CN-122017738-A - Radar intra-pulse modulation signal waveform inversion method and device

CN122017738ACN 122017738 ACN122017738 ACN 122017738ACN-122017738-A

Abstract

The invention provides a radar intra-pulse modulation signal waveform inversion method and a device, which relate to the technical field of radar signal processing and comprise the steps of obtaining signal parameters; the method comprises the steps of generating an intermediate frequency signal according to signal generation functions, generating a time-frequency diagram through short-time Fourier transformation of the intermediate frequency signal, calculating a time range and a frequency range based on waveform parameters of the intermediate frequency signal, respectively normalizing the time range and the frequency range into a coordinate range of the time-frequency diagram to obtain a detection label, training the time-frequency diagram through the detection label according to a target detection model, carrying out positive and negative sample distribution based on a Hungary matching strategy to obtain an improved target detection model, carrying out unknown parameter target detection on the time-frequency diagram based on the improved target detection model to obtain a detection result, and carrying out inversion processing on the intermediate frequency signal based on the detection result to obtain an inversion result. The invention solves the problem of greatly reduced recognition accuracy.

Inventors

WU JIACHENG
ZHOU ZHENGCHUN
TAN RUIJIE
WANG ZICHENG
MENG HUA

Assignees

西南交通大学

Dates

Publication Date: 20260512
Application Date: 20251223

Claims (10)

1. A method for inverting a radar pulse modulated signal waveform, comprising: Acquiring signal parameters, wherein the signal parameters comprise sampling rate, frequency modulation slope and bandwidth; Generating an intermediate frequency signal according to the signal parameters, and generating a time-frequency diagram by performing short-time Fourier transform on the intermediate frequency signal; calculating a time range and a frequency range based on waveform parameters of the intermediate frequency signal respectively, and normalizing the time range and the frequency range to be in a coordinate range of the time-frequency diagram respectively to obtain a detection tag; training the time-frequency diagram through the detection label according to a target detection model, and carrying out positive and negative sample distribution based on a Hungary matching strategy to obtain an improved target detection model; performing target detection of unknown parameters on the time-frequency diagram based on the improved target detection model to obtain a detection result; And carrying out inversion processing on the intermediate frequency signal based on the detection result to obtain an inversion result.
2. The method of claim 1, wherein generating an intermediate frequency signal from the signal parameters according to a signal generation function, generating a time-frequency map by performing short-time fourier transform on the intermediate frequency signal, comprises: generating the signal parameters according to a signal generating function to obtain an intermediate frequency signal; Determining window length and overlapping length according to the characteristics of the intermediate frequency signals, segmenting the intermediate frequency signals into segments according to the window length and the overlapping length, and windowing each segment to obtain windowed segments; Performing fast Fourier transform on the windowed segments to obtain segment spectrums; and carrying out time sequence arrangement according to the segment frequency spectrum to generate a time-frequency diagram.
3. The method of claim 2, wherein the signal generating function includes a chirp waveform function, a pulse waveform function, and a discrete frequency code waveform generating function, and generating the signal parameters according to the signal generating function to obtain an intermediate frequency signal, including: Generating a chirp signal according to the sampling rate, the chirp rate and the bandwidth; generating a simple pulse signal according to the sampling rate and the frequency modulation slope according to a pulse waveform function; Generating a discrete frequency coded signal according to the sampling rate and the bandwidth by a discrete frequency coded waveform generating function; And fusing based on the linear frequency modulation signal, the simple pulse signal and the discrete frequency coding signal to obtain an intermediate frequency signal.
4. The method for inverting the waveform of the radar intra-pulse modulated signal according to claim 1, wherein the training of the time-frequency diagram by the detection tag according to the target detection model, the positive and negative sample distribution by the hungarian matching strategy, and the obtaining of the improved target detection model, comprises: extracting the time-frequency diagram features according to a target detection model to obtain a candidate target set; Based on a Top-N screening mechanism and a Hungary matching algorithm, carrying out similarity measurement matching on candidate targets in a candidate target set and real targets in the detection tag to obtain a positive and negative sample distribution result; respectively modeling a positive sample prediction boundary and a real target in a positive and negative sample distribution result as two-dimensional Gaussian distribution, and calculating a normalized bulldozer distance for predicting the two-dimensional Gaussian distribution and the real two-dimensional Gaussian distribution to obtain a boundary loss function; And based on back propagation, optimizing training is carried out on the target detection model through the boundary loss function, and parameters are updated to obtain an improved target detection model.
5. The method for inverting a radar pulse modulated signal waveform according to claim 4, wherein modeling the positive sample prediction boundary and the real target in the positive and negative sample distribution results as two-dimensional gaussian distributions, respectively, calculating a normalized bulldozer distance for predicting the two-dimensional gaussian distribution and the real two-dimensional gaussian distribution, and obtaining a boundary loss function, comprises: respectively modeling a positive sample prediction boundary and the real target in the positive and negative sample distribution result as two-dimensional Gaussian distribution to obtain a predicted two-dimensional Gaussian distribution and a real two-dimensional Gaussian distribution; Calculating the distance between the predicted two-dimensional Gaussian distribution and the real two-dimensional Gaussian distribution to obtain the distance of the bulldozer; Normalizing the bulldozer distance to obtain a normalized bulldozer distance; And based on the normalized bulldozer distance, measuring the similarity between the prediction boundary and the Gaussian distribution corresponding to the real boundary, and obtaining a boundary loss function.
6. The method for inverting the waveform of the radar pulse modulated signal according to claim 1, wherein inverting the intermediate frequency signal based on the detection result to obtain an inversion result comprises: extracting the position information of the target in the time-frequency diagram based on the target label in the detection result to obtain a target boundary parameter; Converting the normalized coordinates output in the target detection model into actual time and frequency values, and calculating the time span and the frequency variation range of the target in a time-frequency diagram through the target boundary parameters, the actual time and the frequency values to obtain target modulation parameters; extracting targets of the discrete frequency coding signals based on the output of the target detection model, and performing descending order sorting according to the vertical center coordinates of the discrete frequency coding signals to calculate a frequency coding sequence; and fusing the target modulation parameter and the frequency coding sequence, and outputting an inversion result.
7. A radar intra-pulse modulated signal waveform inversion apparatus, comprising: The acquisition module is used for acquiring signal parameters, wherein the signal parameters comprise sampling rate, frequency modulation slope and bandwidth; The generating module is used for generating intermediate frequency signals for the signal parameters according to a signal generating function and generating a time-frequency diagram by carrying out short-time Fourier transform on the intermediate frequency signals; The calculation module is used for respectively calculating a time range and a frequency range based on the waveform parameters of the intermediate frequency signals, and obtaining a detection tag by respectively normalizing the time range and the frequency range into the coordinate range of the time-frequency diagram; The training module is used for training the time-frequency diagram through the detection label according to the target detection model, and carrying out positive and negative sample distribution based on a Hungary matching strategy to obtain an improved target detection model; The detection module is used for carrying out unknown parameter target detection on the time-frequency diagram based on the improved target detection model to obtain a detection result; And the inversion module is used for carrying out inversion processing on the intermediate frequency signal based on the detection result to obtain an inversion result.
8. The radar pulse modulated signal waveform inversion apparatus of claim 7, wherein said generating module comprises: The generating unit is used for generating the signal parameters according to a signal generating function to obtain an intermediate frequency signal; The dividing unit is used for determining window length and overlapping length according to the characteristics of the intermediate frequency signals, dividing the intermediate frequency signals into segments according to the window length and the overlapping length, and windowing each segment to obtain windowed segments; the transformation unit is used for carrying out fast Fourier transformation on the windowed segments to obtain segment spectrums; And the arrangement unit is used for carrying out time sequence arrangement according to the segment frequency spectrum to generate a time-frequency diagram.
9. The radar pulse modulated signal waveform inversion apparatus of claim 8 wherein said signal generation function comprises a chirped waveform function and a pulse waveform function, a discrete frequency coded waveform generation function, said generation unit comprising: a first subunit, configured to generate a chirp signal according to a chirp waveform function for the sampling rate, the chirp rate, and the bandwidth; a second subunit, configured to generate a simple pulse signal according to a pulse waveform function for the sampling rate and the frequency modulation slope; a third subunit, configured to generate a discrete frequency encoded signal according to the sampling rate and the bandwidth by using a discrete frequency encoded waveform generating function; And the fusion subunit is used for fusing the linear frequency modulation signal, the simple pulse signal and the discrete frequency coding signal to obtain an intermediate frequency signal.
10. The radar pulse modulated signal waveform inversion apparatus of claim 7, wherein said training module comprises: the extraction unit is used for extracting the time-frequency diagram features according to a target detection model to obtain a candidate target set; The sorting unit is used for carrying out similarity measurement matching on candidate targets in the candidate target set and real targets in the detection tag based on a Top-N screening mechanism and a Hungary matching algorithm to obtain a positive and negative sample distribution result; the modeling unit is used for modeling a positive sample prediction boundary and a real target in the positive and negative sample distribution result as two-dimensional Gaussian distribution respectively, and calculating and normalizing bulldozer distance to obtain a boundary loss function; And the updating unit is used for carrying out optimization training on the target detection model through the boundary loss function based on back propagation and updating parameters to obtain an improved target detection model.

Description

Radar intra-pulse modulation signal waveform inversion method and device Technical Field The invention relates to the technical field of radar signal processing, in particular to a radar intra-pulse modulation signal waveform inversion method and device. Background In radar signal processing technology, conventional radar signal recognition methods mainly rely on manually designed feature extraction and classification algorithms. However, in processing complex signals, especially in the face of large-scale, high-dimensional radar signal data, the artificially designed feature extraction process often requires a large number of computational operations to be performed, which makes it difficult to meet the application scenario requirements with extremely high real-time requirements. In addition, manually designed feature extraction algorithms are generally constructed based on specific signal models and priori knowledge, and when the feature distribution of actual signals has significant differences with a preset model, the performance of the algorithms can be drastically reduced, so that the recognition accuracy is greatly reduced. Therefore, a method and a device for inverting the waveform of the radar pulse modulation signal are needed, and the problem that the recognition accuracy is greatly reduced is solved. Disclosure of Invention The invention aims to provide a radar intra-pulse modulation signal waveform inversion method and device so as to solve the problems. In order to achieve the above purpose, the technical scheme adopted by the invention is as follows: In a first aspect, the present application provides a radar intra-pulse modulated signal waveform inversion method, including: Acquiring signal parameters, wherein the signal parameters comprise sampling rate, frequency modulation slope and bandwidth; Generating an intermediate frequency signal according to the signal parameters, and generating a time-frequency diagram by performing short-time Fourier transform on the intermediate frequency signal; calculating a time range and a frequency range based on waveform parameters of the intermediate frequency signal respectively, and normalizing the time range and the frequency range to be in a coordinate range of the time-frequency diagram respectively to obtain a detection tag; training the time-frequency diagram through the detection label according to a target detection model, and carrying out positive and negative sample distribution based on a Hungary matching strategy to obtain an improved target detection model; performing target detection of unknown parameters on the time-frequency diagram based on the improved target detection model to obtain a detection result; And carrying out inversion processing on the intermediate frequency signal based on the detection result to obtain an inversion result. In a second aspect, the present application further provides a radar pulse modulated signal waveform inversion apparatus, including: The acquisition module is used for acquiring signal parameters, wherein the signal parameters comprise sampling rate, frequency modulation slope and bandwidth; The generating module is used for generating intermediate frequency signals for the signal parameters according to a signal generating function and generating a time-frequency diagram by carrying out short-time Fourier transform on the intermediate frequency signals; The calculation module is used for respectively calculating a time range and a frequency range based on the waveform parameters of the intermediate frequency signals, and obtaining a detection tag by respectively normalizing the time range and the frequency range into the coordinate range of the time-frequency diagram; The training module is used for training the time-frequency diagram through the detection label according to the target detection model, and carrying out positive and negative sample distribution based on a Hungary matching strategy to obtain an improved target detection model; The detection module is used for carrying out unknown parameter target detection on the time-frequency diagram based on the improved target detection model to obtain a detection result; And the inversion module is used for carrying out inversion processing on the intermediate frequency signal based on the detection result to obtain an inversion result. The beneficial effects of the invention are as follows: The method and the device generate the time-frequency diagram through short-time Fourier transform, are used for simultaneously displaying the time domain and frequency domain information of the signals, so that the characteristics of the signals are more abundant and visual, the problem that the characteristics of the signals are extracted incompletely is solved, and the Hungary matching strategy is used for training a target detection model. Compared with the traditional cross-correlation threshold value distribution and non-maximum value suppression method,