A Transformer-based real-time pulse compression method and system for radar signals

Abstract

This application provides a real-time pulse compression method and system for radar signals based on Transformer. The method involves: acquiring interference signals from the environment to obtain raw spectrum data; using a Transformer network to obtain interference characteristic parameters; converting the nonlinear frequency modulation parameters in the interference characteristic parameters into anti-interference waveform parameters based on preset waveform generation rules; generating and transmitting an anti-interference transmission signal based on the anti-interference waveform parameters, and receiving the echo signal after transmitting the anti-interference transmission signal; processing the time-varying frequency slope characteristics in the anti-interference transmission signal to obtain bandpass filter range and sampling rate configuration parameters; and processing the echo signal based on the bandpass filter range and sampling rate configuration parameters to generate a pulse compression signal. The technical solution provided in this application enhances the target detection and resolution capabilities of radar systems in complex electromagnetic environments.

Description

Technical Field

[0001] This application relates to the field of signal processing technology, and in particular to a method and system for real-time pulse compression of radar signals based on Transformer. Background Technology

[0002] In environments with strong electromagnetic interference, radar systems face severe signal contamination and pulse compression main lobe distortion, leading to a decrease in target detection reliability. Especially in extreme scenarios, it is still necessary to ensure high-fidelity output of the pulse compression main lobe to maintain effective detection and resolution of critical targets. Therefore, there is an urgent need for an adaptive pulse compression technology capable of real-time interference sensing, dynamic adjustment of the transmitted waveform, and optimization of the receiver link processing mechanism.

[0003] To address the aforementioned technical requirements, existing solutions employ a pulse compression architecture combining spectrum sensing and fixed nonlinear frequency-modulated waveforms, and implement receiver bandpass filtering configuration through a pre-set filter bank. This approach first identifies environmental interference using traditional time-frequency analysis methods, then selects a pre-designed waveform as the transmitted signal, and simultaneously selects appropriate filter bandwidth and sampling rate parameters from a finite set based on interference frequency band information for echo signal processing. However, this solution has certain drawbacks. For example, limited by the static characteristics of the waveform generation method, it cannot flexibly adjust the frequency modulation pattern of the transmitted waveform according to real-time interference characteristics, resulting in insufficient adaptability in complex and variable electromagnetic environments. Furthermore, its receiver parameter configuration relies on empirical models and discrete configuration sets, making it difficult to achieve precise matching of time-varying frequency slope characteristics, which can easily lead to target information loss or increased false alarm rates under strong interference conditions. Summary of the Invention

[0004] This application provides a real-time pulse compression method and system for radar signals based on Transformer, which solves the problems of insufficient adaptability in complex and variable electromagnetic environments, difficulty in achieving fine matching of time-varying frequency slope characteristics, and easy loss of target information or increase in false alarm rate under strong interference conditions in the prior art.

[0005] In a first aspect, this application provides a real-time pulse compression method for radar signals based on Transformer, including:

[0006] The interference signals in the environment are collected, and the interference signals are subjected to spectrum analysis to obtain the raw spectrum data;

[0007] Using a Transformer network, time-frequency characteristic analysis is performed on the interference components in the original spectrum data to obtain interference characteristic parameters;

[0008] Based on preset waveform generation rules, the nonlinear frequency modulation parameters in the interference characteristic parameters are converted into anti-interference waveform parameters.

[0009] Based on the anti-interference waveform parameters, an anti-interference transmission signal is generated and transmitted, and the echo signal after transmitting the anti-interference transmission signal is received;

[0010] The time-varying frequency slope characteristics in the anti-interference transmission signal are processed to obtain the bandpass filter range and sampling rate configuration parameters;

[0011] Based on the bandpass filtering range and the sampling rate configuration parameters, the echo signal is processed to generate a pulse compression signal.

[0012] Optionally, a Transformer network is used to perform time-frequency feature analysis on the interference components in the original spectral data to obtain interference feature parameters, including:

[0013] The interference components in the original spectrum data are time-location encoded to generate a location index vector;

[0014] The location index vector and the interference components are fused to obtain a location marker sequence;

[0015] Using the self-attention mechanism of the Transformer network, the attention weights of each feature element in the location label sequence are calculated to generate an attention weight matrix;

[0016] Based on the attention weight matrix, the feature elements of the location marker sequence are weighted and summed to generate an aggregated feature vector;

[0017] The aggregated feature vector is subjected to dimensionality reduction to extract key interference feature components, and these key interference feature components are used as interference feature parameters.

[0018] Optionally, the location index vector and the interference components are fused to obtain a location marker sequence, including:

[0019] According to the preset time window rules, the interference components are divided into multiple time units, where each time unit corresponds to an interference feature vector;

[0020] The location index vector is divided into location index sub-vectors with the same number of time units;

[0021] The interference feature vector of each time unit and the corresponding position index sub-vector are concatenated to obtain multiple time unit enhanced feature vectors;

[0022] All time-unit enhanced feature vectors are linearly combined to obtain multiple linear combination vectors. According to the preset activation rule, all linear combination vectors are non-linearly activated to obtain multiple activated feature vectors. All activated feature vectors are feature-scaled to obtain multiple fused feature vectors.

[0023] Arrange all the fused feature vectors to obtain the location marker sequence.

[0024] Optionally, based on a preset waveform generation rule, the nonlinear frequency modulation parameter in the interference characteristic parameters is converted into anti-interference waveform parameters, including:

[0025] Based on the preset waveform generation rules, the nonlinear frequency modulation parameter in the interference characteristic parameters is decomposed to obtain the frequency variation range parameter and the signal duration parameter.

[0026] The minimum and maximum frequencies are extracted from the frequency variation range parameter, and multiple time nodes are extracted from the signal duration parameter;

[0027] Using the minimum frequency, the maximum frequency, and the time node, a frequency modulation curve representing the time-frequency correspondence is constructed.

[0028] The slope abrupt change is detected in the time-frequency correspondence to determine the slope jump interval. The slope jump interval is then smoothed to obtain the corrected frequency modulation curve.

[0029] The corrected frequency modulation curve is discretely sampled to generate a discrete sequence consisting of time-frequency point pairs, and the discrete sequence is encoded into anti-interference waveform parameters.

[0030] Optionally, based on the anti-interference waveform parameters, generating and transmitting an anti-interference transmission signal, and receiving the echo signal after transmitting the anti-interference transmission signal, includes:

[0031] Based on the anti-interference waveform parameters, the baseband signal is reconstructed to generate an anti-interference baseband signal;

[0032] Based on the time-varying frequency characteristics in the anti-interference baseband signal, the instantaneous frequency change is calculated, and the instantaneous frequency change is converted into a frequency control word;

[0033] According to the frequency control word, the carrier frequency of the radio frequency local oscillator is set to generate a time-varying local oscillator signal;

[0034] The anti-interference baseband signal and the time-varying local oscillator signal are orthogonally modulated to generate an anti-interference radio frequency signal;

[0035] The power of the anti-interference radio frequency signal is amplified to generate a high-power anti-interference transmission signal, which is then transmitted.

[0036] Receive the echo signal after the high-power anti-interference transmission signal is transmitted.

[0037] Optionally, the time-varying frequency slope characteristics in the anti-interference transmitted signal are processed to obtain bandpass filtering range and sampling rate configuration parameters, including:

[0038] The time-varying frequency slope characteristic is subjected to frequency trajectory reconstruction operation to obtain the time-frequency change curve;

[0039] Calculate the derivative of the time-frequency change curve to generate a frequency change rate curve;

[0040] Calculate the absolute value of the maximum slope of the frequency change rate curve, and calculate the minimum sampling rate of the interference signal based on the absolute value of the maximum slope, and use the minimum sampling rate as a sampling rate configuration parameter;

[0041] The minimum and maximum frequency values ​​are extracted from the time-frequency variation curve. Based on the minimum and maximum frequency values, the passband range of the anti-interference transmission signal is calculated, and the passband range is used as the bandpass filtering range.

[0042] Optionally, based on the bandpass filtering range and the sampling rate configuration parameters, the echo signal is processed to generate a pulse compression signal, including:

[0043] Based on the bandpass filtering range, the echo signal is subjected to frequency band limiting to generate a filtered echo signal;

[0044] Based on the sampling rate configuration parameters, the sampling rate of the filtered echo signal is adjusted to generate a resampled echo signal;

[0045] The resampled echo signal is subjected to in-phase component separation to obtain a baseband in-phase signal, and the resampled echo signal is subjected to quadrature component separation to obtain a baseband quadrature signal.

[0046] A matched filter operation is performed on the baseband in-phase signal and the baseband quadrature signal to obtain a pulse compressed complex signal;

[0047] Calculate the magnitude of the pulse compression complex signal to obtain the pulse compression amplitude signal;

[0048] The main lobe signal is extracted from the pulse compression amplitude signal and used as the pulse compression signal.

[0049] Secondly, this application provides a Transformer-based real-time pulse compression system for radar signals, comprising:

[0050] The acquisition module is used to acquire interference signals in the environment, perform spectrum analysis on the interference signals, and obtain raw spectrum data.

[0051] The analysis module is used to perform time-frequency feature analysis on the interference components in the original spectrum data using a Transformer network to obtain interference feature parameters;

[0052] The conversion module is used to convert the nonlinear frequency modulation parameters in the interference characteristic parameters into anti-interference waveform parameters based on preset waveform generation rules.

[0053] The receiving module is configured to generate an anti-interference transmission signal based on the anti-interference waveform parameters, transmit the anti-interference transmission signal, and receive the echo signal after transmitting the anti-interference transmission signal.

[0054] The processing module is used to process the time-varying frequency slope characteristics in the anti-interference transmission signal to obtain the bandpass filter range and sampling rate configuration parameters;

[0055] The compression module processes the echo signal based on the bandpass filtering range and the sampling rate configuration parameters to generate a pulse compressed signal.

[0056] Thirdly, this application provides a computing device, including a processing component and a storage component; the storage component stores one or more computer instructions; the one or more computer instructions are to be invoked and executed by the processing component to implement a Transformer-based real-time pulse compression method for radar signals as described in the first aspect above.

[0057] Fourthly, this application provides a computer storage medium storing a computer program, which, when executed by a computer, implements a Transformer-based real-time pulse compression method for radar signals as described in the first aspect.

[0058] In this application, interference signals in the environment are collected, and spectral analysis is performed on the interference signals to obtain raw spectral data. A Transformer network is used to perform time-frequency feature analysis on the interference components in the raw spectral data to obtain interference feature parameters. Based on preset waveform generation rules, the nonlinear frequency modulation parameters in the interference feature parameters are converted into anti-interference waveform parameters. Based on the anti-interference waveform parameters, an anti-interference transmission signal is generated and transmitted, and the echo signal after the transmission is received. The time-varying frequency slope characteristics in the anti-interference transmission signal are processed to obtain bandpass filter range and sampling rate configuration parameters. Based on the bandpass filter range and sampling rate configuration parameters, the echo signal is processed to generate a pulse compression signal. The technical solution provided in this application realizes real-time perception of the electromagnetic environment in which the radar is located, obtains the distribution characteristics of interference signals in the frequency domain, and provides basic data support for subsequent interference feature extraction and waveform adaptive adjustment. By introducing the Transformer architecture, which possesses strong large-scale modeling capabilities and global dependency modeling capabilities, the complex time-frequency structures hidden in interference signals can be effectively mined, and physically meaningful and generative interference feature parameters can be extracted, improving the precision and adaptability of feature representation. Based on the extracted interference features, nonlinear frequency modulation parameters matching the characteristics of the current environment are dynamically generated, giving the transmitted waveform good anti-interference characteristics, thereby improving the detection stability of the radar system in complex electromagnetic environments. The practical application and closed-loop feedback of the anti-interference waveform are realized, enabling the radar transmitted signal to actively adapt to the current interference environment, while providing echo data highly matched to the transmitted waveform for subsequent receiving and processing stages. Optimal receiver signal processing parameters are derived to ensure that the receiving link can accurately match the characteristics of the transmitted signal, improving pulse compression performance. With the support of highly adaptable receiving parameters, high-quality pulse compression processing is completed, improving main lobe fidelity and sidelobe suppression capabilities, and enhancing the radar's target discrimination and identification capabilities. This application overcomes the limitations of traditional fixed models in feature extraction under complex interference environments by introducing the self-attention mechanism of Transformer networks. It can more accurately capture the non-stationary and nonlinear variations in interference signals, generating more adaptive and physically interpretable interference feature parameters. Compared to existing methods that rely on limited preset templates or empirical models, this application achieves dynamic modeling and continuous optimization of interference features, providing a more precise foundation for the coordinated adjustment of subsequent transmitted waveforms and received parameters. This enhances the radar system's ability to maintain pulse compression mainlobe integrity and reduce false alarm rates in strong electromagnetic interference environments.

[0059] These or other aspects of this application will become more apparent in the following description of the embodiments. Attached Figure Description

[0060] To more clearly illustrate the technical solutions in the embodiments of this application or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0061] Figure 1 A flowchart of a real-time pulse compression method for radar signals based on Transformer provided in this application;

[0062] Figure 2 A schematic diagram of a real-time pulse compression system for radar signals based on Transformer provided in this application;

[0063] Figure 3 This is a schematic diagram of the structure of a computing device provided in this application. Detailed Implementation

[0064] To enable those skilled in the art to better understand the present application, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings.

[0065] In some of the processes described in the specification, claims, and accompanying drawings of this application, multiple operations appearing in a specific order are included. However, it should be clearly understood that these operations may not be executed in the order they appear herein, or may be executed in parallel. The operation numbers, such as 101, 102, etc., are merely used to distinguish different operations and do not themselves represent any execution order. Furthermore, these processes may include more or fewer operations, and these operations may be executed sequentially or in parallel. It should be noted that the descriptions such as "first," "second," etc., in this document are used to distinguish different messages, devices, modules, etc., and do not represent a chronological order, nor do they limit "first" and "second" to different types.

[0066] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this application.

[0067] In environments with strong electromagnetic interference, radar systems face severe signal contamination and pulse compression main lobe distortion, leading to decreased target detection reliability. Main lobe integrity must be maintained, especially under extreme interference-to-signal ratio conditions, posing a significant challenge to the adaptive capabilities of existing methods based on fixed nonlinear frequency-modulated waveforms and preset filter banks. Traditional methods, limited by static waveform design and discrete parameter configuration mechanisms, struggle to dynamically adjust the transmitted signal format and receiver processing strategy according to real-time interference characteristics, resulting in insufficient adaptability and stability in complex and variable electromagnetic environments. To address these issues, this application proposes a Transformer-based real-time pulse compression method for radar signals. By acquiring environmental interference signals and performing deep spectrum analysis, the method utilizes a Transformer network to extract the time-frequency characteristics of interference components, generating physically meaningful interference characteristic parameters. Based on these parameters, a nonlinear frequency-modulated waveform with stronger anti-interference performance is dynamically synthesized. Simultaneously, by combining the time-varying frequency slope characteristics of the transmitted signal, the optimal bandpass filter range and sampling rate configuration parameters are intelligently derived, achieving closed-loop optimization of the transmit and receive links. This improves the detection accuracy and resolution of key targets in complex electromagnetic environments.

[0068] Figure 1 A flowchart of a real-time pulse compression method for radar signals based on Transformer is provided in this application embodiment, as shown below. Figure 1 As shown, the method includes:

[0069] Step 101: Collect interference signals in the environment, perform spectrum analysis on the interference signals, and obtain raw spectrum data.

[0070] In this step, the raw spectrum data refers to the two-dimensional matrix obtained by performing spectrum analysis on the electromagnetic interference signal.

[0071] In this embodiment of the application, interference signals in the environment are collected, and a fast Fourier transform operation is performed on the interference signals to generate raw spectrum data containing frequency, amplitude, and phase information.

[0072] Step 102: Using a Transformer network, perform time-frequency feature analysis on the interference components in the original spectrum data to obtain interference feature parameters.

[0073] In this step, the Transformer network refers to a deep learning architecture that incorporates multi-head self-attention and positional encoding to extract time-frequency coupled features of interference. Interference components refer to the set of anomalous frequency points in the original spectral data whose amplitude exceeds the environmental noise threshold. Interference feature parameters refer to the feature vectors generated by the Transformer network.

[0074] In this embodiment, the encoding layer structure of the Transformer network is used to perform time-frequency joint attention calculation on the interference components in the original spectrum data. By capturing the sparse distribution characteristics of the interference components in the time dimension and the harmonic correlation characteristics in the frequency dimension, an interference feature parameter vector is generated.

[0075] Step 103: Based on the preset waveform generation rules, convert the nonlinear frequency modulation parameters in the interference characteristic parameters into anti-interference waveform parameters.

[0076] In this step, the preset waveform generation rules refer to the set of nonlinear frequency modulation constraints stored in the radar system. The anti-jamming waveform parameters refer to the discrete time and frequency point pair sequence generated based on the interference characteristics.

[0077] In this embodiment of the application, based on the nonlinear constraint conditions in the preset waveform generation rules, the nonlinear frequency modulation parameter in the interference characteristic parameters is smoothed and corrected to generate anti-interference waveform parameters.

[0078] Step 104: Based on the anti-interference waveform parameters, generate an anti-interference transmission signal and transmit the anti-interference transmission signal, and receive the echo signal after transmitting the anti-interference transmission signal.

[0079] In this step, the anti-jamming transmitted signal refers to a radio frequency signal synthesized based on anti-jamming waveform parameters, whose instantaneous frequency changes non-linearly with time to avoid interference frequency bands. The echo signal refers to the delayed and attenuated signal formed after the anti-jamming transmitted signal is reflected by the target.

[0080] In this embodiment of the application, the anti-interference waveform parameters are reconstructed using a baseband method to generate an anti-interference baseband signal, which is then radiated into space via a power amplifier, and the echo signal after transmitting the anti-interference signal is received.

[0081] Step 105: Process the time-varying frequency slope characteristics in the anti-interference transmission signal to obtain the bandpass filter range and sampling rate configuration parameters.

[0082] In this step, the time-varying frequency slope characteristic refers to the first derivative function of the frequency of the anti-interference transmitted signal with respect to time. The bandpass filtering range refers to the filter passband interval determined based on the instantaneous frequency boundary of the signal. The sampling rate configuration parameter refers to the lowest sampling frequency that satisfies the Nyquist sampling theorem.

[0083] In this embodiment, derivative extreme value analysis is performed on the time-varying frequency slope characteristics of the anti-interference transmission signal to calculate the maximum absolute value of the slope to determine the sampling rate configuration parameters, and frequency boundary values ​​are extracted to calculate the bandpass filtering range.

[0084] Step 106: Based on the bandpass filtering range and the sampling rate configuration parameters, process the echo signal to generate a pulse compression signal.

[0085] In this step, the pulse compression signal refers to the time-domain amplitude signal after matched filtering.

[0086] In this embodiment, based on the bandpass filtering range, the echo signal is subjected to frequency band limiting operation, and based on the sampling rate configuration parameters, correlation calculations are performed on the echo signal after matched filtering to generate a pulse compression signal.

[0087] This application embodiment dynamically extracts the time-frequency coupling characteristics of interference through a Transformer network to generate an anti-interference transmission waveform with frequency agility; it derives down-conversion parameters in real time based on the time-varying characteristics of the transmitted signal to achieve adaptive reconstruction of receiver hardware parameters; and finally outputs a high-resolution pulse compression signal in a strong interference environment, solving the problems of interference suppression failure and resolution degradation caused by fixed waveforms and fixed receiving parameters in traditional radar.

[0088] This application provides a specific embodiment. Step 102 involves using a Transformer network to perform time-frequency feature analysis on the interference components in the original spectrum data to obtain interference feature parameters. This specifically includes the following steps:

[0089] Step 201: Perform time-location encoding on the interference components in the original spectrum data to generate a location index vector.

[0090] In this step, the location index vector refers to a mathematical vector generated through time location encoding, which is used to reflect the absolute location information of the interference components on the time axis.

[0091] In this embodiment, time-frequency analysis is performed on the interference components in the original spectrum data, the center timestamp of each time segment is extracted, the center timestamp is divided by the total duration and then multiplied by the preset position coding dimension to generate a position index vector containing time sequence information. The calculation formula of the position index vector is: center timestamp ÷ total duration × preset position coding dimension.

[0092] Step 202: Perform feature fusion on the location index vector and the interference components to obtain a location marker sequence.

[0093] In this step, the location marker sequence refers to the feature sequence after fusing the location index vector with the spectral data of the interference components.

[0094] In this embodiment, feature splicing technology is used to connect the position index sub-vector of the position index vector with the interference feature vector of the interference component along the column direction. After linear combination, the vectors are arranged to obtain the position mark sequence.

[0095] Step 203: Calculate the attention weights of each feature element in the location label sequence using the self-attention mechanism of the Transformer network, and generate an attention weight matrix.

[0096] In this step, attention weights refer to the scalar values ​​calculated by the self-attention mechanism. The attention weight matrix is ​​a two-dimensional matrix composed of all attention weights.

[0097] In this embodiment, the location marker sequence is input into the Transformer encoding layer. Through the self-attention mechanism, the query vector and the key vector are multiplied by a dot product and divided by a scaling factor to calculate the attention weight of each feature element, thus obtaining the attention weight matrix.

[0098] Step 204: Based on the attention weight matrix, perform a weighted summation of each feature element of the position tag sequence to generate an aggregated feature vector.

[0099] In this step, the aggregated feature vector refers to the output vector after the location marker sequence is weighted and summed, which includes feature representations that incorporate global context information, and its dimension is the same as that of the location marker sequence.

[0100] In this embodiment, matrix multiplication is performed on the value vectors of the attention weight matrix and the position tag sequence, and a weighted sum is calculated for each feature element to generate an aggregated feature vector.

[0101] Step 205: Perform dimensionality reduction processing on the aggregated feature vector to extract key interference feature components, and use the key interference feature components as interference feature parameters.

[0102] In this step, the key interference feature components refer to the feature subset after dimensionality reduction, which are used to characterize the essential properties of the interference.

[0103] In this embodiment, a dimensionality reduction algorithm is used to perform feature selection on the aggregated feature vector, retaining the key interference feature components that contribute the most to the variance, and these key interference feature components are used as interference feature parameters.

[0104] This application's embodiments capture the temporal evolution of interference through time-location coding, construct a joint time-frequency representation using feature fusion, quantify the dynamic correlation of frequency band units based on a self-attention mechanism, and finally extract physically interpretable interference feature components, thus solving the defect of traditional spectrum analysis methods that ignore time-frequency coupling characteristics.

[0105] This application provides a specific embodiment. Step 202 involves fusing the location index vector and the interference components to obtain a location marker sequence, specifically including the following steps:

[0106] Step 211: According to the preset time window rules, the interference components are divided into multiple time units, where each time unit corresponds to an interference feature vector.

[0107] In this step, the preset time window rule refers to a pre-defined fixed-duration segmentation strategy. The interference feature vector refers to the mathematical representation of the interference spectrum within a single time unit, reflecting the frequency domain energy distribution of the interference during that period.

[0108] In this embodiment, the interference components are divided into time units according to a preset time window rule. Each segment is subjected to Fourier transform to generate an interference feature vector containing the frequency amplitude spectrum.

[0109] Step 212: Divide the position index vector into position index sub-vectors that are consistent with the number of time units.

[0110] In this step, the position index sub-vector refers to the segment after the position index vector is divided, representing the position weight of the corresponding time unit in the overall sequence.

[0111] In this embodiment, the position index vector is divided into equal segments, with the number of segments equal to the number of time units, to generate position index sub-vectors that correspond one-to-one with the time units.

[0112] Step 213: Concatenate the interference feature vector of each time unit with the corresponding position index sub-vector to obtain multiple time unit enhanced feature vectors.

[0113] In this step, the time unit enhanced feature vector refers to the composite vector obtained by concatenating the interference feature vector and the position index sub-vector.

[0114] In this embodiment, a vector concatenation operation is used to concatenate the interference feature vector and the corresponding position index sub-vector of each time unit along the feature dimension to obtain multiple time unit enhanced feature vectors.

[0115] Step 214: Linearly combine all time unit enhanced feature vectors to obtain multiple linear combination vectors. According to the preset activation rule, nonlinearly activate all linear combination vectors to obtain multiple activated feature vectors. Perform feature scaling on all activated feature vectors to obtain multiple fused feature vectors.

[0116] In this step, the linear combination vector refers to the dot product of the time-unit enhanced feature vector and the weight vector. The preset activation rule refers to the nonlinear mapping strategy. The activated feature vector refers to the scalar result after processing the linear combination vector. The fused feature vector refers to the final output after scaling and offsetting the activated feature vector.

[0117] In this embodiment, a dot product operation is performed on the preset weight vector and the time unit enhanced feature vector to generate a linear combination value. According to the preset activation rule, the linear combination value is non-linearly activated to generate activated feature values. A scaling and offset operation is performed on the activated feature values ​​to generate fused feature values.

[0118] Step 215: Arrange all fused feature vectors to obtain the location marker sequence.

[0119] In this embodiment of the application, all fused feature vectors are sorted in ascending order according to the time unit index to obtain a position marker sequence.

[0120] This application's embodiments achieve refined segmentation of interference through time window rules, utilize vector concatenation to fuse time-frequency features, enhance the expression of key information based on weighted nonlinear transformation, and finally construct a time-preserving position marker sequence to resolve the inherent contradiction between time resolution and frequency resolution in traditional spectrum analysis.

[0121] This application provides a specific embodiment. Step 103 involves converting the nonlinear frequency modulation parameter in the interference characteristic parameters into anti-interference waveform parameters based on a preset waveform generation rule. This specifically includes the following steps:

[0122] Step 301: Based on the preset waveform generation rules, decompose the nonlinear frequency modulation parameter in the interference characteristic parameters to obtain the frequency change range parameter and the signal duration parameter.

[0123] In this step, the frequency variation range parameter refers to the two-dimensional vector representing the frequency variation interval in the interference characteristic parameters, reflecting the signal bandwidth. The signal duration parameter refers to the scalar value representing the signal duration in the interference characteristic parameters.

[0124] In this embodiment, based on a preset waveform generation rule, a parameter decomposition operation is performed on the nonlinear frequency modulation parameter in the interference characteristic parameters to extract the upper and lower boundary values ​​of frequency change, thereby forming the frequency change range parameter, and the signal duration is extracted to form the signal duration parameter.

[0125] Step 302: Extract the minimum and maximum frequencies from the frequency variation range parameters, and extract multiple time nodes from the signal duration parameters.

[0126] In this step, the minimum frequency refers to the lower limit frequency value in the frequency variation range parameter, which determines the lowest carrier frequency of the transmitted signal. The maximum frequency refers to the upper limit frequency value in the frequency variation range parameter, which determines the highest carrier frequency of the transmitted signal. A time node refers to a set of time points selected at equal intervals within the signal's duration.

[0127] In this embodiment, extreme value detection technology is used to identify the minimum and maximum frequencies from the frequency variation range parameters, and multiple time nodes are extracted from the signal duration parameters through an equal-interval sampling method.

[0128] Step 303: Using the minimum frequency, the maximum frequency, and the time node, construct a frequency modulation curve that characterizes the time-frequency correspondence.

[0129] In this step, the frequency modulation curve refers to the continuous function curve connecting the time node and the corresponding frequency value, reflecting the instantaneous frequency change law of the signal.

[0130] In this embodiment, cubic spline interpolation is performed based on the minimum frequency, maximum frequency, and time node to construct a frequency modulation curve that represents the time-frequency correspondence, with the horizontal axis representing the time node and the vertical axis representing the corresponding frequency value.

[0131] Step 304: Detect abrupt slope changes in the time-frequency correspondence, determine the slope jump interval, and smooth the slope jump interval to obtain the corrected frequency modulation curve.

[0132] In this step, the slope jump interval refers to the time period during which the first derivative of the frequency modulation curve changes abruptly. The corrected frequency modulation curve refers to the frequency modulation curve after smoothing.

[0133] In this embodiment, the slope abrupt change point of the time-frequency correspondence is detected, and the slope jump interval is marked with the criterion that the difference between the derivatives of adjacent nodes exceeds a preset threshold. Piecewise polynomial fitting is used for smoothing correction to generate the corrected frequency modulation curve.

[0134] Step 305: Discretely sample the corrected frequency modulation curve to generate a discrete sequence consisting of time-frequency point pairs, and encode the discrete sequence into anti-interference waveform parameters.

[0135] In this step, the discrete sequence refers to the discretized representation of the corrected frequency modulation curve.

[0136] In this embodiment of the application, the modified frequency modulation curve is discretely sampled to generate a discrete sequence composed of time-frequency point pairs. The discrete sequence is then converted into anti-interference waveform parameters through binary quantization encoding.

[0137] This application's embodiments construct physically realizable frequency modulation curves through cubic spline interpolation and eliminate slope jumps based on derivative continuity detection; waveform parameters adapted to the hardware platform are generated through discrete sampling, solving the problem of pulse compression and sidelobe elevation caused by slope discontinuity in traditional nonlinear frequency modulation waveforms.

[0138] This application provides a specific embodiment. Step 104 involves generating and transmitting an anti-interference transmission signal based on the anti-interference waveform parameters, and receiving the echo signal after transmitting the anti-interference transmission signal. The specific steps include:

[0139] Step 401: Based on the anti-interference waveform parameters, reconstruct the baseband signal to generate an anti-interference baseband signal.

[0140] In this step, the anti-interference baseband signal refers to the complex signal generated based on the waveform parameters.

[0141] In this embodiment, a baseband reconstruction operation is performed based on the anti-interference waveform parameters to generate an anti-interference baseband signal.

[0142] Step 402: Based on the time-varying frequency characteristics in the anti-interference baseband signal, calculate the instantaneous frequency change and convert the instantaneous frequency change into a frequency control word.

[0143] In this step, the time-varying frequency characteristic refers to the derivative function of the instantaneous frequency of the baseband signal with respect to time. The instantaneous frequency change refers to the frequency difference between adjacent sampling points. The frequency control word refers to the digital instruction that controls the frequency synthesizer.

[0144] In this embodiment, the frequency difference between adjacent sampling points is calculated based on the time-varying frequency characteristics of the anti-interference baseband signal, and the difference is multiplied by a preset resolution parameter to generate a frequency control word.

[0145] Step 403: Set the carrier frequency of the radio frequency local oscillator source according to the frequency control word to generate a time-varying local oscillator signal.

[0146] In this step, the radio frequency local oscillator source refers to a signal generator employing direct digital frequency synthesis technology. The carrier frequency refers to the output frequency of the local oscillator signal. The time-varying local oscillator signal refers to a radio frequency signal whose carrier frequency changes non-linearly with time.

[0147] In this embodiment of the application, the carrier frequency of the radio frequency local oscillator is set according to the frequency control word, and its output is controlled to be a time-varying local oscillator signal that is synchronously agile with the baseband signal.

[0148] Step 404: Orthogonally modulate the anti-interference baseband signal and the time-varying local oscillator signal to generate an anti-interference radio frequency signal.

[0149] In this step, the anti-interference radio frequency signal refers to the product of the baseband signal and the local oscillator signal modulated together.

[0150] In this embodiment, one component of the anti-interference baseband signal is multiplied by the time-varying local oscillator signal, and the other two components are multiplied by the quadrature phase-shifted local oscillator signal. The two products are then combined to generate an anti-interference radio frequency signal.

[0151] Step 405: Amplify the power of the anti-interference radio frequency signal to generate a high-power anti-interference transmission signal and transmit it.

[0152] In this step, the high-power anti-interference transmission signal refers to the radio frequency signal after power amplification.

[0153] In this embodiment, the anti-interference radio frequency signal is amplified to generate a high-power anti-interference transmission signal and then transmitted.

[0154] Step 406: Receive the echo signal after the high-power anti-interference transmission signal is transmitted.

[0155] In this embodiment of the application, the electromagnetic wave energy reflected by the target is captured by the receiving antenna, and an echo signal is generated after the signal-to-noise ratio is improved.

[0156] This application embodiment achieves synchronous agility of local oscillator signal and baseband waveform through dynamic frequency control word, and uses quadrature modulation to ensure signal phase continuity; finally, it generates time-varying spectrum transmission signal with adjustable power, solving the technical defect of traditional fixed frequency radar being easily identified by jammers.

[0157] This application provides a specific embodiment. Step 105 involves processing the time-varying frequency slope characteristics in the anti-interference transmission signal to obtain the bandpass filter range and sampling rate configuration parameters. This specifically includes the following steps:

[0158] Step 501: Perform a frequency trajectory restoration operation on the time-varying frequency slope characteristic to obtain the time-frequency change curve.

[0159] In this step, the time-frequency change curve refers to the continuous function curve restored by integration, which reflects the complete frequency change process of the transmitted signal.

[0160] In this embodiment of the application, a numerical integration algorithm is used to continuously accumulate the time-varying frequency slope characteristics, converting the discrete slope values ​​into a continuous frequency trajectory, and generating a time-frequency change curve with time on the horizontal axis and instantaneous frequency on the vertical axis.

[0161] Step 502: Calculate the derivative of the time-frequency change curve to generate a frequency change rate curve.

[0162] In this step, the frequency change rate curve refers to the first derivative function of the time-frequency change curve.

[0163] In this embodiment, the central difference method is used to perform differential calculations on the time-frequency change curve, and the frequency difference between adjacent sampling points is divided by the time interval to generate a discrete frequency change rate sequence. The sequence points are then connected to obtain the frequency change rate curve.

[0164] Step 503: Calculate the maximum absolute value of the slope of the frequency change rate curve, and calculate the minimum sampling rate of the interference signal based on the maximum absolute value of the slope, and use the minimum sampling rate as the sampling rate configuration parameter.

[0165] In this step, the maximum slope absolute value refers to the absolute value of the largest positive or negative slope in the frequency change rate curve, characterizing the intensity of the most dramatic frequency jump in the signal. The minimum sampling rate refers to the theoretical lowest sampling frequency calculated based on the maximum slope absolute value.

[0166] In this embodiment of the application, the maximum absolute value of the slope is obtained by traversing the frequency change rate curve, and the minimum sampling rate of the interference signal is calculated. The minimum sampling rate is calculated as the maximum absolute value of the slope × the pulse width × 2.5, and the minimum sampling rate that satisfies the Nyquist theorem is generated as the sampling rate configuration parameter.

[0167] Step 504: Extract the minimum and maximum frequency values ​​from the time-frequency variation curve, calculate the passband range of the anti-interference transmission signal based on the minimum and maximum frequency values, and use the passband range as the bandpass filtering range.

[0168] In this step, the minimum frequency value refers to the trough frequency value in the time-frequency variation curve, which determines the lower limit reference of the signal bandwidth. The maximum frequency value refers to the peak frequency value in the time-frequency variation curve, which determines the upper limit reference of the signal bandwidth. The passband range refers to the effective passband interval of the bandpass filter.

[0169] In this embodiment, the minimum point of the detection time-frequency change curve is taken as the minimum frequency value, the maximum point is taken as the maximum frequency value, the minimum frequency value minus a fixed proportion of bandwidth is taken as the lower limit of the passband range, and the maximum frequency value plus a fixed proportion of bandwidth is taken as the upper limit of the passband range, thus obtaining the passband range as the bandpass filtering range.

[0170] This application embodiment reconstructs the complete signal spectrum characteristics by restoring the frequency trajectory, ensures distortion-free signal acquisition by dynamically calculating the sampling rate based on the maximum slope, and generates a fault-tolerant passband range by combining boundary frequency extension, thus solving the problem of high-speed frequency conversion signal distortion caused by traditional fixed receiving parameters.

[0171] This application provides a specific embodiment. Step 106 involves processing the echo signal based on the bandpass filter range and the sampling rate configuration parameters to generate a pulse compression signal. This specifically includes the following steps:

[0172] Step 601: Based on the bandpass filtering range, perform a frequency band limiting operation on the echo signal to generate a filtered echo signal.

[0173] In this step, the filtered echo signal refers to the signal processed by the bandpass filter, reflecting the effective frequency band energy of the target reflected echo.

[0174] In this embodiment, based on the bandpass filtering range, the upper and lower cutoff frequencies of the passband are set, and a spectrum filtering operation is performed on the echo signal to suppress out-of-band noise and generate a filtered echo signal.

[0175] Step 602: Based on the sampling rate configuration parameters, adjust the sampling rate of the filtered echo signal to generate a resampled echo signal.

[0176] In this step, the resampled echo signal refers to the discrete signal after the sampling rate is adjusted.

[0177] In this embodiment, based on the sampling rate configuration parameters, the system clock frequency is reset, and interpolation or decimation operations are performed on the filtered echo signal to generate a resampled echo signal.

[0178] Step 603: Perform in-phase component separation on the resampled echo signal to obtain a baseband in-phase signal, and simultaneously perform quadrature component separation on the resampled echo signal to obtain a baseband quadrature signal.

[0179] In this step, the baseband in-phase signal refers to the in-phase component signal output by quadrature demodulation, reflecting the real part of the target's scattering characteristics. The baseband quadrature signal refers to the quadrature component signal output by quadrature demodulation, reflecting the imaginary part of the target's scattering characteristics.

[0180] In this embodiment, the resampled echo signal is mixed and separated from a preset reference carrier to obtain a baseband in-phase signal, and then mixed and separated from a preset phase-shifted reference carrier to obtain a baseband quadrature signal.

[0181] Step 604: Perform matched filtering on the baseband in-phase signal and the baseband quadrature signal to obtain a pulse compressed complex signal.

[0182] In this step, the pulse compressed complex signal refers to the complex signal output by the matched filter.

[0183] In this embodiment, the baseband in-phase signal is convolved with the real part of the transmitted waveform complex conjugate, and the baseband quadrature signal is convolved with the imaginary part, and then combined to obtain a pulse compressed complex signal.

[0184] Step 605: Calculate the magnitude of the pulse compression complex signal to obtain the pulse compression amplitude signal.

[0185] In this step, the pulse compression amplitude signal refers to the real signal composed of the complex signal magnitude.

[0186] In this embodiment, the square root operation is performed after summing the squares of the real and imaginary parts of the pulse-compressed complex signal to obtain the pulse-compressed amplitude signal.

[0187] Step 606: Extract the main lobe signal from the pulse compression amplitude signal and use the main lobe signal as the pulse compression signal.

[0188] In this step, the main lobe signal refers to the effective target response pulse in the pulse compression amplitude signal that satisfies the condition that the amplitude exceeds the noise floor by a multiple and the half-power width is less than the inverse of the bandwidth.

[0189] In this embodiment, threshold detection is performed on the pulse compression amplitude signal to extract the effective main lobe signal that satisfies the condition that the amplitude exceeds the noise floor by a multiple and the pulse width is less than the inverse of the bandwidth. The main lobe signal is then used as the pulse compression signal.

[0190] This application embodiment suppresses out-of-band interference through dynamic programmable filtering and ensures signal integrity through resampling; it extracts full target scattering information based on orthogonal decoupling and improves range resolution through matched filtering; finally, it outputs a high-confidence main lobe signal, solving the technical bottleneck of side lobes masking weak targets in traditional pulse compression.

[0191] Figure 2 This application provides a schematic diagram of the structure of a real-time pulse compression system for radar signals based on Transformer, as shown in the embodiment. Figure 2 As shown, the system includes:

[0192] Acquisition module 21 is used to acquire interference signals in the environment, perform spectrum analysis on the interference signals, and obtain raw spectrum data;

[0193] Analysis module 22 is used to perform time-frequency feature analysis on the interference components in the original spectrum data using a Transformer network to obtain interference feature parameters;

[0194] The conversion module 23 is used to convert the nonlinear frequency modulation parameter in the interference characteristic parameters into anti-interference waveform parameters based on a preset waveform generation rule.

[0195] The receiving module 24 is used to generate an anti-interference transmission signal based on the anti-interference waveform parameters, transmit the anti-interference transmission signal, and receive the echo signal after transmitting the anti-interference transmission signal;

[0196] Processing module 25 is used to process the time-varying frequency slope characteristics in the anti-interference transmission signal to obtain bandpass filter range and sampling rate configuration parameters;

[0197] Compression module 26 processes the echo signal based on the bandpass filtering range and the sampling rate configuration parameters to generate a pulse compressed signal.

[0198] Figure 2 The aforementioned Transformer-based real-time pulse compression system for radar signals can perform... Figure 1 The implementation principle and technical effects of the Transformer-based real-time pulse compression method for radar signals described in the illustrated embodiment will not be repeated here. The specific methods by which each module and unit performs its operations in the Transformer-based real-time pulse compression system for radar signals described in the above embodiments have been described in detail in the embodiments related to this method, and will not be elaborated upon here.

[0199] In one possible design, Figure 2 The Transformer-based real-time pulse compression system for radar signals shown in the embodiment can be implemented as a computing device, such as... Figure 3 As shown, the computing device may include a storage component 31 and a processing component 32;

[0200] The storage component 31 stores one or more computer instructions, wherein the one or more computer instructions are invoked and executed by the processing component 32.

[0201] The processing component 32 is used for the above Figure 1 The embodiment describes a real-time pulse compression method for radar signals based on Transformer.

[0202] The processing component 32 may include one or more processors to execute computer instructions to complete all or part of the steps in the above-described method. Alternatively, the processing component may be implemented as 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 above-described method.

[0203] Storage component 31 is configured to store various types of data to support operations at the terminal. The storage component can be implemented by 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.

[0204] Of course, computing devices may also include other components, such as input / output interfaces, display components, communication components, etc.

[0205] Input / output interfaces provide interfaces between processing components and peripheral interface modules, which can be output devices, input devices, etc.

[0206] The communication components are configured to facilitate wired or wireless communication between computing devices and other devices.

[0207] The computing device can be a physical device or an elastic computing host provided by a cloud computing platform. In this case, the computing device can refer to a cloud server, and the aforementioned processing components, storage components, etc., can be basic server resources rented or purchased from the cloud computing platform.

[0208] This application also provides a computer storage medium storing a computer program, which, when executed by a computer, can perform the above-described functions. Figure 1 The embodiment shown is a real-time pulse compression method for radar signals based on Transformer.

[0209] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0210] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0211] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0212] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A real-time pulse compression method for radar signals based on Transformer, characterized in that, The method comprises the following steps: Collecting interference signals in the environment, performing spectral analysis on the interference signals to obtain original spectrum data; Using a Transformer network to analyze the time-frequency characteristics of the interference components in the original spectrum data to obtain interference characteristic parameters; Based on the preset waveform generation rule, the nonlinear frequency modulation parameters in the interference characteristic parameters are converted into anti-interference waveform parameters; Based on the anti-interference waveform parameters, an anti-interference transmitting signal is generated and transmitted, and an echo signal after transmitting the anti-interference transmitting signal is received; Processing the time-varying frequency slope characteristics in the anti-interference transmitting signal to obtain a band-pass filtering range and a sampling rate configuration parameter; Based on the band-pass filtering range and the sampling rate configuration parameter, the echo signal is processed to generate a pulse compression signal; Using a Transformer network to analyze the time-frequency characteristics of the interference components in the original spectrum data to obtain interference characteristic parameters, including: Performing time position encoding on the interference components in the original spectrum data to generate a position index vector; Feature fusion is performed on the position index vector and the interference components to obtain a position mark sequence; Using the self-attention mechanism of the Transformer network, the attention weights of each feature element in the position mark sequence are calculated to generate an attention weight matrix; Based on the attention weight matrix, the weighted sum of each feature element in the position mark sequence is performed to generate an aggregated feature vector; The aggregated feature vector is processed to extract key interference feature components, and the key interference feature components are taken as interference characteristic parameters; Processing the time-varying frequency slope characteristics in the anti-interference transmitting signal to obtain a band-pass filtering range and a sampling rate configuration parameter, including: Performing frequency trajectory restoration operation on the time-varying frequency slope characteristics to obtain a time-frequency change curve; Calculating the derivative of the time-frequency change curve to generate a frequency change rate curve; Calculating the maximum slope absolute value of the frequency change rate curve to calculate the minimum sampling rate of the interference signal according to the maximum slope absolute value, and taking the minimum sampling rate as the sampling rate configuration parameter; From the time-frequency change curve, the minimum frequency value and the maximum frequency value are extracted, and based on the minimum frequency value and the maximum frequency value, the passband range of the anti-interference transmitting signal is calculated, and the passband range is taken as the band-pass filtering range.

2. The method of claim 1, wherein, Feature fusion is performed on the position index vector and the interference components to obtain a position mark sequence, including: According to the preset time window rule, the interference components are divided into a plurality of time units, wherein each time unit corresponds to an interference feature vector; The position index vector is divided into position index sub-vectors consistent with the number of time units; The interference feature vector of each time unit and the corresponding position index sub-vector are spliced to obtain a plurality of time unit enhanced feature vectors; Linearly combine all time unit enhancement feature vectors to obtain multiple linear combination vectors, perform nonlinear activation on all linear combination vectors according to a preset activation rule to obtain multiple activated feature vectors, and perform feature scaling on all activated feature vectors to obtain multiple fusion feature vectors; Arrange all fusion feature vectors to obtain a position marker sequence.

3. The method of claim 1, wherein, Based on a preset waveform generation rule, the nonlinear frequency modulation parameter in the interference characteristic parameter is converted into an anti-interference waveform parameter, including: Based on a preset waveform generation rule, the nonlinear frequency modulation parameter in the interference characteristic parameter is decomposed to obtain a frequency variation range parameter and a signal time width parameter; From the frequency variation range parameter, the minimum frequency and the maximum frequency are extracted, and from the signal time width parameter, multiple time nodes are extracted; Using the minimum frequency, the maximum frequency and the time nodes, a frequency modulation curve representing the time-frequency correspondence relationship is constructed; The time-frequency correspondence relationship is detected for slope jump to determine a slope jump interval, and the slope jump interval is modified for smoothness to obtain a modified frequency modulation curve; The modified frequency modulation curve is discretely sampled to generate a discrete sequence composed of time-frequency point pairs, and the discrete sequence is encoded into an anti-interference waveform parameter.

4. The method of claim 1, wherein, Based on the anti-interference waveform parameter, an anti-interference transmission signal is generated and transmitted, and an echo signal after transmitting the anti-interference transmission signal is received, including: Based on the anti-interference waveform parameter, a baseband signal is reconstructed to generate an anti-interference baseband signal; Based on the time-varying frequency characteristic in the anti-interference baseband signal, an instantaneous frequency variation amount is calculated, and the instantaneous frequency variation amount is converted into a frequency control word; According to the frequency control word, the carrier frequency of the radio frequency local oscillator source is set to generate a time-varying local oscillator signal; The anti-interference baseband signal and the time-varying local oscillator signal are quadrature modulated to generate an anti-interference radio frequency signal; The power of the anti-interference radio frequency signal is amplified to generate a high-power anti-interference transmission signal and transmit it; The echo signal after transmitting the high-power anti-interference transmission signal is received.

5. The method of claim 1, wherein, Based on the bandpass filtering range and the sampling rate configuration parameter, the echo signal is processed to generate a pulse compression signal, including: Based on the bandpass filtering range, a band limiting operation is performed on the echo signal to generate a filtered echo signal; Based on the sampling rate configuration parameter, the sampling rate of the filtered echo signal is adjusted to generate a resampled echo signal; The in-phase component separation operation is performed on the resampled echo signal to obtain a baseband in-phase signal, and the quadrature component separation operation is performed on the resampled echo signal to obtain a baseband quadrature signal; The baseband in-phase signal and the baseband quadrature signal are subjected to a matched filtering operation to obtain a pulse compression complex signal; The modulus value of the pulse compression complex signal is calculated to obtain a pulse compression amplitude signal; The main lobe signal is extracted from the pulse compression amplitude signal, and the main lobe signal is taken as the pulse compression signal.

6. A real-time pulse compression system for radar signals based on Transformer, characterized in that, Including: The collection module is used for collecting an interference signal in an environment, performing spectrum analysis on the interference signal, and obtaining original spectrum data; The analysis module is used for performing time-frequency feature analysis on interference components in the original spectrum data by using a Transformer network, and obtaining interference feature parameters; The conversion module is used for converting nonlinear frequency modulation parameters in the interference feature parameters into anti-interference waveform parameters based on a preset waveform generation rule; The receiving module is used for generating an anti-interference transmission signal based on the anti-interference waveform parameters, transmitting the anti-interference transmission signal, and receiving an echo signal after the anti-interference transmission signal is transmitted; The processing module is used for processing time-varying frequency slope characteristics in the anti-interference transmission signal, and obtaining a band-pass filtering range and a sampling rate configuration parameter; The compression module is used for processing the echo signal based on the band-pass filtering range and the sampling rate configuration parameter, and generating a pulse compression signal. The analysis module is used for performing time-frequency feature analysis on interference components in the original spectrum data by using a Transformer network, and obtaining interference feature parameters, including: Performing time position encoding on the interference components in the original spectrum data to generate a position index vector; Performing feature fusion on the position index vector and the interference components to obtain a position mark sequence; Using a self-attention mechanism of the Transformer network, calculating attention weights of each feature element in the position mark sequence to generate an attention weight matrix; Based on the attention weight matrix, performing weighted summation on each feature element of the position mark sequence to generate an aggregated feature vector; Performing dimension reduction processing on the aggregated feature vector to extract a key interference feature component, and taking the key interference feature component as an interference feature parameter; The processing module is used for processing time-varying frequency slope characteristics in the anti-interference transmission signal, and obtaining a band-pass filtering range and a sampling rate configuration parameter, including: Performing a frequency trajectory restoration operation on the time-varying frequency slope characteristics to obtain a time-frequency change curve; Calculating a derivative of the time-frequency change curve to generate a frequency change rate curve; Calculating a maximum slope absolute value of the frequency change rate curve to calculate a minimum sampling rate of the interference signal according to the maximum slope absolute value, and taking the minimum sampling rate as a sampling rate configuration parameter; Extracting a minimum frequency value and a maximum frequency value from the time-frequency change curve, and calculating a passband range of the anti-interference transmission signal based on the minimum frequency value and the maximum frequency value, and taking the passband range as a band-pass filtering range.

7. A computing device, comprising: The storage component stores one or more computer instructions; the one or more computer instructions are used to be called and executed by the processing component to implement the radar signal real-time pulse compression method based on the Transformer according to any one of claims 1-5.

8. A computer storage medium, characterized in that The computer program is stored in the computer and is executed by the computer to implement the radar signal real-time pulse compression method based on the Transformer according to any one of claims 1-5.

Images