Signal sorting method, device, apparatus, and storage medium

Abstract

The application relates to a signal sorting method and device, computer equipment, a storage medium and a computer program product. The method comprises the following steps: acquiring an initial pulse signal; preprocessing the initial pulse signal to obtain an initial pulse repetition interval of the initial pulse signal; establishing a corresponding polar coordinate image according to a preset relationship between an initial pulse description word of the initial pulse signal and a polar coordinate system and a preset relationship between the initial pulse repetition interval and the polar coordinate system; inputting the polar coordinate image into a pulse classification model to obtain a target center point of the initial pulse signal; and determining a signal source of the initial pulse signal according to the target center point. The method can identify the emission source of a radar pulse signal.

Description

Technical Field

[0001] This application relates to the field of radar technology, and in particular to a signal sorting method, apparatus, computer equipment, storage medium, and computer program product. Background Technology

[0002] With the increasing number and variety of radar systems, different radar types have different uses, carriers, and levels of danger. This diversification of radar types is gradually increasing the complexity of the environment. Radar signal processing is a crucial step in electronic intelligence and electronic support systems operating in complex environments.

[0003] Therefore, how to sort radar signals to obtain the sources of radar signal transmission is an urgent problem to be solved. Summary of the Invention

[0004] Therefore, it is necessary to provide a signal sorting method, apparatus, computer equipment, computer-readable storage medium, and computer program product capable of identifying the source of radar pulse signals, in order to address the aforementioned technical problems.

[0005] In a first aspect, this application provides a signal sorting method, the method comprising: acquiring an initial pulse signal; preprocessing the initial pulse signal to obtain an initial pulse repetition interval of the initial pulse signal; establishing a corresponding polar coordinate image based on a preset relationship between the initial pulse descriptor of the initial pulse signal and a polar coordinate system, and a preset relationship between the initial pulse repetition interval and a polar coordinate system; inputting the polar coordinate image into a pulse classification model to obtain a target center point of the initial pulse signal; and determining the signal source of the initial pulse signal based on the target center point.

[0006] In one embodiment, the preprocessing of the initial pulse signal to obtain the initial pulse repetition interval of the initial pulse signal includes: interpolating the initial pulse signal; sampling and calculating the interpolated initial pulse signal to obtain a corresponding first pulse repetition frequency; and if the threshold value of the first pulse repetition frequency is greater than a target value, obtaining the initial pulse repetition interval based on the first pulse repetition frequency.

[0007] In one embodiment, sampling and calculating the interpolated initial pulse signal to obtain the corresponding first pulse repetition frequency includes: uniformly sampling and performing fast Fourier transform calculation on the interpolated initial pulse signal to obtain the corresponding first pulse repetition frequency.

[0008] In one embodiment, the preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system includes: the initial angle of arrival of the initial pulse descriptor corresponds to the polar angle of the polar coordinate system, the initial carrier of the initial pulse descriptor corresponds to the color of the polar coordinate system, and the initial pulse width of the initial pulse descriptor corresponds to the radius of the polar coordinate system; the preset relationship between the initial pulse repetition interval and the polar coordinate system includes: the initial pulse repetition interval corresponds to the polar radius of the polar coordinate system.

[0009] In one embodiment, the establishment of the pulse classification model includes: acquiring training pulse signals and labels; the labels being information corresponding to a reference signal source of the training pulse signals; preprocessing the training pulse signals to obtain the training pulse repetition interval of the training pulse signals; establishing a corresponding polar coordinate training image based on a preset relationship between the training pulse descriptor of the training pulse signals and the polar coordinate system, and a preset relationship between the training pulse repetition interval and the polar coordinate system; inputting the polar coordinate training image and the labels into an initial pulse classification model to obtain the predicted target center point of the training pulse signals; determining the predicted signal source of the training pulse signals based on the predicted target center point; obtaining the target loss based on the information of the predicted signal source, the labels, and the target loss function; and adjusting the parameters of the initial pulse classification model based on the target loss to obtain the pulse classification model.

[0010] In one embodiment, the target loss function includes a category prediction loss function, a width prediction loss function, and a location prediction loss function; obtaining the target loss based on the information of the prediction signal source, the label, and the target loss function includes: obtaining the target loss based on the information of the prediction signal source, the label, the category prediction loss function, the width prediction loss function, and the location prediction loss function.

[0011] In one embodiment, the tag includes a reference carrier, a reference pulse width, a reference angle of arrival, and a reference pulse repetition interval of the reference signal source; obtaining the target loss based on the information of the predicted signal source, the tag, the category prediction loss function, the width prediction loss function, and the position prediction loss function includes: obtaining a first loss based on the predicted carrier of the predicted signal source, the reference carrier of the reference signal source, and the category prediction loss function; obtaining a second loss based on the predicted pulse width of the predicted signal source, the reference pulse width of the reference signal source, and the width prediction loss function; obtaining a third loss based on the predicted angle of arrival, the predicted pulse repetition interval, and the scaling scale of the predicted signal source, the reference angle of arrival and the reference pulse repetition interval of the reference signal source, and the position prediction loss function; and obtaining the target loss based on the first loss, the second loss, and the third loss.

[0012] Secondly, this application also provides a signal sorting device, the device comprising: an acquisition module for acquiring an initial pulse signal; a preprocessing module for preprocessing the initial pulse signal to obtain an initial pulse repetition interval of the initial pulse signal; an image building module for building a corresponding polar coordinate image based on a preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system, and a preset relationship between the initial pulse repetition interval and the polar coordinate system; a center point acquisition module for inputting the polar coordinate image into a pulse classification model to obtain a target center point of the initial pulse signal; and a signal source determination module for determining the signal source of the initial pulse signal based on the target center point.

[0013] Thirdly, this application also provides a computer device. The computer device includes a memory and a processor, the memory storing a computer program, and the processor executing the computer program to implement the steps of the method described above.

[0014] Fourthly, this application also provides a computer-readable storage medium. The computer-readable storage medium stores a computer program thereon, which, when executed by a processor, implements the steps of the above-described method.

[0015] The aforementioned signal sorting method, apparatus, computer equipment, and storage medium establish a corresponding polar coordinate image based on a preset relationship between the initial pulse descriptor and the initial pulse repetition interval of the initial pulse signal and the polar coordinate system. Based on the periodicity of the polar coordinate image, initial pulse signals with the same initial pulse repetition interval can be concentrated in a specific region. This does not increase complexity with the increase in observation time and the number of received initial pulse signals. Furthermore, with the increase in observation time, the accuracy of the pulse classification model in obtaining the corresponding target center point can be improved. The polar coordinate image is input into the pulse classification model to obtain the target center point of the initial pulse signal; the signal source of the initial pulse signal is then determined based on the target center point, i.e., the transmission source of the initial pulse signal. In other words, the pulse classification model enables the identification of the signal sources of different initial pulse signals. Attached Figure Description

[0016] Figure 1 This is a diagram illustrating the application environment of the signal sorting method in one embodiment;

[0017] Figure 2 This is a flowchart illustrating a signal sorting method in one embodiment;

[0018] Figure 3 This is a schematic diagram of the signal sorting process in another embodiment;

[0019] Figure 4 This is a flowchart illustrating the signal sorting method in another embodiment;

[0020] Figure 5 This is a schematic diagram of a polar coordinate image in one embodiment;

[0021] Figure 6 This is a schematic diagram of a polar coordinate image in another embodiment;

[0022] Figure 7 This is a schematic diagram of a polar coordinate image in another embodiment;

[0023] Figure 8 This is a schematic diagram of a polar coordinate image in another embodiment;

[0024] Figure 9 This is a schematic diagram of a polar coordinate image in another embodiment;

[0025] Figure 10 This is a schematic diagram of a polar coordinate image in another embodiment;

[0026] Figure 11 This is a schematic diagram of the target polar coordinate image in another embodiment;

[0027] Figure 12 This is a schematic diagram of the target polar coordinate image in another embodiment;

[0028] Figure 13 This is a structural block diagram of a signal sorting device in one embodiment;

[0029] Figure 14 This is an internal structural diagram of a computer device in one embodiment. Detailed Implementation

[0030] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0031] The signal sorting method provided in this application embodiment can be applied to, for example... Figure 1 In the application environment shown, terminal 102 communicates with server 104 via a network. A data storage system can store the data that server 104 needs to process. The data storage system can be integrated onto server 104 or placed in the cloud or on other network servers. Terminal 102 acquires an initial pulse signal; preprocesses the initial pulse signal to obtain the initial pulse repetition interval; establishes a corresponding polar coordinate image based on the preset relationship between the initial pulse descriptor and the polar coordinate system, and the preset relationship between the initial pulse repetition interval and the polar coordinate system; inputs the polar coordinate image into a pulse classification model to obtain the target center point of the initial pulse signal; and determines the signal source of the initial pulse signal based on the target center point. Terminal 102 can be, but is not limited to, various personal computers, laptops, smartphones, tablets, IoT devices, and portable wearable devices. Server 104 can be implemented using a standalone server or a server cluster composed of multiple servers.

[0032] In one embodiment, such as Figure 2 As shown, a signal sorting method is provided, which can be applied to... Figure 1 Taking the terminal in the example, the explanation includes the following steps:

[0033] Step 202: Obtain the initial pulse signal.

[0034] The initial pulse signal is a pulse signal received from multiple unknown signal sources. Optionally, these unknown signal sources are unknown radar radiation sources. Combined with... Figure 3 To clarify, the initial pulse signal includes an initial pulse signal received from at least one unknown signal source. In other words, the initial pulse signal can include multiple pulse signals, and the sources of these multiple pulse signals can be different.

[0035] Optionally, the terminal acquires the received initial pulse signal.

[0036] Step 204: Preprocess the initial pulse signal to obtain the initial pulse repetition interval of the initial pulse signal.

[0037] The preprocessing includes converting the original pulse signal into a spectrum. The initial pulse repetition interval is the time interval between two initial pulse signals emitted by the same unknown signal source, obtained from the spectrum. Figure 3 The potential pulse interval of the initial pulse signal in the signal.

[0038] Specifically, the terminal preprocesses the initial pulse signal to obtain the spectrum of the converted original pulse signal, and obtains the initial pulse repetition interval based on the spectrum.

[0039] Optionally, the initial pulse signal is preprocessed to obtain the initial pulse repetition interval of the initial pulse signal, including: interpolating the initial pulse signal; sampling and calculating the interpolated initial pulse signal to obtain the corresponding first pulse repetition frequency; if the threshold value of the first pulse repetition frequency is greater than the target value, the initial pulse repetition interval is obtained according to the first pulse repetition frequency.

[0040] Interpolation can be understood as interpolating the arrival time of the initial pulse signal to obtain a mapping diagram describing the arrival time of the initial pulse signal and the amplitude of the initial pulse signal. When the initial pulse signal is received, the amplitude corresponding to the arrival time of the initial pulse signal is the peak value, and when the initial pulse signal is not received, the amplitude corresponding to the arrival time of the initial pulse signal is 0.

[0041] Specifically, the terminal performs interpolation modeling on the arrival time of the initial pulse signal. Figure 4 The original signal is interpolated to obtain a mapping diagram describing the arrival time and amplitude of the initial pulse signal. The interpolated initial pulse signal is sampled and calculated to obtain the corresponding first pulse repetition frequency. If the threshold value of the first pulse repetition frequency is greater than the target value, the initial pulse repetition interval is obtained according to the first pulse repetition frequency.

[0042] Optionally, the initial pulse signal after interpolation is sampled and calculated to obtain the corresponding first pulse repetition frequency, including: uniformly sampling and fast Fourier transforming the initial pulse signal after interpolation to obtain the corresponding first pulse repetition frequency.

[0043] Specifically, the terminal performs uniform sampling on the interpolated initial pulse signal. Figure 4 The interpolated signal is sampled uniformly as shown, with a sampling frequency F. s Satisfy F s ≥2F maxWhere Fmax is the upper limit of the pulse repetition frequency in the signal environment. This can be understood as the density of the current pulse signal. The terminal performs a Fast Fourier Transform (FFT) on the sampled initial pulse signal. Figure 4 The FFT operation is performed on the sampled data as shown to obtain the corresponding spectrum, and the corresponding first pulse repetition frequency is obtained based on the spectrum.

[0044] The target value is a pre-set specific value. It should be noted that the target value in this application can be set according to actual needs and updated as needs change. Therefore, this application does not limit the specific numerical value of the target value. Threshold value D th The calculation formula (1) for (i) is shown below:

[0045]

[0046] In the above formula, α and β are adjustable parameters, PRF i Let be the repetition frequency of the i-th first pulse. It should be noted that during the Fast Fourier Transform (FFT), only the pulse repetition frequency corresponding to the initial pulse signal has relatively high energy. Therefore, through FFT, the first pulse repetition frequency corresponding to the potential initial pulse signal can be found. By setting a target value, the first pulse repetition frequency is filtered. If the energy corresponding to the first pulse repetition frequency is greater than the target value, then the reciprocal of the first pulse repetition frequency is set as the initial pulse repetition interval (PRI), i.e. Figure 4 The PRI is saved as shown. If the energy corresponding to the first pulse repetition frequency is less than the target value, then proceed. Figure 4 The example shown is to discard PRIs that do not meet the criteria.

[0047] Therefore, the above formula (1) is used to calculate the energy of each first pulse repetition frequency. Optionally, the adjustable parameters are α = 0.5 and β = 1000.

[0048] Step 206: Establish the corresponding polar coordinate image based on the preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system, and the preset relationship between the initial pulse repetition interval and the polar coordinate system.

[0049] Among them, the initial pulse description word (PDW) is the initial pulse signal description word formed by numbers with fixed length, fixed grid and positioning meanings, obtained after intra-pulse analysis of the initial pulse signal.

[0050] Optionally, the preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system includes: the polar angle of the polar coordinate system corresponding to the initial angle of arrival of the initial pulse descriptor, the color of the polar coordinates of the polar coordinate system corresponding to the initial carrier of the initial pulse descriptor, and the radius of the polar coordinates of the polar coordinate system corresponding to the initial pulse width of the initial pulse descriptor; the preset relationship between the initial pulse repetition interval and the polar coordinate system includes: the polar radius of the polar coordinate system corresponding to the initial pulse repetition interval.

[0051] Upon receiving the initial pulse signal, the initial pulse signal is analyzed by machine measurement to obtain the initial pulse descriptor for each initial pulse signal. The initial pulse descriptor includes: initial angle of arrival (DOA), initial carrier frequency (CF), and initial pulse width (PW). The initial angle of arrival (DOA) is the direction of arrival of the initial pulse signal as measured by the machine, i.e., the direction angle at which each initial pulse signal arrives at the array reference element, also known as the direction of arrival. The initial carrier frequency (CF) is the transmission frequency of the initial pulse signal as measured by the machine. The initial pulse width (PW) is the duration of the transmitted initial pulse signal as measured by the machine.

[0052] The initial arrival angle of the aforementioned initial pulse descriptor corresponds to the polar angle of the polar coordinate system, including: the terminal pre-establishes a one-to-one correspondence between the angular coordinate system corresponding to the initial arrival angle of the pulse descriptor and the angular coordinate system corresponding to the polar angle of the polar coordinate system.

[0053] The color of the polar coordinates in the polar coordinate system corresponding to the initial carrier of the aforementioned initial pulse descriptor includes: the terminal pre-establishing a one-to-one correspondence between the frequencies of different initial carriers and the colors of the polar coordinates in the polar coordinate system. For example, the terminal pre-establishes a correspondence between carrier frequencies of 3-30MHz and the color of the polar coordinates in the polar coordinate image as yellow. When the initial carrier frequency of the initial pulse signal is 25MHz, the terminal sets the color of the polar coordinates in the polar coordinate image corresponding to the initial pulse signal to yellow.

[0054] The initial pulse width of the aforementioned initial pulse descriptor corresponds to the radius of the polar coordinates in the polar coordinate system, including: the terminal pre-constructs a corresponding mapping relationship between different initial pulse widths and the radius of the polar coordinates in the polar coordinate system.

[0055] The aforementioned initial pulse repetition interval corresponds to the polar radius in the polar coordinate system, including: the terminal pre-establishes a one-to-one correspondence between the length of different initial pulse repetition intervals and the length of the polar radius in the polar coordinate image.

[0056] The terminal constructs the initial angle of arrival, initial carrier, initial pulse width, and initial pulse repetition interval of the initial pulse signal according to the aforementioned preset relationship, as follows: Figures 5-10 The polar coordinate image shown is Figure 4 The initial pulse signal is mapped onto polar coordinates. It should be noted that... Figures 5-10 The black dot in the diagram represents the polar coordinates. In practice, the polar coordinates may have different colors, and the specific color of the polar coordinates is determined based on the carrier frequency of the initial pulse signal.

[0057] Step 208: Input the polar coordinate image into the pulse classification model to obtain the target center point of the initial pulse signal.

[0058] Among them, the pulse classification model is a target recognition and classification network, which is a pre-trained model used to process the input polar coordinate image to obtain the corresponding target center point.

[0059] The terminal inputs the polar coordinate image obtained from the initial pulse signal into the pulse classification model. The pulse classification model processes the polar coordinate image, as follows: Figure 3 The pulse segmentation shown yields a target polar coordinate image of the initial pulse signal. The target center point is then obtained from this polar coordinate image. The target polar coordinate image is shown below. Figure 11 and Figure 12 As shown, where, Figure 11 Radar_1 and Figure 12 Radar_2 in the diagram represents the target center point. It should be noted that... Figure 11 and Figure 12 The polar coordinates in the target polar coordinate image may have different colors during the specific implementation process, and the specific color of the polar coordinates is determined according to the target carrier frequency of the initial pulse signal.

[0060] Step 210: Determine the signal source of the initial pulse signal based on the target center point.

[0061] The terminal obtains the target angle of arrival, target carrier, target pulse width, and target pulse repetition interval based on the polar angle, color, radius, and radius of the target center point in the target polar coordinate image. It then determines the signal source based on these parameters, i.e., it determines the signal source of the initial pulse signal based on the target center point. In other words, it obtains... Figure 3 The sorting results are shown.

[0062] It should be noted that the initial angle of arrival, initial carrier, initial pulse width, and initial pulse repetition interval of the initial pulse signal are measured values ​​of the initial pulse signal and cannot accurately describe the initial pulse signal. However, the target angle of arrival, target carrier, target pulse width, and target pulse repetition interval of the initial pulse signal obtained after pulse ripple model processing are more accurate pulse description words for the initial pulse signal.

[0063] The aforementioned signal sorting method establishes a corresponding polar coordinate image based on a preset relationship between the initial pulse descriptor and the initial pulse repetition interval of the initial pulse signal and the polar coordinate system. Based on the periodicity of the polar coordinate image, initial pulse signals with the same initial pulse repetition interval can be concentrated in a specific region, without increasing complexity with the increase in observation time and the number of received initial pulse signals; furthermore, with the increase in observation time, the accuracy of the pulse classification model in obtaining the corresponding target center point can be improved. The polar coordinate image is input into the pulse classification model to obtain the target center point of the initial pulse signal; the signal source of the initial pulse signal is then determined based on the target center point, i.e., the emission source of the initial pulse signal. The pulse classification model learns the features of the aforementioned polar coordinate image through its training process, enabling accurate identification of the signal source of different initial pulse signals using the pulse classification model.

[0064] In one embodiment, the pulse classification model is established in the following way: Figure 4 The training-specific target recognition and sorting network includes: acquiring training pulse signals and labels; labels being information corresponding to the reference signal source of the training pulse signals; preprocessing the training pulse signals to obtain the training pulse repetition interval; establishing corresponding polar coordinate training images based on the preset relationship between the training pulse descriptor and the polar coordinate system, and the preset relationship between the training pulse repetition interval and the polar coordinate system; inputting the polar coordinate training images and labels into the initial pulse classification model to obtain the predicted target center point of the training pulse signals; determining the predicted signal source of the training pulse signals based on the predicted target center point; obtaining the target loss based on the information of the predicted signal source, labels, and target loss function; and adjusting the parameters of the initial pulse classification model based on the target loss to obtain the pulse classification model.

[0065] In the pulse classification model establishment process, multiple known reference signal sources transmit pulse signals. The label corresponds to the information of the pulse signal transmitted by the known reference signal source. The training pulse signal is the pulse signal received by the receiver from the aforementioned known reference signal sources. Optionally, the aforementioned known reference signal sources are known radar radiation sources. It should be noted that the training pulse signal includes a pulse signal transmitted by at least one known reference signal source; that is, the training pulse signal can include multiple pulse signals, and the transmitting sources of these multiple pulse signals can be different. The training pulse repetition interval is the time interval between two training pulse signals emitted by the same unknown signal source, obtained from the spectrum training diagram. The aforementioned spectrum training diagram is the spectrum training diagram obtained after preprocessing the training pulse signals. The training pulse description word (PDW) is a training pulse signal description word formed by numbers with fixed length, fixed interval, and location meaning, obtained after intra-pulse analysis of the training pulse signal. Regarding the training pulse description word, when a training pulse signal is received, the training pulse signal is analyzed by machine measurement to obtain the training pulse description word for each training pulse signal. The training pulse descriptor includes: training angle of arrival (DOA), training carrier, and training pulse width. The training angle of arrival (DOA) is the direction of arrival of the training pulse signal measured by the machine, i.e., the direction angle at which each training pulse signal arrives at the array reference element, or simply the direction of arrival. The training carrier (RF) is the transmission frequency of the training pulse signal measured by the machine. The training pulse width (PW) is the duration of the transmitted training pulse signal measured by the machine. The initial pulse classification model is a target recognition classification network, a pre-set model used to learn and process the input polar coordinate training image and labels to obtain the corresponding training target center point. The predicted signal source is the transmission source of the training pulse signal predicted based on the training pulse signal and the initial pulse classification model. The target loss function estimates the degree of inconsistency between the information of the predicted signal source corresponding to the training pulse signal and the label corresponding to the training pulse signal, and is used to adjust the parameters of the initial pulse classification model.

[0066] The terminal acquires the received training pulse signal and its corresponding tag. The terminal preprocesses the training pulse signal to obtain a spectrum training diagram of the converted training pulse signal, and obtains the training pulse repetition interval based on the spectrum training diagram.

[0067] The preset relationship between the training pulse descriptor of the training pulse signal and the polar coordinate system includes: the polar angle of the polar coordinate system corresponding to the training arrival angle of the training pulse descriptor, the color of the polar coordinate system corresponding to the training carrier of the training pulse descriptor, and the radius of the polar coordinate system corresponding to the training pulse width of the training pulse descriptor; the preset relationship between the training pulse repetition interval and the polar coordinate system includes: the polar radius of the polar coordinate system corresponding to the training pulse repetition interval.

[0068] The terminal constructs the training pulse signal's pulse training descriptor, including the specific values ​​of the training angle of arrival, training carrier, training pulse width, and training pulse repetition interval, according to the aforementioned preset relationship, as follows: Figures 5-10 The polar coordinate training image shown.

[0069] The terminal inputs the polar coordinate training image obtained from the training pulse signal into the initial pulse classification model. The initial pulse classification model processes the polar coordinate training image to obtain the target polar coordinate training image of the training pulse signal. Based on the above target polar coordinate training image, the predicted target center point is obtained.

[0070] The terminal obtains the predicted angle of arrival, predicted carrier, predicted pulse width, and predicted pulse repetition interval based on the polar angle, color, radius, and radius of the predicted target center point in the target polar coordinate training image. It then determines the predicted signal source based on these parameters, i.e., the predicted signal source for the training pulse signal based on the predicted target center point.

[0071] The terminal obtains the target loss based on the information of the predicted signal source, the label, and the target loss function. The information of the predicted signal source includes the predicted angle of arrival, predicted carrier, predicted pulse width, and predicted pulse repetition interval. The label includes the reference angle of arrival, reference carrier, reference pulse width, and reference pulse repetition interval of the reference signal source. It is understood that there is a correspondence between the reference signal source and the predicted signal source; the reference signal source is a known signal source that transmits radar pulse signals, which are received by the receiver. The radar pulse signals transmitted by the known signal source received by the receiver are the training pulse signals. The terminal predicts the corresponding transmitting source based on the training pulse signals, i.e., the predicted signal source. The terminal compares the target loss with the preset loss and adjusts the parameters of the initial pulse classification model based on the comparison results to obtain the pulse classification model. It should be noted that the preset loss is a pre-set standard for measuring the accuracy of the predicted signal source. This standard can be updated and adjusted according to actual conditions, and this application does not limit the specific value of the preset loss.

[0072] The above signal sorting method trains an initial pulse classification model by training pulse signals, labels, and a target loss function to obtain a pulse classification model, which makes the pulse classification model more accurate in classifying the signal source of pulse signals.

[0073] In one embodiment, the target loss function includes a category prediction loss function, a width prediction loss function, and a location prediction loss function; the target loss is obtained based on the information of the predicted signal source, the label, and the target loss function, including: obtaining the target loss based on the information of the predicted signal source, the label, the category prediction loss function, the width prediction loss function, and the location prediction loss function.

[0074] Optionally, in one embodiment, the tag includes a reference carrier, a reference pulse width, a reference angle of arrival, and a reference pulse repetition interval of the reference signal source; obtaining the target loss based on the information of the predicted signal source, the tag, the category prediction loss function, the width prediction loss function, and the position prediction loss function includes: obtaining a first loss based on the predicted carrier of the predicted signal source, the reference carrier of the reference signal source, and the category prediction loss function; obtaining a second loss based on the predicted pulse width of the predicted signal source, the reference pulse width of the reference signal source, and the width prediction loss function; obtaining a third loss based on the predicted angle of arrival, the predicted pulse repetition interval, and the scaling scale of the predicted signal source, the reference angle of arrival, the reference pulse repetition interval, and the position prediction loss function of the reference signal source; and obtaining the target loss based on the first loss, the second loss, and the third loss.

[0075] It can be understood that the reference carrier, reference pulse width, reference angle of arrival, and reference pulse repetition interval of the pulse signal emitted by the reference signal source correspond to the predicted angle of arrival, predicted carrier, predicted pulse width, and predicted pulse repetition interval of the prediction signal source. These represent different data observed from different observation angles of the pulse signal emitted by the reference signal source. The calculation formula (2) for the category prediction loss function is as follows:

[0076]

[0077] L in formula (2) CThe first loss is α and β, which are hyperparameters. α is used to balance positive and negative samples, i.e., whether the predicted sample is a positive sample (radar signal pulse, i.e., training pulse signal) or a negative sample (background content). The more negative samples there are, the smaller the weight of α, thus reducing the influence of negative samples. β is used to balance easy and hard samples. For easy and hard samples, the more difficult the sample, i.e., the sample with the lower prediction probability value, the greater its weight in the category prediction loss function. N is the number of keypoints in the polar coordinate training image, i.e., the number of training pulse signal samples. The subscript xyc of the summation symbol represents all coordinate points on all heatmaps (i.e., the target polar coordinate training image). To predict the carrier, Y xyc (1-Y) is the reference carrier. xyc ) β In order to suppress 0 <Y xyc The loss function for negative samples with a value less than 1 is the loss function for those points near the center point of the Gaussian heatmap.

[0078] The formula (3) for calculating the width prediction loss function is as follows:

[0079]

[0080] L in formula (3) size This is the second loss. To predict the size of the polar radius in the polar coordinate training image corresponding to the pulse width, S pw This refers to the labeled dimensions in the polar coordinate image corresponding to the reference pulse width. Specifically, it is based on the predicted bounding box width and height values ​​in the polar coordinate training image. The size of the prediction box of the training pulse signal can be determined, and thus the prediction pulse width PW of the training pulse signal can be obtained, as shown in formula (4).

[0081]

[0082] The calculation formula (5) for the location prediction loss function is as follows:

[0083]

[0084] L in formula (5) local As the third loss, This represents the offset of the predicted target center point, where p is the coordinate of the reference target center point, and R is the scaling factor. These are the scaled approximate integer coordinates of the predicted target center point. The reference target center point is the one obtained by mapping the reference angle of arrival and reference pulse repetition interval to a polar coordinate image.

[0085] Specifically, based on the predicted offset value of the target center point in formula (5) The predicted angle of arrival and predicted pulse repetition interval of the training pulse signal can be obtained from formulas (6) and (7).

[0086]

[0087]

[0088] In Equation (6), DOA is the predicted angle of arrival, and in Equation (7), PRI is the predicted pulse repetition interval.

[0089] The formula for calculating the target loss (8) is as follows:

[0090] L total =L C +λ size L size +λ local L local (8)

[0091] L in formula (8) total The target loss is λ. size With λ local These are the weights of the loss function.

[0092] With L total As the function value decreases, the accuracy of recognizing descriptors of the training pulse signals also improves. The weights W of the CenterNet network, i.e., the initial pulse classification model, can be minimized by the loss function L. total To achieve this, the calculation formula (9) for the weight W of the initial pulse classification model is as follows:

[0093]

[0094] W in formula (9) last These are the weights calculated in the previous iteration.

[0095] Optionally, the weights W of the initial pulse classification model are determined using the stochastic gradient descent (SGD) algorithm. A transfer learning approach is employed, specifically cite{cantrell9}. Using the same network structure, the weights of the network model are pre-trained on the open-source dataset COCO. The aim is to enable the CenterNet feature extraction network to better extract the PDW features of the pulse signal for radar signal sorting. The transfer learning method is used to randomly initialize the network model parameters, ensuring a high similarity between the initial weights and the desired weights. Therefore, the training model converges faster, and the accuracy of radar pulse signal classification is higher.

[0096] like Figure 4 As shown, optionally, the data corresponding to the initial pulse signal of the signal source correctly identified by the pulse classification model can be stored as training data for training the initial pulse classification model.

[0097] It should be understood that although the steps in the flowcharts of the embodiments described above are shown sequentially according to the arrows, these steps are not necessarily executed in the order indicated by the arrows. Unless explicitly stated herein, there is no strict order restriction on the execution of these steps, and they can be executed in other orders. Moreover, at least some steps in the flowcharts of the embodiments described above may include multiple steps or multiple stages. These steps or stages are not necessarily completed at the same time, but can be executed at different times. The execution order of these steps or stages is not necessarily sequential, but can be performed alternately or in turn with other steps or at least some of the steps or stages of other steps.

[0098] Based on the same inventive concept, this application also provides a signal sorting device for implementing the signal sorting method described above. The solution provided by this device is similar to the implementation described in the above method; therefore, the specific limitations in one or more signal sorting device embodiments provided below can be found in the limitations of the signal sorting method described above, and will not be repeated here.

[0099] In one embodiment, such as Figure 13 As shown, a signal sorting device is provided, including: an acquisition module 100, a preprocessing module 200, an image building module 300, a center point acquisition module 400, and a signal source determination module 500, wherein:

[0100] Acquisition module 100 is used to acquire the initial pulse signal.

[0101] The preprocessing module 200 is used to preprocess the initial pulse signal to obtain the initial pulse repetition interval of the initial pulse signal.

[0102] The image building module 300 is used to build a corresponding polar coordinate image based on the preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system, and the preset relationship between the initial pulse repetition interval and the polar coordinate system.

[0103] The center point acquisition module 400 is used to input the polar coordinate image into the pulse classification model to obtain the target center point of the initial pulse signal.

[0104] The signal source determination module 500 is used to determine the signal source of the initial pulse signal based on the target center point.

[0105] In one embodiment, the preprocessing module includes: an interpolation module for interpolating the initial pulse signal; a first pulse repetition frequency acquisition module for sampling and calculating the interpolated initial pulse signal to obtain the corresponding first pulse repetition frequency; and a determination module for setting the first pulse repetition frequency as the initial pulse repetition interval if the threshold value of the first pulse repetition frequency is greater than the target value.

[0106] In one embodiment, the first pulse repetition frequency acquisition module includes: a first submodule, used to uniformly sample and perform fast Fourier transform calculation on the interpolated initial pulse signal to obtain the corresponding first pulse repetition frequency.

[0107] In one embodiment, the image establishment module includes: a first preset relationship module for the polar angle of the initial arrival angle of the initial pulse descriptor corresponding to the polar coordinate system, the color of the polar coordinate of the initial carrier of the initial pulse descriptor corresponding to the polar coordinate system, and the radius of the polar coordinate of the initial pulse width of the initial pulse descriptor corresponding to the polar coordinate system; and a second preset relationship module for the polar radius of the initial pulse repetition interval corresponding to the polar coordinate system.

[0108] In one embodiment, the system includes: a second acquisition module for acquiring training pulse signals and labels; the labels are information corresponding to a reference signal source of the training pulse signals; a second preprocessing module for preprocessing the training pulse signals to obtain the training pulse repetition interval of the training pulse signals; a second image building module for building a corresponding polar coordinate training image based on a preset relationship between the training pulse descriptor of the training pulse signals and the polar coordinate system, and a preset relationship between the training pulse repetition interval and the polar coordinate system; a second center point acquisition module for inputting the polar coordinate training image and labels into an initial pulse classification model to obtain the predicted target center point of the training pulse signals; a second signal source determination module for determining the predicted signal source of the training pulse signals based on the predicted target center point; and a loss adjustment module for obtaining a target loss based on the information of the predicted signal source, the labels, and the target loss function; and adjusting the parameters of the initial pulse classification model based on the target loss to obtain a pulse classification model.

[0109] In one embodiment, the loss adjustment module includes a loss calculation module, used to obtain the target loss based on the information of the predicted signal source, label, category prediction loss function, width prediction loss function and location prediction loss function.

[0110] In one embodiment, the loss adjustment module includes: a first loss determination module, configured to obtain a first loss based on the predicted carrier of the predicted signal source, the reference carrier of the reference signal source, and a category prediction loss function; a second loss determination module, configured to obtain a second loss based on the predicted pulse width of the predicted signal source, the reference pulse width of the reference signal source, and a width prediction loss function; a third loss determination module, configured to obtain a third loss based on the predicted angle of arrival, the predicted pulse repetition interval, and the scaling scale of the predicted signal source, the reference angle of arrival, the reference pulse repetition interval, and the position prediction loss function of the reference signal source; and a target loss determination module, configured to obtain a target loss based on the first loss, the second loss, and the third loss.

[0111] Each module in the aforementioned signal sorting device can be implemented entirely or partially through software, hardware, or a combination thereof. These modules can be embedded in the processor of a computer device in hardware form or independent of it, or stored in the memory of a computer device in software form, so that the processor can call and execute the operations corresponding to each module.

[0112] In one embodiment, a computer device is provided, which may be a terminal, and its internal structure diagram may be as follows: Figure 14 As shown, the computer device includes a processor, memory, communication interface, display screen, and input devices connected via a system bus. The processor provides computing and control capabilities. The memory includes non-volatile storage media and internal memory. The non-volatile storage media stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs stored in the non-volatile storage media. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, mobile cellular networks, NFC (Near Field Communication), or other technologies. When executed by the processor, the computer program implements a signal sorting method. The display screen can be an LCD screen or an e-ink screen. The input devices can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the computer device casing, or an external keyboard, touchpad, or mouse.

[0113] Those skilled in the art will understand that Figure 14 The structure shown is merely a block diagram of a portion of the structure related to the present application and does not constitute a limitation on the computer device to which the present application is applied. Specific computer devices may include more or fewer components than those shown in the figure, or combine certain components, or have different component arrangements.

[0114] In one embodiment, a computer device is provided, including a memory and a processor. The memory stores a computer program, and the processor executes the computer program to perform the following steps: acquiring an initial pulse signal; preprocessing the initial pulse signal to obtain an initial pulse repetition interval of the initial pulse signal; establishing a corresponding polar coordinate image based on a preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system, and a preset relationship between the initial pulse repetition interval and the polar coordinate system; inputting the polar coordinate image into a pulse classification model to obtain a target center point of the initial pulse signal; and determining the signal source of the initial pulse signal based on the target center point.

[0115] In one embodiment, the preprocessing of the initial pulse signal to obtain the initial pulse repetition interval when the processor executes the computer program includes: interpolating the initial pulse signal; sampling and calculating the interpolated initial pulse signal to obtain the corresponding first pulse repetition frequency; and setting the first pulse repetition frequency as the initial pulse repetition interval if the threshold value of the first pulse repetition frequency is greater than the target value.

[0116] In one embodiment, the sampling and calculation of the interpolated initial pulse signal to obtain the corresponding first pulse repetition frequency implemented by the processor executing the computer program includes: uniformly sampling and fast Fourier transform calculation of the interpolated initial pulse signal to obtain the corresponding first pulse repetition frequency.

[0117] In one embodiment, the preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system implemented by the processor when executing the computer program includes: the initial angle of arrival of the initial pulse descriptor corresponding to the polar angle of the polar coordinate system, the initial carrier of the initial pulse descriptor corresponding to the color of the polar coordinates of the polar coordinate system, and the initial pulse width of the initial pulse descriptor corresponding to the radius of the polar coordinates of the polar coordinate system; the preset relationship between the initial pulse repetition interval and the polar coordinate system includes: the initial pulse repetition interval corresponding to the polar radius of the polar coordinate system.

[0118] In one embodiment, the method for establishing a pulse classification model implemented by the processor when executing a computer program includes: acquiring training pulse signals and labels; the labels being information corresponding to the reference signal source of the training pulse signals; preprocessing the training pulse signals to obtain the training pulse repetition interval of the training pulse signals; establishing a corresponding polar coordinate training image based on a preset relationship between the training pulse descriptor of the training pulse signals and the polar coordinate system, and a preset relationship between the training pulse repetition interval and the polar coordinate system; inputting the polar coordinate training image and labels into an initial pulse classification model to obtain the predicted target center point of the training pulse signals; determining the predicted signal source of the training pulse signals based on the predicted target center point; obtaining the target loss based on the information of the predicted signal source, the labels, and the target loss function; and adjusting the parameters of the initial pulse classification model based on the target loss to obtain the pulse classification model.

[0119] In one embodiment, the target loss function implemented by the processor when executing the computer program includes a category prediction loss function, a width prediction loss function, and a location prediction loss function; obtaining the target loss based on the information of the predicted signal source, the label, and the target loss function includes: obtaining the target loss based on the information of the predicted signal source, the label, the category prediction loss function, the width prediction loss function, and the location prediction loss function.

[0120] In one embodiment, when the processor executes the computer program, the implemented tag includes a reference carrier, reference pulse width, reference angle of arrival, and reference pulse repetition interval of the reference signal source; the target loss is obtained based on the information of the predicted signal source, the tag, the category prediction loss function, the width prediction loss function, and the position prediction loss function, including: obtaining a first loss based on the predicted carrier of the predicted signal source, the reference carrier of the reference signal source, and the category prediction loss function; obtaining a second loss based on the predicted pulse width of the predicted signal source, the reference pulse width of the reference signal source, and the width prediction loss function; obtaining a third loss based on the predicted angle of arrival, the predicted pulse repetition interval, and the scaling scale of the predicted signal source, the reference angle of arrival and the reference pulse repetition interval of the reference signal source, and the position prediction loss function; and obtaining the target loss based on the first loss, the second loss, and the third loss.

[0121] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, it performs the following steps: acquiring an initial pulse signal; preprocessing the initial pulse signal to obtain an initial pulse repetition interval of the initial pulse signal; establishing a corresponding polar coordinate image based on a preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system, and a preset relationship between the initial pulse repetition interval and the polar coordinate system; inputting the polar coordinate image into a pulse classification model to obtain the target center point of the initial pulse signal; and determining the signal source of the initial pulse signal based on the target center point.

[0122] In one embodiment, the computer program, when executed by a processor, preprocesses an initial pulse signal to obtain an initial pulse repetition interval, including: interpolating the initial pulse signal; sampling and calculating the interpolated initial pulse signal to obtain a corresponding first pulse repetition frequency; and setting the first pulse repetition frequency as the initial pulse repetition interval if the threshold value of the first pulse repetition frequency is greater than a target value.

[0123] In one embodiment, the sampling and calculation of the interpolated initial pulse signal to obtain the corresponding first pulse repetition frequency, implemented by the computer program when executed by the processor, includes: uniformly sampling and calculating the interpolated initial pulse signal using a fast Fourier transform to obtain the corresponding first pulse repetition frequency.

[0124] In one embodiment, the preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system implemented when the computer program is executed by the processor includes: the initial angle of arrival of the initial pulse descriptor corresponding to the polar angle of the polar coordinate system, the initial carrier of the initial pulse descriptor corresponding to the color of the polar coordinate system, and the initial pulse width of the initial pulse descriptor corresponding to the radius of the polar coordinate system; the preset relationship between the initial pulse repetition interval and the polar coordinate system includes: the initial pulse repetition interval corresponding to the polar radius of the polar coordinate system.

[0125] In one embodiment, the method for establishing a pulse classification model implemented when a computer program is executed by a processor includes: acquiring training pulse signals and labels; the labels being information corresponding to the reference signal source of the training pulse signals; preprocessing the training pulse signals to obtain the training pulse repetition interval of the training pulse signals; establishing a corresponding polar coordinate training image based on a preset relationship between the training pulse descriptor of the training pulse signals and the polar coordinate system, and a preset relationship between the training pulse repetition interval and the polar coordinate system; inputting the polar coordinate training image and labels into an initial pulse classification model to obtain the predicted target center point of the training pulse signals; determining the predicted signal source of the training pulse signals based on the predicted target center point; obtaining the target loss based on the information of the predicted signal source, the labels, and the target loss function; and adjusting the parameters of the initial pulse classification model based on the target loss to obtain the pulse classification model.

[0126] In one embodiment, the target loss function implemented when the computer program is executed by the processor includes a category prediction loss function, a width prediction loss function, and a location prediction loss function; obtaining the target loss based on the information of the prediction signal source, the label, and the target loss function includes: obtaining the target loss based on the information of the prediction signal source, the label, the category prediction loss function, the width prediction loss function, and the location prediction loss function.

[0127] In one embodiment, when the computer program is executed by a processor, the implemented tag includes a reference carrier, a reference pulse width, a reference angle of arrival, and a reference pulse repetition interval of a reference signal source; and a target loss is obtained based on the information of the predicted signal source, the tag, a category prediction loss function, a width prediction loss function, and a position prediction loss function, including: obtaining a first loss based on the predicted carrier of the predicted signal source, the reference carrier of the reference signal source, and the category prediction loss function; obtaining a second loss based on the predicted pulse width of the predicted signal source, the reference pulse width of the reference signal source, and the width prediction loss function; obtaining a third loss based on the predicted angle of arrival, the predicted pulse repetition interval, and the scaling scale of the predicted signal source, the reference angle of arrival and the reference pulse repetition interval of the reference signal source, and the position prediction loss function; and obtaining the target loss based on the first loss, the second loss, and the third loss.

[0128] In one embodiment, a computer program product is provided, including a computer program that, when executed by a processor, performs the following steps: acquiring an initial pulse signal; preprocessing the initial pulse signal to obtain an initial pulse repetition interval of the initial pulse signal; establishing a corresponding polar coordinate image based on a preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system, and a preset relationship between the initial pulse repetition interval and the polar coordinate system; inputting the polar coordinate image into a pulse classification model to obtain a target center point of the initial pulse signal; and determining the signal source of the initial pulse signal based on the target center point.

[0129] In one embodiment, the computer program, when executed by a processor, preprocesses an initial pulse signal to obtain an initial pulse repetition interval, including: interpolating the initial pulse signal; sampling and calculating the interpolated initial pulse signal to obtain a corresponding first pulse repetition frequency; and setting the first pulse repetition frequency as the initial pulse repetition interval if the threshold value of the first pulse repetition frequency is greater than a target value.

[0130] In one embodiment, the sampling and calculation of the interpolated initial pulse signal to obtain the corresponding first pulse repetition frequency, implemented by the computer program when executed by the processor, includes: uniformly sampling and calculating the interpolated initial pulse signal using a fast Fourier transform to obtain the corresponding first pulse repetition frequency.

[0131] In one embodiment, the preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system implemented when the computer program is executed by the processor includes: the initial angle of arrival of the initial pulse descriptor corresponding to the polar angle of the polar coordinate system, the initial carrier of the initial pulse descriptor corresponding to the color of the polar coordinate system, and the initial pulse width of the initial pulse descriptor corresponding to the radius of the polar coordinate system; the preset relationship between the initial pulse repetition interval and the polar coordinate system includes: the initial pulse repetition interval corresponding to the polar radius of the polar coordinate system.

[0132] In one embodiment, the method for establishing a pulse classification model implemented when a computer program is executed by a processor includes: acquiring training pulse signals and labels; the labels being information corresponding to the reference signal source of the training pulse signals; preprocessing the training pulse signals to obtain the training pulse repetition interval of the training pulse signals; establishing a corresponding polar coordinate training image based on a preset relationship between the training pulse descriptor of the training pulse signals and the polar coordinate system, and a preset relationship between the training pulse repetition interval and the polar coordinate system; inputting the polar coordinate training image and labels into an initial pulse classification model to obtain the predicted target center point of the training pulse signals; determining the predicted signal source of the training pulse signals based on the predicted target center point; obtaining the target loss based on the information of the predicted signal source, the labels, and the target loss function; and adjusting the parameters of the initial pulse classification model based on the target loss to obtain the pulse classification model.

[0133] In one embodiment, the target loss function implemented when the computer program is executed by the processor includes a category prediction loss function, a width prediction loss function, and a location prediction loss function; obtaining the target loss based on the information of the prediction signal source, the label, and the target loss function includes: obtaining the target loss based on the information of the prediction signal source, the label, the category prediction loss function, the width prediction loss function, and the location prediction loss function.

[0134] In one embodiment, when the computer program is executed by a processor, the implemented tag includes a reference carrier, a reference pulse width, a reference angle of arrival, and a reference pulse repetition interval of a reference signal source; and a target loss is obtained based on the information of the predicted signal source, the tag, a category prediction loss function, a width prediction loss function, and a position prediction loss function, including: obtaining a first loss based on the predicted carrier of the predicted signal source, the reference carrier of the reference signal source, and the category prediction loss function; obtaining a second loss based on the predicted pulse width of the predicted signal source, the reference pulse width of the reference signal source, and the width prediction loss function; obtaining a third loss based on the predicted angle of arrival, the predicted pulse repetition interval, and the scaling scale of the predicted signal source, the reference angle of arrival and the reference pulse repetition interval of the reference signal source, and the position prediction loss function; and obtaining the target loss based on the first loss, the second loss, and the third loss.

[0135] It should be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, data stored, data displayed, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties.

[0136] Those skilled in the art will understand that all or part of the processes in the methods of the above embodiments can be implemented by a computer program instructing related hardware. The computer program can be stored in a non-volatile computer-readable storage medium, and when executed, it can include the processes of the embodiments of the above methods. Any references to memory, databases, or other media used in the embodiments provided in this application can include at least one of non-volatile and volatile memory. Non-volatile memory can include read-only memory (ROM), magnetic tape, floppy disk, flash memory, optical memory, high-density embedded non-volatile memory, resistive random access memory (ReRAM), magnetic random access memory (MRAM), ferroelectric random access memory (FRAM), phase change memory (PCM), graphene memory, etc. Volatile memory can include random access memory (RAM) or external cache memory, etc. By way of illustration and not limitation, RAM can take many forms, such as Static Random Access Memory (SRAM) or Dynamic Random Access Memory (DRAM). The databases involved in the embodiments provided in this application may include at least one type of relational database and non-relational database. Non-relational databases may include, but are not limited to, blockchain-based distributed databases. The processors involved in the embodiments provided in this application may be general-purpose processors, central processing units, graphics processing units, digital signal processors, programmable logic devices, quantum computing-based data processing logic devices, etc., and are not limited to these.

[0137] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0138] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this patent application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A signal sorting method, characterized in that, The method includes: Obtain the initial pulse signal; The initial pulse signal is preprocessed to obtain the initial pulse repetition interval of the initial pulse signal; A corresponding polar coordinate image is established based on the preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system, and the preset relationship between the initial pulse repetition interval and the polar coordinate system. The polar coordinate image is input into the pulse classification model to obtain the target center point of the initial pulse signal; The signal source of the initial pulse signal is determined based on the target center point; The preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system includes: The initial angle of arrival of the initial pulse descriptor corresponds to the polar angle of the polar coordinate system, the initial carrier of the initial pulse descriptor corresponds to the color of the polar coordinate system, and the initial pulse width of the initial pulse descriptor corresponds to the radius of the polar coordinate system. The preset relationship between the initial pulse repetition interval and the polar coordinate system includes: the initial pulse repetition interval corresponds to the polar radius of the polar coordinate system.

2. The method according to claim 1, characterized in that, The preprocessing of the initial pulse signal to obtain the initial pulse repetition interval of the initial pulse signal includes: The initial pulse signal is interpolated. The interpolated initial pulse signal is sampled and calculated to obtain the corresponding first pulse repetition frequency; If the threshold value of the first pulse repetition frequency is greater than the target value, the initial pulse repetition interval is obtained based on the first pulse repetition frequency.

3. The method according to claim 2, characterized in that, The step of sampling and calculating the interpolated initial pulse signal to obtain the corresponding first pulse repetition frequency includes: The initial pulse signal after interpolation is uniformly sampled and calculated using a fast Fourier transform to obtain the corresponding first pulse repetition frequency.

4. The method according to claim 1, characterized in that, The pulse classification model is established in the following ways: Acquire training pulse signals and labels; the labels are information corresponding to the reference signal source of the training pulse signals; The training pulse signal is preprocessed to obtain the training pulse repetition interval of the training pulse signal; A corresponding polar coordinate training image is established based on the preset relationship between the training pulse descriptor of the training pulse signal and the polar coordinate system, and the preset relationship between the training pulse repetition interval and the polar coordinate system. The polar coordinate training image and the label are input into the initial pulse classification model to obtain the predicted target center point of the training pulse signal; the predicted signal source of the training pulse signal is determined based on the predicted target center point. The target loss is obtained based on the information of the predicted signal source, the label, and the target loss function; The parameters of the initial pulse classification model are adjusted based on the target loss to obtain the pulse classification model.

5. The method according to claim 4, characterized in that, The target loss function includes a category prediction loss function, a width prediction loss function, and a location prediction loss function; The step of obtaining the target loss based on the information of the predicted signal source, the label, and the target loss function includes: The target loss is obtained based on the information of the predicted signal source, the label, the category prediction loss function, the width prediction loss function, and the position prediction loss function.

6. The method according to claim 5, characterized in that, The tag includes the reference carrier, reference pulse width, reference angle of arrival, and reference pulse repetition interval of the reference signal source; The step of obtaining the target loss based on the information of the predicted signal source, the label, the category prediction loss function, the width prediction loss function, and the position prediction loss function includes: The first loss is obtained based on the predicted carrier of the predicted signal source, the reference carrier of the reference signal source, and the category prediction loss function; The second loss is obtained based on the predicted pulse width of the predicted signal source, the reference pulse width of the reference signal source, and the width prediction loss function; Based on the predicted angle of arrival, predicted pulse repetition interval, and scaling scale of the predicted signal source, the reference angle of arrival and reference pulse repetition interval of the reference signal source, and the position prediction loss function, a third loss is obtained; The target loss is obtained based on the first loss, the second loss, and the third loss.

7. A signal sorting device, characterized in that, The device includes: The acquisition module is used to acquire the initial pulse signal; A preprocessing module is used to preprocess the initial pulse signal to obtain the initial pulse repetition interval of the initial pulse signal; An image building module is used to build a corresponding polar coordinate image based on a preset relationship between the initial pulse descriptor of the initial pulse signal and the polar coordinate system, and a preset relationship between the initial pulse repetition interval and the polar coordinate system; the initial angle of arrival of the initial pulse descriptor corresponds to the polar angle of the polar coordinate system, the initial carrier of the initial pulse descriptor corresponds to the color of the polar coordinate system, and the initial pulse width of the initial pulse descriptor corresponds to the radius of the polar coordinate system; the initial pulse repetition interval corresponds to the polar radius of the polar coordinate system. The center point acquisition module is used to input the polar coordinate image into the pulse classification model to obtain the target center point of the initial pulse signal; The signal source determination module is used to determine the signal source of the initial pulse signal based on the target center point.

8. A computer device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method according to any one of claims 1 to 6.

9. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

10. A computer program product, comprising a computer program, characterized in that, When the computer program is executed by a processor, it implements the steps of the method according to any one of claims 1 to 6.

Images