally the hardware consumption is reduced. The method is simple in algorithm and easy for engineering realization.

Description

technical field [0001] The present invention relates to FPGA technology. Background technique [0002] Discrete fourier transform (DFT for short) and inverse discrete fourier transform (IDFT for short) are important transformation tools in the field of digital signal processing, both in signal spectrum analysis and OFDM system modulation and demodulation It is realized by DFT / IDFT. However, direct calculation of DFT / IDFT requires a lot of calculations. Its fast algorithm (fast fourier transform referred to as FFT) uses the properties of Fourier transform coefficients to greatly reduce the amount of multiplication operations, making Fourier transform widely used in engineering. [0003] Since cooley-tukey proposed the DFT fast algorithm, there has been more in-depth research on the DFT algorithm, mainly including the prime factorization method, the winogard algorithm (WFTA), the circular convolution algorithm, etc., but these algorithms mainly focus on the DFT transformation...

AI Technical Summary

Problems solved by technology

[0003] Since cooley-tukey proposed the DFT fast algorithm, there has been more in-depth research on the DFT algorithm, mainly including the prime factorization method, the winogard algorithm (WFTA), the circular convolution algorithm, etc., but these algorithms mainly focus on the DFT transformation itself, and there are few studies on it. Research on the throughput of DFT transformation. When higher throughput is required, FPGA usually implements it in parallel processing.

[0004] When the FPGA implementation of the prior art is implemented, simply adopting a parallel method to improve the FFT throughput rate will consume a large amount of hardware resources

Images