Radar intra-pulse modulation signal waveform inversion method and device

Abstract

The invention provides a radar intra-pulse modulation signal waveform inversion method and device, and relates to the technical field of radar signal processing, and the method comprises the steps: obtaining a signal parameter; generating an intermediate frequency signal for the signal parameter according to the signal generation function, and performing short-time Fourier transform on the intermediate frequency signal to generate a time-frequency diagram; respectively calculating a time range and a frequency range based on waveform parameters of the intermediate frequency signal, and respectively normalizing the time range and the frequency range into a coordinate range of a time-frequency diagram to obtain a detection label; according to the target detection model, training the time-frequency diagram through detection labels, and carrying out positive and negative sample distribution based on a Hungary matching strategy to obtain an improved target detection model; performing unknown parameter target detection on the time-frequency graph based on an improved target detection model to obtain a detection result; and performing inversion processing on the intermediate frequency signal based on the detection result to obtain an inversion result. According to the invention, the problem that the recognition accuracy is greatly reduced is solved.

Description

Technical Field

[0001] This invention relates to the field of radar signal processing technology, and more specifically, to a method and apparatus for inverting the waveform of an intra-pulse modulated radar signal. Background Technology

[0002] In radar signal processing technology, traditional radar signal recognition methods mainly rely on manually designed feature extraction and classification algorithms. However, when processing complex signals, especially large-scale, high-dimensional radar signal data, manually designed feature extraction processes often require a large number of computational operations, making them unsuitable for applications with extremely high real-time requirements. Furthermore, manually designed feature extraction algorithms are typically built upon specific signal models and prior knowledge. When the feature distribution of the actual signal differs significantly from the preset model, the performance of these algorithms deteriorates sharply, leading to a substantial decrease in recognition accuracy.

[0003] Therefore, there is an urgent need for a method and device for inverting the waveform of radar pulse modulation signals, which solves the problem of a significant reduction in recognition accuracy. Summary of the Invention

[0004] The purpose of this invention is to provide a method and apparatus for inverting the waveform of radar intra-pulse modulated signals, thereby improving the aforementioned problems. To achieve the above objective, the technical solution adopted by this invention is as follows: In a first aspect, this application provides a method for inverting the waveform of an intra-pulse modulated radar signal, including: Acquire signal parameters, including sampling rate, frequency modulation slope, and bandwidth; An intermediate frequency (IF) signal is generated from the signal parameters according to the signal generation function, and a time-frequency diagram is generated by performing a short-time Fourier transform on the IF signal. The time range and frequency range are calculated based on the waveform parameters of the intermediate frequency signal. The detection tag is obtained by normalizing the time range and frequency range to the coordinate range of the time-frequency graph. Based on the target detection model, the time-frequency map is trained using the detection labels, and positive and negative samples are assigned based on the Hungarian matching strategy to obtain an improved target detection model. The improved target detection model is used to perform target detection with unknown parameters on the time-frequency map to obtain the detection results; Based on the detection results, the intermediate frequency signal is inverted to obtain the inversion result.

[0005] Secondly, this application also provides a radar intra-pulse modulation signal waveform inversion device, comprising: The acquisition module is used to acquire signal parameters, including sampling rate, frequency modulation slope, and bandwidth. The generation module is used to generate an intermediate frequency signal based on the signal parameters according to the signal generation function, and to generate a time-frequency diagram by performing a short-time Fourier transform on the intermediate frequency signal. The calculation module is used to calculate the time range and frequency range based on the waveform parameters of the intermediate frequency signal, and obtain the detection tag by normalizing the time range and frequency range to the coordinate range of the time-frequency graph. The training module is used to train the time-frequency map based on the target detection model using the detection labels, and to perform positive and negative sample allocation based on the Hungarian matching strategy to obtain an improved target detection model. The detection module is used to perform unknown parameter target detection on the time-frequency map based on the improved target detection model, and obtain the detection result; The inversion module is used to perform inversion processing on the intermediate frequency signal based on the detection results to obtain the inversion result.

[0006] The beneficial effects of this invention are as follows: This invention generates a time-frequency diagram using short-time Fourier transform, simultaneously displaying the time and frequency domain information of a signal. This makes the signal features richer and more intuitive, solving the problem of incomplete signal feature extraction. The invention also employs a Hungarian matching strategy for training the target detection model. Compared to traditional cross-union ratio (CUNR) threshold allocation and non-maximum suppression methods, the Hungarian matching strategy can accurately allocate positive samples, avoiding ambiguity in sample allocation caused by overlapping small targets. This precise sample allocation further optimizes the training process of the target detection model, enabling the improved model to more accurately identify target signals during detection, effectively reducing false detection and false negative rates. Based on the detection results of the improved target detection model, inversion processing of the intermediate frequency signal yields more accurate inversion results, thus better restoring the original radar intra-pulse modulation signal waveform. In summary, this invention solves the problem of significantly reduced recognition accuracy.

[0007] Other features and advantages of the invention will be set forth in the following description, and will be apparent in part from the description, or may be learned by practicing embodiments of the invention. The objects and other advantages of the invention may be realized and obtained by means of the structures particularly pointed out in the written description, claims, and drawings. Attached Figure Description

[0008] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly introduced below. It should be understood that the following drawings only show some embodiments of the present invention and should not be regarded as a limitation on the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.

[0009] Figure 1 This is a schematic diagram of the radar intra-pulse modulation signal waveform inversion method described in this embodiment of the invention; Figure 2 This is a schematic diagram of the radar intra-pulse modulation signal waveform inversion device described in this embodiment of the invention.

[0010] The diagram is labeled as follows: 800, Radar pulse modulation signal waveform inversion device; 801, Processor; 802, Memory; 803, Multimedia component; 804, I / O interface; 805, Communication component. Detailed Implementation

[0011] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of the present invention. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0012] It should be noted that similar reference numerals and letters in the following figures indicate similar items; therefore, once an item is defined in one figure, it does not need to be further defined and explained in subsequent figures. Furthermore, in the description of this invention, terms such as "first," "second," etc., are used only to distinguish descriptions and should not be construed as indicating or implying relative importance.

[0013] Example 1: This embodiment provides a method for inverting the waveform of an intra-pulse modulated radar signal.

[0014] See Figure 1 The figure shows that the method includes steps S1 to S6, including: S1: Acquire signal parameters, including sampling rate, frequency modulation slope, and bandwidth; In this step, the signal parameters also include pulse repetition frequency, bandwidth carrier frequency, coding sequence, and sweep direction.

[0015] S2: Generate an intermediate frequency signal from the signal parameters according to the signal generation function, and generate a time-frequency diagram by performing a short-time Fourier transform on the intermediate frequency signal; To clarify the specific method for obtaining the time-frequency diagram, step S2 includes S21 to S24, specifically: S21: Generate the signal parameters according to the signal generation function to obtain the intermediate frequency signal; In this step, the signal generation function includes a linear frequency modulated waveform function, a pulse waveform function, and a discrete frequency coded waveform generation function.

[0016] To clarify the specific method for acquiring the intermediate frequency signal, step S21 includes S211 to S214, specifically: S211: Generate a linear frequency modulated signal based on the sampling rate, the frequency modulation slope, and the bandwidth according to the linear frequency modulated waveform function; In this step, input the sampling rate. FM slope Pulse repetition frequency ,bandwidth Bandwidth and carrier frequency The frequency sweep direction is then used to generate the linear frequency modulated waveform function to obtain the linear frequency modulated signal.

[0017] The expression for the linear frequency modulated signal is: (1) In the above formula (1), It is a linear frequency modulated signal. The amplitude of the transmitted signal, It is a cosine function. Pi For bandwidth carrier frequency, For time variables, This represents the frequency modulation slope.

[0018] S212: Generate a simple pulse signal based on the sampling rate and the frequency modulation slope according to the pulse waveform function; In this step, input the sampling rate. FM slope Pulse repetition frequency Bandwidth and carrier frequency The signal is generated in the pulse waveform function to obtain a simple pulse signal. The expression for the simple pulse signal is: (2) In the above formula (2), It is a simple pulse signal. The amplitude of the transmitted signal, It is an exponential function. The imaginary unit, Pi For bandwidth carrier frequency, It is a time variable.

[0019] S213: Generate a discrete frequency encoded signal based on the sampling rate and the bandwidth according to the discrete frequency encoded waveform generation function; In this step, frequency segmentation and encoding are first performed according to the discrete frequency encoded waveform generation function, and the bandwidth is... Average score There are 1 frequency band, and the frequency interval of each frequency band is 1. According to the encoded sequence Assign a specific frequency value to each frequency band to generate a frequency vector. The duration of each frequency band is ,in It is a constant. The total duration of the entire signal is... Complex signals passing through each frequency band splicing, based on pulse repetition frequency and sampling rate Calculate the number of periodic samples for each pulse. And generate multiple discrete frequency encoded signals.

[0020] S214: The intermediate frequency signal is obtained by fusing the linear frequency modulated signal, the simple pulse signal, and the discrete frequency encoded signal.

[0021] In this step, the linear frequency modulated signal, the simple pulse signal, and the discrete frequency encoded signal are fused together, and each signal is categorized to obtain the intermediate frequency signal.

[0022] S22: Determine the window length and overlap length based on the characteristics of the intermediate frequency signal, segment the intermediate frequency signal using the window length and overlap length, and perform windowing processing on each segment to obtain windowed segments; In this step, the intermediate frequency signal is divided according to the window length. Each segment moves the sampling points by the overlap length. By selecting an appropriate window function, each segment is multiplied by the window function to obtain the windowed segment.

[0023] S23: Perform a fast Fourier transform on the windowed segment to obtain the segment spectrum; In this step, the spectrum obtained after the fast Fourier transform of each windowed segment is the segment spectrum.

[0024] The expression for the short-time Fourier transform is: (3) In the above formula (3), This is the result of the short-time Fourier transform. The integral symbol is used. The input time-domain signal, For window functions, It is a complex exponential function. The imaginary unit, For time variables, Pi For frequency variables, It is the integral variable.

[0025] S24: Arrange the spectrum segments in time order to generate a time-frequency diagram.

[0026] S3: Calculate the time range and frequency range based on the waveform parameters of the intermediate frequency signal, and obtain the detection tag by normalizing the time range and frequency range to the coordinate range of the time-frequency graph. To clarify the specific method for obtaining the detection tags, step S3 includes S31 to S34, specifically: S31: Calculate the time range of the intermediate frequency signal based on the start and end times in the waveform parameters; In this step, the start time is: (4) In the above formula (4), For the first The start time of each signal Sampling rate, For the first The starting index of each signal. Lead time; The end time is: In the above formula (5), For the first The end time of each signal For the first The start time of each signal This represents the frequency modulation slope.

[0027] S32: Calculate the frequency range of the intermediate frequency signal based on the start and end frequencies in the waveform parameters; In this step, the start and end frequencies of the linear frequency modulation signal are: (6) In the above formula (6), This is the end frequency of the linear frequency modulated signal. The starting frequency of the linear frequency modulated signal. For bandwidth carrier frequency, The pulse bandwidth; The start and end frequencies of the simple pulse signal are: (7) In the above formula (7), This is the starting frequency of the pulse signal. The end frequency of the pulse signal. For bandwidth carrier frequency, Frequency range; The start and end frequencies of the discrete frequency encoded signal are: (8) In the above formula (8), The starting frequency of the discrete frequency encoded signal. The end frequency of the discrete frequency encoded signal. For bandwidth carrier frequency, This refers to the frequency range.

[0028] The frequency range expression for each frequency band is: (9) In the above formula (9), For the frequency range of each frequency band, The first frequency vector is the frequency vector. The frequency value of each element, This is a time parameter.

[0029] S33: Normalize the time range and the frequency range to the coordinate range of the time-frequency graph to obtain the normalized coordinate range; S34: Determine the category label of the signal modulation type in the time-frequency diagram based on the labeling category of the intermediate frequency signal, and generate a detection label by combining it with the normalized coordinate range.

[0030] S4: Based on the target detection model, the time-frequency map is trained using the detection labels, and positive and negative samples are assigned based on the Hungarian matching strategy to obtain an improved target detection model; To clarify the specific acquisition method of the improved target detection model, step S4 includes S41 to S44, specifically: S41: Extract the features from the time-frequency map based on the target detection model to obtain a set of candidate targets; In this step, based on the target detection model, the time-frequency map features are extracted, and possible target regions are identified, forming a candidate target set containing N candidate targets. Through multi-scale feature modeling capabilities, the local structure and target features in the time-frequency map are effectively captured.

[0031] S42: Based on the Top-N screening mechanism and the Hungarian matching algorithm, the candidate targets in the candidate target set are matched with the real targets in the detection labels to obtain the positive and negative sample allocation results; In this step, the N candidate targets are optimally assigned to the class-labeled real targets using the one-to-one Hungarian matching algorithm. Let the set of class-labeled real targets be denoted as . .

[0032] Due to the sparsity of targets on the time-frequency graph, Often much smaller Therefore, in order to achieve matching between candidate targets and real targets, the set is first... Complete the set to obtain the complete set. , where when index hour, .

[0033] At this point, for the set With sets Matching is performed based on minimizing the cost function, with the objective function being: (10) In the above formula (10), This is the index of the candidate target for the best match. To find the parameter that minimizes the subsequent expression, To From 1 to Summation, For matching cost function, For the first A label for a real target. For the first Real targets and candidate targets Predicted labels, For candidate target index, A set of candidate target indices; In other words, the matching cost function comprehensively considers the positional error, scale difference, and class prediction error between the candidate target and the real target, ensuring that each real target is matched with only one optimal candidate target. The unmatched Top-N candidate targets are regarded as negative samples and used for penalties during training to enhance the model's discriminative ability.

[0034] S43: In the positive and negative sample allocation results, the predicted boundary of the positive sample and the real target are modeled as two-dimensional Gaussian distributions respectively. The normalized bulldozer distance is calculated on the predicted two-dimensional Gaussian distribution and the real two-dimensional Gaussian distribution to obtain the boundary loss function. To clarify the specific method for obtaining the boundary loss function, step S43 includes S431 to S434, specifically: S431: Model the positive sample prediction boundary and the real target in the positive and negative sample allocation results as two-dimensional Gaussian distributions respectively to obtain the predicted two-dimensional Gaussian distribution and the real two-dimensional Gaussian distribution; In this step, the center coordinates and covariance matrix of the prediction boundary are extracted from the positive sample allocation results, and the prediction two-dimensional Gaussian distribution is obtained by modeling through the probability density function of the two-dimensional Gaussian distribution. Similarly, the center coordinates and covariance matrix of the real boundary are extracted from the real target, and the real two-dimensional Gaussian distribution is obtained by modeling it using the probability density function of the two-dimensional Gaussian distribution.

[0035] S432: Calculate the distance between the predicted two-dimensional Gaussian distribution and the actual two-dimensional Gaussian distribution to obtain the bulldozer distance; In this step, the difference between the two Gaussian distributions is quantified by the bulldozer distance.

[0036] S433: Normalize the bulldozer distance to obtain the normalized bulldozer distance; In this step, the bulldozer distance is normalized to a fixed range to eliminate the influence of different data ranges on distance calculation and improve the accuracy of similarity measurement.

[0037] S434: Based on the normalized bulldozer distance, measure the similarity between the Gaussian distributions corresponding to the predicted boundary and the true boundary to obtain the boundary loss function.

[0038] In this step, the boundary loss function is optimized for small target detection in the radar time-frequency map. By modeling the bounding box as a two-dimensional Gaussian distribution and calculating the bulldozer distance, it can not only more accurately reflect the energy diffusion characteristics of the signal in the time-frequency map, but its normalized distance metric also gives the algorithm excellent scale robustness.

[0039] S44: Based on backpropagation, the target detection model is optimized and trained using the boundary loss function, and the parameters are updated to obtain an improved target detection model.

[0040] S5: Based on the improved target detection model, perform unknown parameter target detection on the time-frequency map to obtain the detection result; S6: Based on the detection results, the intermediate frequency signal is inverted to obtain the inversion result.

[0041] To clarify the specific method for obtaining the inversion results, step S6 includes S61 to S64, specifically: S61: Extract the target's position information in the time-frequency map based on the target label in the detection result to obtain the target boundary parameters; In this step, the location information includes the target center coordinates, width, and height, and the target label is used to accurately extract the target's position and pulse modulation category in the time-frequency graph; The target label includes the target category and normalized coordinate information.

[0042] S62: Convert the normalized coordinates output in the target detection model into actual time and frequency values, and calculate the time span and frequency variation range of the target in the time-frequency diagram using the target boundary parameters, actual time and frequency values ​​to obtain the target modulation parameters; In this step, the target modulation parameters are: (11) In the above formula (11), For bandwidth carrier frequency, This represents the minimum ordinate value of the target in the time-frequency graph. The frequency range of the time-frequency graph. For pulse bandwidth, The maximum ordinate value of the target in the time-frequency graph. For frequency range, The maximum x-coordinate value of the target in the time-frequency graph. The minimum x-coordinate value of the target in the time-frequency graph. The time range of the time-frequency graph. This is the pulse repetition frequency.

[0043] S63: Extract the target from the discrete frequency encoded signal based on the output of the target detection model, sort the discrete frequency encoded signal in descending order according to the vertical center coordinates of the discrete frequency encoded signal, and calculate the frequency encoded sequence; S64: Fuse the target modulation parameters and the frequency coding sequence to output the inversion result.

[0044] Example 2: This embodiment provides a radar intra-pulse modulation signal waveform inversion device, the device comprising: The acquisition module is used to acquire signal parameters, including sampling rate, frequency modulation slope, and bandwidth. The generation module is used to generate an intermediate frequency signal based on the signal parameters according to the signal generation function, and to generate a time-frequency diagram by performing a short-time Fourier transform on the intermediate frequency signal. To clarify the specific methods for obtaining the generated modules, the following are included: The generation unit is used to generate the signal parameters according to the signal generation function to obtain the intermediate frequency signal; To clarify the specific methods for obtaining the generated units, the following are included: The first subunit is used to generate a linear frequency modulated signal based on the sampling rate, the frequency modulation slope, and the bandwidth according to the linear frequency modulated waveform function; The second subunit is used to generate a simple pulse signal based on the sampling rate and the frequency modulation slope according to the pulse waveform function; The third subunit is used to generate a discrete frequency encoded signal based on the sampling rate and the bandwidth according to the discrete frequency encoded waveform generation function; The fusion subunit is used to fuse the linear frequency modulated signal, the simple pulse signal, and the discrete frequency coded signal to obtain an intermediate frequency signal.

[0045] The segmentation unit is used to determine the window length and overlap length according to the characteristics of the intermediate frequency signal, segment the intermediate frequency signal into segments using the window length and overlap length, and perform windowing processing on each segment to obtain windowed segments; The transformation unit is used to perform a fast Fourier transform on the windowed segment to obtain the segment spectrum; The arrangement unit is used to arrange the spectrum of the segments in time order to generate a time-frequency diagram.

[0046] The calculation module is used to calculate the time range and frequency range based on the waveform parameters of the intermediate frequency signal, and obtain the detection tag by normalizing the time range and frequency range to the coordinate range of the time-frequency graph. The training module is used to train the time-frequency map based on the target detection model using the detection labels, and to perform positive and negative sample allocation based on the Hungarian matching strategy to obtain an improved target detection model. To clarify the specific methods for obtaining the training modules, the following are included: The extraction unit is used to extract features from the time-frequency map based on the target detection model to obtain a set of candidate targets; The sorting unit is used to perform similarity measurement matching between candidate targets in the candidate target set and real targets in the detection label based on the Top-N screening mechanism and Hungarian matching algorithm to obtain positive and negative sample allocation results; The modeling unit is used to model the positive sample prediction boundary and the real target in the positive and negative sample allocation results as two-dimensional Gaussian distributions respectively, calculate the normalized bulldozer distance on the predicted two-dimensional Gaussian distribution and the real two-dimensional Gaussian distribution, and obtain the boundary loss function. The update unit is used to optimize and train the target detection model based on backpropagation and update the parameters through the boundary loss function to obtain an improved target detection model.

[0047] The detection module is used to perform unknown parameter target detection on the time-frequency map based on the improved target detection model, and obtain the detection result; The inversion module is used to perform inversion processing on the intermediate frequency signal based on the detection results to obtain the inversion result.

[0048] It should be noted that the specific manner in which each module performs its operation in the apparatus described in the above embodiments has been described in detail in the embodiments of the method, and will not be elaborated here.

[0049] Example 3: Corresponding to the above method embodiments, this embodiment also provides a radar intra-pulse modulation signal waveform inversion device. The radar intra-pulse modulation signal waveform inversion device described below and the radar intra-pulse modulation signal waveform inversion method described above can be referred to each other.

[0050] Figure 2 This is a target image of a radar intra-pulse modulated signal waveform inversion device 800, illustrated according to an exemplary embodiment. For example... Figure 2 As shown, the radar intra-pulse modulated signal waveform inversion device 800 may include: a processor 801 and a memory 802. The radar intra-pulse modulated signal waveform inversion device 800 may also include one or more of the following: a multimedia component 803, an I / O interface 804, and a communication component 805.

[0051] The processor 801 controls the overall operation of the radar intra-pulse modulated signal waveform inversion device 800 to complete all or part of the steps in the aforementioned radar intra-pulse modulated signal waveform inversion method. The memory 802 stores various types of data to support the operation of the radar intra-pulse modulated signal waveform inversion device 800. This data may include, for example, instructions for any application or method operating on the radar intra-pulse modulated signal waveform inversion device 800, as well as application-related data such as contact data, sent and received messages, images, audio, and video. The memory 802 can be implemented using any type of volatile or non-volatile storage device or a combination thereof, such as Static Random Access Memory (SRAM), Electrically Erasable Programmable Read-Only Memory (EEPROM), Erasable Programmable Read-Only Memory (EPROM), Programmable Read-Only Memory (PROM), Read-Only Memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The multimedia component 803 may include a screen and an audio component. The screen may be, for example, a touchscreen, and the audio component is used to output and / or input audio signals. For example, the audio component may include a microphone for receiving external audio signals. The received audio signals may be further stored in the memory 802 or transmitted via the communication component 805. The audio component also includes at least one speaker for outputting audio signals. I / O interface 804 provides an interface between processor 801 and other interface modules, such as keyboards, mice, and buttons. These buttons can be virtual or physical. Communication component 805 is used for wired or wireless communication between the radar intra-pulse modulation signal waveform inversion device 800 and other devices. Wireless communication includes, for example, Wi-Fi, Bluetooth, Near Field Communication (NFC), 2G, 3G, or 4G, or a combination thereof. Therefore, the corresponding communication component 805 may include a Wi-Fi module, a Bluetooth module, or an NFC module.

[0052] In an exemplary embodiment, the radar intra-pulse modulated signal waveform inversion device 800 may be implemented by one or more application-specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field-programmable gate arrays (FPGAs), controllers, microcontrollers, microprocessors, or other electronic components to perform the radar intra-pulse modulated signal waveform inversion method described above.

[0053] Example 4: Corresponding to the above method embodiments, this embodiment also provides a medium. The medium described below can be referred to in correspondence with the radar intra-pulse modulation signal waveform inversion method described above.

[0054] A medium storing a computer program, which, when executed by a processor, implements the steps of the radar intra-pulse modulation signal waveform inversion method described in the above method embodiments.

[0055] The medium can specifically be any medium capable of storing program code, such as a USB flash drive, external hard drive, read-only memory (ROM), random access memory (RAM), magnetic disk, or optical disk.

[0056] The above description is merely a preferred embodiment of the present invention and is not intended to limit the invention. Various modifications and variations can be made to the present invention by those skilled in the art. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

[0057] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be determined by the scope of the claims.

Claims

1. A method for inverting the waveform of an intra-pulse modulated radar signal, characterized in that, include: Acquire signal parameters, including sampling rate, frequency modulation slope, and bandwidth; An intermediate frequency (IF) signal is generated from the signal parameters according to the signal generation function, and a time-frequency diagram is generated by performing a short-time Fourier transform on the IF signal. The time range and frequency range are calculated based on the waveform parameters of the intermediate frequency signal. The detection tag is obtained by normalizing the time range and frequency range to the coordinate range of the time-frequency graph. Based on the target detection model, the time-frequency map is trained using the detection labels, and positive and negative samples are assigned based on the Hungarian matching strategy to obtain an improved target detection model. The improved target detection model is used to perform target detection with unknown parameters on the time-frequency map to obtain the detection results; Based on the detection results, the intermediate frequency signal is inverted to obtain the inversion result.

2. The radar intra-pulse modulation signal waveform inversion method according to claim 1, characterized in that, An intermediate frequency (IF) signal is generated from the signal parameters according to the signal generation function. A time-frequency diagram is generated by performing a short-time Fourier transform on the IF signal, including: The signal parameters are generated according to the signal generation function to obtain the intermediate frequency signal; The window length and overlap length are determined based on the characteristics of the intermediate frequency signal. The intermediate frequency signal is then segmented using the window length and overlap length, and each segment is windowed to obtain windowed segments. Perform a Fast Fourier Transform on the windowed segment to obtain the segment spectrum; The frequency spectrum of the segments is arranged in chronological order to generate a time-frequency diagram.

3. The radar intra-pulse modulation signal waveform inversion method according to claim 2, characterized in that, The signal generation function includes a linear frequency modulated waveform function, a pulse waveform function, and a discrete frequency encoded waveform generation function. The signal parameters are generated according to the signal generation function to obtain an intermediate frequency signal, including: A linear frequency modulated signal is generated based on the sampling rate, the frequency modulation slope, and the bandwidth according to the linear frequency modulated waveform function. A simple pulse signal is generated based on the sampling rate and the frequency modulation slope according to the pulse waveform function; A discrete frequency encoded signal is generated based on the sampling rate and the bandwidth using a discrete frequency encoded waveform generation function; An intermediate frequency signal is obtained by fusing the linear frequency modulated signal, the simple pulse signal, and the discrete frequency encoded signal.

4. The radar intra-pulse modulation signal waveform inversion method according to claim 1, characterized in that, Based on the target detection model, the time-frequency map is trained using the detection labels, and positive and negative samples are assigned using a Hungarian matching strategy to obtain an improved target detection model, including: Based on the target detection model, the time-frequency map features are extracted to obtain a set of candidate targets; Based on the Top-N screening mechanism and the Hungarian matching algorithm, the candidate targets in the candidate target set are matched with the real targets in the detection labels to obtain the positive and negative sample allocation results. In the positive and negative sample allocation results, the predicted boundary of the positive sample and the real target are modeled as two-dimensional Gaussian distributions respectively. The normalized bulldozer distance is calculated on the predicted two-dimensional Gaussian distribution and the real two-dimensional Gaussian distribution to obtain the boundary loss function. Based on backpropagation, the target detection model is optimized and trained using the boundary loss function, and the parameters are updated to obtain an improved target detection model.

5. The radar intra-pulse modulation signal waveform inversion method according to claim 4, characterized in that, The positive sample prediction boundary and the true target in the positive and negative sample allocation results are modeled as two-dimensional Gaussian distributions, respectively. The normalized bulldozer distance is calculated on the predicted and true two-dimensional Gaussian distributions to obtain the boundary loss function, including: The positive sample prediction boundary and the real target in the positive and negative sample allocation results are modeled as two-dimensional Gaussian distributions respectively to obtain the predicted two-dimensional Gaussian distribution and the real two-dimensional Gaussian distribution; The distance between the predicted two-dimensional Gaussian distribution and the actual two-dimensional Gaussian distribution is calculated to obtain the bulldozer distance; The bulldozer distance is normalized to obtain the normalized bulldozer distance; Based on the normalized bulldozer distance, the similarity between the Gaussian distributions corresponding to the predicted boundary and the true boundary is measured to obtain the boundary loss function.

6. The radar intra-pulse modulation signal waveform inversion method according to claim 1, characterized in that, Based on the detection results, the intermediate frequency signal is inverted to obtain the inversion result, including: Based on the target label in the detection results, the position information of the target in the time-frequency map is extracted to obtain the target boundary parameters; The normalized coordinates output from the target detection model are converted into actual time and frequency values. The time span and frequency variation range of the target in the time-frequency diagram are calculated using the target boundary parameters, actual time and frequency values ​​to obtain the target modulation parameters. Based on the output of the target detection model, the target is extracted using discrete frequency encoded signals, and the discrete frequency encoded signals are sorted in descending order according to their vertical center coordinates to calculate the frequency encoded sequence. The target modulation parameters and the frequency coding sequence are fused together to output the inversion result.

7. A radar intra-pulse modulated signal waveform inversion device, characterized in that, include: The acquisition module is used to acquire signal parameters, including sampling rate, frequency modulation slope, and bandwidth. The generation module is used to generate an intermediate frequency signal based on the signal parameters according to the signal generation function, and to generate a time-frequency diagram by performing a short-time Fourier transform on the intermediate frequency signal. The calculation module is used to calculate the time range and frequency range based on the waveform parameters of the intermediate frequency signal, and obtain the detection tag by normalizing the time range and frequency range to the coordinate range of the time-frequency graph. The training module is used to train the time-frequency map based on the target detection model using the detection labels, and to perform positive and negative sample allocation based on the Hungarian matching strategy to obtain an improved target detection model. The detection module is used to perform unknown parameter target detection on the time-frequency map based on the improved target detection model, and obtain the detection result; The inversion module is used to perform inversion processing on the intermediate frequency signal based on the detection results to obtain the inversion result.

8. The radar intra-pulse modulation signal waveform inversion device according to claim 7, characterized in that, The generation module includes: The generation unit is used to generate the signal parameters according to the signal generation function to obtain the intermediate frequency signal; The segmentation unit is used to determine the window length and overlap length according to the characteristics of the intermediate frequency signal, segment the intermediate frequency signal into segments using the window length and overlap length, and perform windowing processing on each segment to obtain windowed segments; The transformation unit is used to perform a fast Fourier transform on the windowed segment to obtain the segment spectrum; The arrangement unit is used to arrange the spectrum of the segments in time order to generate a time-frequency diagram.

9. The radar intra-pulse modulated signal waveform inversion device according to claim 8, characterized in that, The signal generation function includes a linear frequency modulated waveform function, a pulse waveform function, and a discrete frequency coded waveform generation function. The generation unit includes: The first subunit is used to generate a linear frequency modulated signal based on the sampling rate, the frequency modulation slope, and the bandwidth according to the linear frequency modulated waveform function; The second subunit is used to generate a simple pulse signal based on the sampling rate and the frequency modulation slope according to the pulse waveform function; The third subunit is used to generate a discrete frequency encoded signal based on the sampling rate and the bandwidth according to the discrete frequency encoded waveform generation function; The fusion subunit is used to fuse the linear frequency modulated signal, the simple pulse signal, and the discrete frequency coded signal to obtain an intermediate frequency signal.

10. The radar intra-pulse modulated signal waveform inversion device according to claim 7, characterized in that, The training module includes: The extraction unit is used to extract features from the time-frequency map based on the target detection model to obtain a set of candidate targets; The sorting unit is used to perform similarity measurement matching between candidate targets in the candidate target set and real targets in the detection label based on the Top-N screening mechanism and Hungarian matching algorithm to obtain positive and negative sample allocation results; The modeling unit is used to model the positive sample prediction boundary and the real target in the positive and negative sample allocation results as two-dimensional Gaussian distributions respectively, calculate the normalized bulldozer distance on the predicted two-dimensional Gaussian distribution and the real two-dimensional Gaussian distribution, and obtain the boundary loss function. The update unit is used to optimize and train the target detection model based on backpropagation and update the parameters through the boundary loss function to obtain an improved target detection model.

Images