Single-frequency signal short sample high precision frequency measurement method and device based on relatively prime perception

Abstract

The invention discloses a single-frequency signal short sample high precision frequency measurement method and device based on relatively prime perception; the method comprises the following steps: carrying out discrete fourier transformation (DFT) for a multichannel output sequence formed by multichannel sparse signals, thus obtaining a DET converted output sequence; carrying out spectrum peak scanning for the DET converted output sequence so as to obtain a spectrum peak position index, using the index to obtain a peak DET spectral line, using a phase difference method to carry out spectrum correction, thus obtaining corrected frequency; using the corrected frequency as a remainder, using a Chinese remainder theorem in a closed parse form to process, thus reconstructing an original signal frequency. The device comprises a digital signal processor, an output drive, and a display circuit thereof; the digital signal processor is used for calculating so as to reconstruct the original signal frequency; the output drive and the display circuit thereof can display the original signal frequency. The single-frequency signal short sample high precision frequency measurement method and device can reduce spectrum estimation time delay, and improve sample utilization rate.

Description

technical field [0001] The invention relates to the field of digital signal processing, in particular to signal undersampling, mutual prime perception, polyphase filtering, frequency spectrum correction, and parameter reconstruction based on the Chinese remainder theorem. Background technique [0002] Spectrum estimation is related to many actual physical quantity measurements (such as radar wind speed measurement in atmospheric dynamics research [1] , Speed ​​measurement in vibration analysis [2] etc.) closely related signal processing methods [3,4] , is widely used in various disciplines such as electronics, electricity and automation, and is the basic problem of digital signal processing. Only when the traditional spectrum estimation satisfies Shannon sampling theorem (that is, at least two samples are required to be collected in a signal cycle), the signal will not be distorted, and it is possible to recover the signal effectively. However, in many engineering applica...

AI Technical Summary

Problems solved by technology

[0005] However, the method mentioned in literature [11] has two defects that cannot be ignored. One is that the frequency estimation accuracy is not high enough. This is because no spectral correction is added in the spectrum estimation process, and the final spectral line can only be roughly positioned at the integer position , the fractional shift of the signal frequency is artificially ignored, resulting in inevitable errors in spectrum estimation; the second is that it consumes a lot of samples, which is due to the need to do cross-correlation processing on the two signals. In order to obtain accurate spectrum estimation, it takes A large number of snapshots (snapshot, the number of snapshots is equal to the product of two downsampling factors, that is, MN) averages, which greatly increases the time delay of spectrum estimation and increases the requirements and cost of hardware equipment. Not conducive to large-scale programming and practical application

Due to the complexity of real signals, these cases are likely to occur in large numbers, seriously affecting the performance of spectrum estimation

Images