Radar target detection method and device based on compressed sensing

Abstract

The invention provides a radar target detection method and device based on compressed sensing. The radar target detection method comprises the following steps: acquiring a radar echo signal corresponding to a radar transmitting signal; determining the standard echo information of the to-be-measured target on a plurality of preset distance units according to the characteristics of the radar emission signal; according to the radar echo signal and the standard echo information, a convex optimization problem based on the compressed sensing theory is solved, a convex optimization intermediate result is obtained, and the convex optimization intermediate result comprises position information of the to-be-measured target and initial signal amplitude information; obtaining a compensation coefficient corresponding to the convex optimization intermediate result according to the convex optimization intermediate result and the standard echo information; correcting the initial signal amplitude information according to the compensation coefficient to obtain corrected signal amplitude information; and obtaining a radar detection result of the to-be-detected target according to the position information of the to-be-detected target and the corrected signal amplitude information. According to the invention, the problem of poor reconstruction precision of the signal amplitude in the prior art can be solved.

Description

Technical Field

[0001] The present application relates to the field of radar signal processing technology, and in particular to a radar target detection method and device based on compressed sensing. Background Art

[0002] Compressed sensing (CS) technology has been a significant breakthrough in signal processing in recent years. Since its introduction in 2006 by researchers such as Candes and Donoho, this method has garnered widespread attention for its exceptional performance in sparse signal reconstruction. CS models signal reconstruction as a sparse optimization problem and employs techniques such as L1 norm minimization to achieve high-precision signal reconstruction from limited observational data. In radar systems, the introduction of CS provides a new approach to complex signal processing. This is particularly true for linear frequency modulation (LFM) signals, such as chirp signals, which exhibit linear frequency modulation characteristics. Traditional algorithms are limited to linear processing frameworks. However, CS technology, through a sparse representation of target echoes, accurately reconstructs target information, including parameters such as range, velocity, and intensity, while suppressing sidelobes. Its nonlinear characteristics significantly mitigate the sidelobe effects in chirp signal processing, laying a theoretical foundation for improving radar system detection capabilities.

[0003] Traditional matched filtering algorithms, when processing chirp signals, suffer from significant sidelobe problems. Target information can be easily obscured by strong clutter or the sidelobes of other targets, resulting in reduced target detection performance. Compressed sensing methods, by contrast, excel at suppressing sidelobes, but their nonlinear processing characteristics also present new challenges. While some commonly used sparse optimization methods can effectively reconstruct the position of sparse signals, their estimated amplitudes differ from the true signal, resulting in poor signal amplitude reconstruction accuracy. To address this issue, previous studies have attempted to analyze the statistical properties of compressed sensing results to achieve higher-precision sparse reconstruction. However, most of these methods are based on Fourier or Gaussian measurement matrices, and thus remain limited when used in complex signal scenarios. This has hindered the practical application of compressed sensing technology in chirp signal processing. Summary of the Invention

[0004] The present application provides a radar target detection method and device based on compressed sensing, which is used to solve the defect of poor signal amplitude reconstruction accuracy in the sparse optimization method in the prior art, and can effectively improve the accuracy of sparse reconstruction of the original signal, thereby improving the radar detection accuracy.

[0005] This application provides a radar target detection method based on compressed sensing, comprising the following steps: Obtain the radar echo signal returned by the radar transmission signal after reaching the target to be measured; Determining standard echo information of the target to be measured at a plurality of preset distance units according to the characteristics of the radar transmission signal; Solving a convex optimization problem based on compressed sensing theory according to the radar echo signal and the standard echo information to obtain a convex optimization intermediate result, wherein the convex optimization intermediate result includes position information of the target to be measured and initial signal amplitude information; Obtaining a compensation coefficient corresponding to the convex optimization intermediate result according to the convex optimization intermediate result and the standard echo information; Correcting the initial signal amplitude information according to the compensation coefficient to obtain corrected signal amplitude information; A radar detection result of the target to be detected is obtained according to the position information of the target to be detected and the corrected signal amplitude information.

[0006] According to the radar target detection method based on compressed sensing provided by the present application, obtaining the radar detection result of the target to be detected based on the position information of the target to be detected and the corrected signal amplitude information includes: Determining the noise intensity of the target to be measured in the radar space according to the compensation coefficient, the standard echo information, the convex optimization intermediate result, and the radar echo signal; Determining a detection threshold based on the noise intensity and a preset false alarm rate; The position information of the target to be detected and the corrected signal amplitude information are screened according to the detection threshold to obtain a radar detection result of the target to be detected.

[0007] According to the radar target detection method based on compressed sensing provided by the present application, the position information of the target to be detected and the corrected signal amplitude information are screened according to the detection threshold to obtain the radar detection result of the target to be detected, including: Determining a test statistic corresponding to the target to be measured at each of the distance units according to the position information of the target to be measured and the corrected signal amplitude information; The detection statistics are screened according to the detection threshold to obtain a target detection statistic, wherein the probability that the target to be detected exists in the distance unit corresponding to the target detection statistic is greater than a preset probability; According to the target detection statistic, a radar detection result of the target to be detected is determined from the position information of the target to be detected and the corrected signal amplitude information.

[0008] According to the radar target detection method based on compressed sensing provided by the present application, the radar transmission signal is represented by the following formula (1), the standard echo information is recorded by an observation matrix, and each element in the observation matrix is ​​determined by the following formula (2): (1) (2) in, is the carrier frequency, is the linear modulation frequency, , is the pulse width, is the bandwidth, is the row number of the element in the observation matrix, is the column number of the element in the observation matrix, is the pulse repetition interval, is the sampling rate, is the time interval between two samples.

[0009] According to the radar target detection method based on compressed sensing provided by the present application, the radar echo signal includes a plurality of to-be-processed segments, which are obtained by pre-dividing the radar echo signal into different processing units based on a preset coherent processing time, and then dividing each processing unit into different processing units based on a pulse repetition interval of the radar transmit signal. One to-be-processed segment corresponds to radar echo information of one pulse, and the convex optimization intermediate result is calculated by the following formulas (3) and (4): (3) (4) in, is the intermediate result of convex optimization, is the radar echo information corresponding to a single segment to be processed, is the observation matrix, for The corresponding solution vector, is the total number of distance units, is the regularization parameter of the sparse penalty term in the convex optimization problem, |·| is the modulus of the complex number, and arg represents the optimal solution vector.

[0010] According to the radar target detection method based on compressed sensing provided by the present application, obtaining the compensation coefficient corresponding to the convex optimization intermediate result according to the convex optimization intermediate result and the standard echo information includes: According to the convex optimization intermediate result and the standard echo information, the compensation coefficient corresponding to the convex optimization intermediate result is obtained by the following formula (5) and formula (6); (5) (6) in, is the compensation coefficient, is the number of columns of the observation matrix, is the number of rows in the observation matrix, is the intermediate result of convex optimization, is the regularization parameter of the sparse penalty term in the convex optimization problem, is the normalized observation matrix, eig( ) represents the eigenvalue of the Gram matrix of the normalized observation matrix A. It is the first intermediate variable in the calculation process.

[0011] According to the radar target detection method based on compressed sensing provided by the present application, the initial signal amplitude information is corrected according to the compensation coefficient to obtain the corrected signal amplitude information, including: According to the compensation coefficient, the initial signal amplitude information is corrected by the following formula (7) to obtain the corrected signal amplitude information; (7) in, is the intermediate result of the modified convex optimization, contains the corrected signal amplitude information, is the compensation coefficient, is the normalized observation matrix, is the intermediate result of convex optimization, is the normalized radar echo information corresponding to a single segment to be processed.

[0012] According to the radar target detection method based on compressed sensing provided by the present application, determining the noise intensity of the target to be detected in the radar space according to the compensation coefficient, the standard echo information, the convex optimization intermediate result, and the radar echo signal includes: According to the compensation coefficient, the standard echo information, the convex optimization intermediate result and the radar echo signal, the noise intensity of the target to be measured in the radar space is determined by the following formula (8) and formula (9); (8) (9) in, is the noise intensity of the target to be measured in the radar space, is the compensation coefficient, is the normalized observation matrix, is the intermediate result of convex optimization, is the radar echo information corresponding to the normalized single to-be-processed segment, is the number of rows of the normalized observation matrix, is the number of columns of the normalized observation matrix, is the standard deviation of the noise of the radar echo signal, is a first intermediate variable in the calculation process, is a second intermediate variable in the calculation process, eig( ) represents solving the eigenvalue of the Gram matrix of the normalized observation matrix A.

[0013] According to the radar target detection method based on compressed sensing provided in the application, the detection threshold is determined according to the noise intensity and a preset false alarm rate, which comprises: The detection threshold is determined according to the noise intensity and a preset false alarm rate by the following formula (10) and formula (11); (10) (11) Wherein, is the detection threshold, is the preset false alarm rate, is the inverse cumulative distribution function of the standard normal distribution, is the noise intensity of the to-be-detected target in the radar space.

[0014] The application also provides a radar target detection device based on compressed sensing, which comprises: A first acquisition module is configured to acquire a radar echo signal returned by a radar transmission signal after reaching a to-be-detected target; A determination module is configured to determine standard echo information of the to-be-detected target on a plurality of preset distance units according to a feature of the radar transmission signal; A second acquisition module is configured to solve a convex optimization problem based on compressed sensing theory according to the radar echo signal and the standard echo information, and obtain a convex optimization intermediate result, wherein the convex optimization intermediate result comprises position information and initial signal amplitude information of the to-be-detected target; A third acquisition module is configured to obtain a compensation coefficient corresponding to the convex optimization intermediate result according to the convex optimization intermediate result and the standard echo information; A correction module is configured to correct the initial signal amplitude information according to the compensation coefficient, and obtain corrected signal amplitude information; A fourth acquisition module is configured to obtain a radar detection result of the to-be-detected target according to the position information of the to-be-detected target and the corrected signal amplitude information.

[0015] The present application also provides an electronic device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, a radar target detection method based on compressed sensing as described above is implemented.

[0016] The present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the compressed sensing-based radar target detection method as described above is implemented.

[0017] The present application also provides a computer program product, including a computer program, which, when executed by a processor, implements any of the above-described radar target detection methods based on compressed sensing.

[0018] This application provides a radar target detection method and device based on compressed sensing. Based on compressed sensing technology, the recovered result exhibits a certain degree of sparsity. Compared with traditional matched filtering methods, this method retains the advantages of compressed sensing in recovering the original radar echo signal, effectively suppressing the high sidelobe problem commonly found in traditional processing methods. This not only improves the radar's target resolution capability but also reduces the impact of clutter interference on detection performance, ensuring the stable operation of the radar system. Furthermore, for radar signals (such as CHIRP signals), this application adds a step to compensate for the intermediate results of convex optimization, achieving more accurate target scene reconstruction and effectively reducing amplitude deviations in target parameter estimation, resulting in a recovered result closer to the true one. This improvement significantly improves the accuracy of sparse reconstruction of the original radar echo signal, laying the foundation for achieving higher detection performance. It effectively addresses the problem in related art where some commonly used sparse optimization methods can effectively reconstruct the position of sparse signals, but their estimated results have amplitude errors compared to the true signal, resulting in poor signal amplitude reconstruction accuracy. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the present application or the prior art, a brief introduction will be given below to the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present application. For ordinary technicians in this field, other drawings can be obtained based on these drawings without any creative work.

[0020] Figure 1 This is a flowchart of a radar target detection method based on compressed sensing shown in an embodiment of the present application.

[0021] Figure 2 This is an overall flow chart of a radar target detection method based on compressed sensing shown in an embodiment of the present application.

[0022] Figure 3 This is a schematic diagram of target scene simulation results reconstructed using different methods shown in an embodiment of the present application.

[0023] Figure 4 This is a schematic diagram of the false alarm rate when the method provided in one embodiment of the present application detects a target scene.

[0024] Figure 5 This is a structural block diagram of a radar target detection device based on compressed sensing shown in an embodiment of the present application.

[0025] Figure 6 This is a schematic diagram of the physical structure of an electronic device shown in an embodiment of the present application. DETAILED DESCRIPTION

[0026] To make the objectives, technical solutions, and advantages of this application more clear, the technical solutions of this application will be clearly and completely described below in conjunction with the accompanying drawings. Obviously, the embodiments described are only part of the embodiments of this application, not all of them. Based on the embodiments of this application, all other embodiments obtained by ordinary technicians in this field without making any creative efforts are within the scope of protection of this application.

[0027] Traditional matched filtering algorithms, when processing chirp signals, suffer from significant sidelobe problems. Target information can easily be obscured by strong clutter or the sidelobes of other targets, resulting in reduced target detection performance. Compressed sensing methods, by contrast, excel at suppressing sidelobes, but their nonlinear processing characteristics also present new challenges. First, while some commonly used sparse optimization methods can effectively reconstruct the position of sparse signals, their estimated amplitudes differ from the true signal, resulting in poor signal amplitude reconstruction accuracy. Second, compressed sensing estimation results often lack a clear analytical expression, and the probability distribution of target parameters is unknown, making it difficult to control false alarm rates. To address these issues, previous studies have attempted to analyze the statistical properties of compressed sensing results to achieve higher-precision sparse reconstruction. However, these methods, which mostly rely on Fourier or Gaussian measurement matrices, still have limitations when dealing with complex signal scenarios, hindering the practical application of compressed sensing technology in chirp signal processing.

[0028] The following will describe step by step how the solution of this application solves the two problems mentioned above.

[0029] Figure 1 This is a flow chart of a radar target detection method based on compressed sensing shown in an embodiment of the present application. Figure 1 The radar target detection method based on compressed sensing of the present application may include the following steps: Step 101: Acquire a radar echo signal returned by a radar transmission signal after reaching a target to be measured.

[0030] Among them, the radar transmission signal mainly refers to the Chirp signal.

[0031] In this application, after obtaining the radar echo signal, some preprocessing operations need to be performed on the radar echo signal, including the following steps Step 1a to Step 1d: Step 1a: The radar echo signal received by the radar receiver is divided into different processing units according to a preset coherent processing time.

[0032] Among them, the coherent processing interval (CPI), also known as the coherent accumulation time, refers to the period of time during which the radar system continuously transmits a series of pulses and can maintain a stable phase relationship between these pulses.

[0033] For example, to detect a long-range, slow-flying aircraft (the target), a radar system transmits a chirp pulse every 1 millisecond (ms). The radar system's coherent processing time is set to 100 ms. Then, during step 1a, the received continuous, original analog echo signal can be segmented into 100-ms segments to produce multiple processing units, each equivalent to a data block.

[0034] Step 1b: For the radar echo signal within the same coherent processing time, divide it into different segments to be processed using the pulse interval of the transmitted signal as the time unit.

[0035] Step 1b is to separate the radar echo signal according to the period of a single pulse within a data block corresponding to a coherent processing time, and obtain independent echo data segments corresponding to each transmitted pulse.

[0036] Pulse interval refers to the Pulse Repetition Interval (PRI) or Pulse Repetition Period (PRP), which is the time interval from the start of the radar transmitting one pulse to the start of the next pulse.

[0037] For example, for a 100-millisecond data block (containing all echo signals received within 100 milliseconds), execute step 1b, starting from millisecond 0 of the data block and truncating to millisecond 1 to obtain the first segment to be processed. This segment corresponds to the echo received within 1 millisecond after the radar transmits the first pulse at millisecond 0. Next, the segment to be processed is started from millisecond 1 and ends at millisecond 2 to obtain the second segment to be processed, corresponding to the echo of the second transmitted pulse. Repeat these steps to obtain multiple segments to be processed. Ultimately, the 100-millisecond processing unit is divided into 100 segments to be processed, each of 1 millisecond length.

[0038] Step 1c: The echo information of all pulse segments is stored in the form of complex numbers in MATLAB as the raw data for one-time processing.

[0039] The purpose of step 1c is to digitize and structure all segmented and discretized radar echo segments to form a two-dimensional data matrix that can be directly accessed by the computer, hereinafter referred to as the original echo data matrix.

[0040] Specifically, each segment to be processed is sampled at high speed and converted into a series of discrete digital points. Then, each sampling point is represented by a complex number. This representation method can encode the amplitude of the signal in one number at the same time. and phase Then, all the complex sampling data of the segments to be processed are organized into a two-dimensional data matrix in the MATLAB computing environment.

[0041] For example, through step 1b, 100 segments to be processed are obtained, each with a length of 1 millisecond. For the first 1 millisecond segment to be processed, it is sampled through the analog-to-digital converter to obtain 10,000 discrete sampling points, each of which is represented as a complex number. In this way, the first segment becomes a one-dimensional array (or vector) containing 10,000 complex numbers. Repeat the same operation for the remaining 99 segments, and finally obtain 100 such one-dimensional arrays, each containing 10,000 complex numbers. Finally, create a two-dimensional matrix in MATLAB to store this data, whose dimensions are .

[0042] In the raw echo data matrix, rows represent pulse numbers, and columns represent sampling moments (sampling points). That is, each row of data corresponds to the echo of an independent pulse transmission event, and each column of data corresponds to a sampling time point in the process of receiving a single echo.

[0043] Step 1d: Use MATLAB's std function to calculate the noise intensity of the radar echo signal, that is, the standard deviation of the noise signal .

[0044] The purpose of step 1d is to statistically analyze the entire original radar echo signal to quantitatively estimate the average power or intensity of the background noise therein, so as to provide a basis for setting key parameters for subsequent compressed sensing algorithms.

[0045] Specifically, for the two-dimensional data matrix obtained in step 1c and stored in MATLAB, the built-in statistical function std in MATLAB is called to calculate the elements in the matrix, and finally a single numerical value (scalar) is obtained, which represents the standard deviation (noise intensity) of the noise signal in the radar echo signal within the current coherent processing time.

[0046] Step 102, according to the characteristics of the radar transmitting signal, determine the standard echo information of the target to be measured on a plurality of preset distance units.

[0047] In the present application, the CHIRP observation matrix is used to record the standard echo information of the target to be measured on a plurality of preset distance units.

[0048] Specifically, the process of constructing the CHIRP observation matrix A according to the radar transmitting signal can include the following steps 2a-2b: Step 2a: Obtain the parameters of the radar transmitting signal, including the carrier frequency (Carrier Frequency), the center frequency (Center Frequency), the sampling rate (Sampling Rate), the bandwidth (Bandwidth), the pulse signal duty cycle (Duty Cycle), and the pulse repetition interval.

[0049] Step 2b: According to the plurality of transmitting signal parameters obtained in step 2a, construct a compressed sensing observation matrix (i.e. CHIRP observation matrix) corresponding to the transmitting signal, and store it in MATLAB.

[0050] Step 2b is to use the accurate radar parameters obtained in step 2a to generate a digital matrix that can represent the radar sensing ability through mathematical modeling and calculation, and load it into the calculation software to provide a core model for subsequent signal sparse decomposition.

[0051] In the present application, the radar transmitting signal is a linear frequency modulation signal (CHIRP signal), and its waveform is: wherein, represents the carrier frequency, represents the linear frequency or sweep rate, which represents the speed of the frequency change with time, , represents the pulse width, represents the bandwidth. An ideal signal waveform is described as linearly increasing from to in time and frequency.

[0052] wherein the value of any element in the observation matrix A is calculated by the following formula: wherein, represents the row number of the element in the observation matrix, represents the column number of the element in the observation matrix, represents the pulse repetition interval; represents the sampling rate; represents the time interval between two samplings.

[0053] The column vectors of the observation matrix are the transmission signal vectors after different time delays. Each column of the observation matrix is a transmission signal after a time delay, such as the first column is the transmission signal at time 0, the second column is the transmission signal after half a sampling time interval, the third column is the transmission signal after one sampling time interval, and each column is delayed by half a sampling time interval.

[0054] In the observation matrix, the row represents the sampling time, and the column represents the distance unit. In a radar system, time delay and distance are equivalent and can be converted. According to , as long as the time delay of the echo is known, , the speed of light, a specific time delay must correspond to a specific distance. For example, for the second column , the time delay (half a sampling time interval) corresponds to a distance of meters, which indicates that the physical meaning of the second column of data is the echo template of a target located at c x Δt / 4 meters.

[0055] Step 103, according to the radar echo signal and the standard echo information, solving the convex optimization problem based on the compressed sensing theory, obtaining the convex optimization intermediate result, and the convex optimization intermediate result includes the position information and the initial signal amplitude information of the target to be measured.

[0056] In an embodiment, step 103 can include the following steps: Divide the radar echo signal into different processing units in units of a preset coherent processing time; ​​Each processing unit is divided into different to-be-processed segments in units of pulse repetition intervals of radar transmitting signals, and each to-be-processed segment corresponds to radar echo information of one pulse. According to each to-be-processed segment and the standard echo information, a convex optimization problem based on a compressed sensing theory is solved to obtain a convex optimization intermediate result.

[0057] The process of dividing the to-be-processed segments can refer to steps 1a-1c described above.

[0058] In the present application, step 103 is performed, and a convex optimization problem based on a compressed sensing theory is solved according to the original echo data matrix and the observation matrix A to obtain a convex optimization intermediate result. Specifically, the complete process of obtaining the convex optimization intermediate result of the compressed sensing model can include the following steps 3a-3b: Step 3a: according to the calculated noise intensity in the original radar echo data , a suitable regularization parameter of a sparse penalty term of the convex optimization problem is set.

[0059] In the present application, according to the noise intensity obtained in step 1d, the regularization parameter of the sparse penalty term of the convex optimization problem is set.

[0060] Step 3b: using the cvx toolbox of MATLAB, taking the original radar echo data in one pulse interval, the regularization parameter of the sparse penalty term of the convex optimization problem, and the observation matrix as inputs, solving the corresponding convex optimization problem to obtain the convex optimization intermediate result of the compressed sensing signal.

[0061] In this step, according to the data in one row of the original echo data matrix (i.e., the radar echo information corresponding to a single to-be-processed segment) obtained in step 1c, the regularization parameter and the observation matrix , the convex optimization intermediate result of the compressed sensing signal is obtained by the following formula: wherein, is the convex optimization intermediate result, |·| represents the modulus of a complex number, represents the dimension of a single solution vector corresponding to the data in one row , i.e., the total number of distance units, the single solution vector is a column vector, represents taking the optimal solution vector.

[0062] Finally, based on the optimal solution vectors corresponding to all the segments to be processed, the convex optimization intermediate results are obtained. The convex optimization intermediate results can reflect the position information of the target to be measured and the initial signal amplitude information.

[0063] In the convex optimization intermediate result matrix, the rows represent distance units and the columns represent pulse numbers.

[0064] Step 104: Obtain a compensation coefficient corresponding to the convex optimization intermediate result according to the convex optimization intermediate result and the standard echo information.

[0065] Step 105: Correct the initial signal amplitude information according to the compensation coefficient to obtain corrected signal amplitude information.

[0066] In this application, steps 104 and 105 are used to perform secondary processing on the convex optimization intermediate result obtained in step 103 to obtain the target scene reconstruction result. Although the convex optimization intermediate result obtained in step 103 has high resolution and successfully suppresses sidelobes, it suffers from inaccurate signal amplitude estimation (i.e., inaccurate initial signal amplitude information estimation). The goal of steps 104 and 105 is to correct the signal amplitude deviation, thereby obtaining a target scene reconstruction result that is closer to the actual situation in terms of both position and signal amplitude.

[0067] Specifically, the process of performing secondary processing on the intermediate result of convex optimization to obtain the target scene reconstruction result may include the following steps 4a to 4d: Step 4a: Using MATLAB Function, calculates the eigenvalue distribution and eigenvector of the Gram matrix (Gram matrix) of the column vector of the observation matrix.

[0068] Step 4a calculates the eigenvalues ​​and eigenvectors of the Gram matrix (AᴴA) of the observation matrix A. Aᴴ is the conjugate transpose of A. The Gram matrix AᴴA describes the similarity between different columns in the observation matrix A. In other words, each element in the Gram matrix (i.e. Row, No. ), equal to the element of the column of the A matrix Column vector and The inner product of column vectors is used to measure the similarity between two vectors.

[0069] Step 4b: Normalize the measurement matrix and the original radar echo data according to the maximum eigenvalue, so that the maximum eigenvalue of the Gram matrix of the column vector of the normalized measurement matrix is ​​1.

[0070] In step 4b, by dividing the observation matrix A and the original radar echo data y in the original echo data matrix by the square root of the maximum eigenvalue of the Gram matrix, the maximum eigenvalue of the corresponding Gram matrix (AnormᴴAnorm) of the obtained normalized observation matrix is ​​exactly equal to 1. This makes the subsequent algorithm (the solution of step 4c) more stable, converges faster, and reduces sensitivity to parameters.

[0071] Step 4c: Substitute the compensation coefficient calculation formula into the convex optimization intermediate results and the sparse penalty term parameters of the convex optimization problem. , the normalized measurement matrix, the eigenvalue of the Gram matrix of the normalized measurement matrix, solve the equations, and obtain the compensation coefficient of the convex optimization result.

[0072] Specifically, the compensation coefficients are obtained by solving the following fixed point equations: : in, is the number of columns of the normalized observation matrix A, is the number of rows of the normalized observation matrix A, For convex optimization intermediate results, is the regularization parameter of the sparse penalty term for the convex optimization problem, It means to solve the eigenvalue of the Gram matrix of the normalized observation matrix A, which can be obtained through the matlab function Solve it.

[0073] where s The definition is as follows: Step 4d: According to the target scene calculation formula, substitute the original radar echo data, the normalized observation matrix, the compensation coefficient, and the convex optimization intermediate result to obtain the reconstructed new observation scene.

[0074] Specifically, the calculation formula is as follows: in, is the compensation coefficient, is the normalized observation matrix, For convex optimization intermediate results, is the normalized raw radar echo data.

[0075] Steps 4a to 4d are the key bridge between preliminary sparse reconstruction and reliable target detection. By analyzing the intrinsic characteristics of the observation matrix A and solving a precise compensation coefficient , performing secondary refinement on the convex optimization intermediate result with amplitude deviation obtained in the previous step can correct the target signal amplitude and restore it to a close-to-true amplitude. At the same time, it makes the noise characteristics of the final result analyzable, providing a technical basis for subsequently solving the problem of difficulty in constant false alarm detection.

[0076] In this application, the target scene refers to the digital description of all the target information to be detected in the radar detection space, which uses a vector to fully describe the locations of the targets within the detection range, as well as the intensity and phase information of these targets. The target scene calculated in step 4d is This is the target scene vector. The new reconstructed vector obtained in this step is That is, a more accurate and realistic radar map.

[0077] Step 106: Obtain a radar detection result of the target to be detected based on the position information of the target to be detected and the corrected signal amplitude information.

[0078] In this application, the target scene vector calculated in step 4d is contains the position information of the target to be measured and the corrected signal amplitude information, so step 106 is executed according to the target scene vector Get the radar detection result of the target to be detected.

[0079] This application is based on compressed sensing technology, and the recovered results after processing have a certain degree of sparsity. Compared with traditional matched filtering methods, this application retains the advantages of compressed sensing methods for the recovery of original radar echo signals, namely, it effectively suppresses the high sidelobe problem commonly found in traditional processing methods. This not only improves the radar's target resolution capability, but also reduces the impact of clutter interference on detection performance, providing a guarantee for the stable operation of the radar system. In addition, for CHIRP transmission signals, this application adds a step to compensate for the intermediate results of convex optimization, achieving more accurate target scene reconstruction, effectively reducing the amplitude deviation in target parameter estimation, and obtaining a recovery result that is closer to the true result. This improvement significantly improves the accuracy of sparse reconstruction of original radar echo signals in the case of Chirp signals, laying the foundation for achieving higher detection performance. It can effectively solve the first problem mentioned in the related art, namely that although some commonly used sparse optimization methods can effectively reconstruct the position of sparse signals, their estimated results have errors in signal amplitude compared to the true signal, resulting in poor reconstruction accuracy of the signal amplitude.

[0080] In combination with the above embodiments, in one implementation, step 106 may include: Step 1061: Determine the noise intensity of the target to be measured in the radar space according to the compensation coefficient, the standard echo information, the convex optimization intermediate result, and the radar echo signal.

[0081] In this application, the noise intensity of the target to be measured in the radar space is determined, that is, the noise level of the target scene is calculated. The complete process includes the following step 5a: Step 5a: According to the scene noise calculation formula and the scene noise intermediate variable calculation formula, use the previously obtained results as input to calculate the noise level of the target scene, that is, the standard deviation of the overall noise (standard deviation of target scene noise).

[0082] The scene noise calculation formula is as follows: in, is the compensation coefficient, is the normalized observation matrix, For convex optimization intermediate results, is the normalized radar echo information corresponding to a single segment to be processed, is the number of rows of the normalized observation matrix, is the number of columns of the normalized observation matrix, is the standard deviation of the noise of the radar echo signal obtained in step 1d, It is another variable in the equation group of the compensation coefficient calculation formula, which is called the first intermediate variable in the calculation process. It is the intermediate variable of scene noise, which is called the second intermediate variable in the calculation process in this paper.

[0083] The calculation formula of the intermediate variable of scene noise is as follows: in, is the compensation coefficient, is the number of columns of the normalized observation matrix, is the first intermediate variable in the calculation process, eig( ) represents the eigenvalue of the Gram matrix of the normalized observation matrix A. is the normalized measurement matrix. The core purpose of step 5a is to accurately and quantitatively calculate the final value after the complex transformation and correction of steps 4a to 4d. The equivalent intensity (standard deviation) of the background noise in .

[0084] In order to solve the problem that it is difficult to set the detection threshold of traditional compressed sensing results due to unknown statistical characteristics, step 5a uses rigorous mathematical formulas to accurately derive the equivalent noise intensity in the final target scene after nonlinear reconstruction and correction. This result provides a key quantitative basis for setting a scientific detection threshold with a constant false alarm rate characteristic in the subsequent steps, thus solving the core technical problem of the difficulty in controlling false alarms in compressed sensing radar in engineering applications.

[0085] Step 1062: Determine a detection threshold based on the noise intensity and a preset false alarm rate.

[0086] In the present application, the step of calculating the detection threshold may include step 6a: Step 6a: Substitute the preset false alarm rate into the threshold calculation formula , calculate the value of the detection threshold under this false alarm rate.

[0087] The threshold calculation formula is as follows: in, is the detection threshold, Indicates the set false alarm rate, Represents the inverse cumulative distribution function of the standard normal distribution, which is defined as follows: in, It represents the noise level of the target scene, that is, the noise intensity of the target to be measured in the radar space.

[0088] In step 6a, the preset performance indicator (false alarm rate ), and the final scene noise level calculated accurately in step 5a Combined with statistical formulas, a scientific detection threshold is calculated , which can ensure that the detection system has a constant false alarm characteristic and provide an objective and quantitative basis for the final target judgment.

[0089] Step 1063: Filter the position information of the target to be measured and the corrected signal amplitude information according to the detection threshold to obtain a radar detection result of the target to be measured.

[0090] Execute step 1063, according to the detection threshold , the target scene vector containing the position information of the target to be measured and the corrected signal amplitude information (calculated in step 4d) is screened to obtain the radar detection result of the target to be detected.

[0091] Specifically, step 1063 may include: Step A: Determine the test statistic corresponding to the target to be measured at each distance unit based on the position information of the target to be measured and the corrected signal amplitude information.

[0092] In this application, the is a complex vector. Directly comparing this complex vector with the single detection threshold (a real number) calculated later is not feasible. Therefore, a test statistic must be calculated first. This statistic is a scalar (a single number) that comprehensively reflects the confidence (probability) of the target's presence or the signal energy strength at a specific range cell.

[0093] Specifically, the process of calculating the test statistic may include the following step 7a: Step 7a: According to the test statistic calculation formula, obtain the test statistic of each detection unit to form a test statistic vector.

[0094] The test statistic calculation formula is as follows: in, Represents the target scene vector. For the calculation formula of complex vector modulus, please refer to the previous article for details.

[0095] In this application, a detection unit refers to a basic, discrete distance unit. If the detection space is divided into 300 distance units, then there are 300 detection units, which is the same as The test statistic is a confidence score that quantifies the probability of the target existing at a specific detection unit (range unit). The test statistic of each detection unit refers to the specific, non-negative real number score corresponding to each range unit and representing the signal energy at that position. The test statistic for each detection unit is .

[0096] Specifically, first calculate the first detection unit and get the test statistic , then calculate the second detection unit and get , repeat this process for all N detection units, and get N independent , , ..., Finally, , , ..., , combined in sequence to form a test statistic vector.

[0097] Step B: Filter the detection statistics according to the detection threshold to obtain the target detection statistics. The probability that the target to be detected exists in the distance unit corresponding to the target detection statistics is greater than the preset probability.

[0098] Step C: Determine the radar detection result of the target to be detected based on the target detection statistics, the position information of the target to be detected and the corrected signal amplitude information.

[0099] In this application, steps B-C are used to implement target detection, and the entire process includes the following steps 8a-8b: Step 8a: Take the test statistic vector and the detection threshold as input, traverse the test statistic vector, and compare each element of the vector (test statistic) with the detection threshold. If the test statistic is greater than the detection threshold, it is considered that there is a target at that point (indicating that the probability of the target to be detected in the corresponding distance unit is greater than the preset probability), so the sequence number of the element is recorded.

[0100] In step 8a, each element in the test statistic vector, i.e. The test statistic for each test unit , and the detection threshold For comparison, if , indicating that There is a target on each detection unit, and then the system will record the serial number of this detection unit. .like , then the detection unit is determined to have only noise and is directly ignored. Finally, the detection result is a list containing the serial numbers of all detection units that are determined to have targets.

[0101] Step 8b: According to the parameters of the radar transmission signal, convert the element number of the detection result to the corresponding target distance and output the detected target distance.

[0102] After determining the vector element numbers for all detected targets, step 8b converts these abstract, internal computer numbers back into real-world measurements with clear physical meaning. Specifically, based on the radar system's preset parameters, such as the precision of the range cells defined when constructing the observation matrix (e.g., each range cell represents 0.5 meters), a simple multiplication or mapping operation is performed on each detected target's number and that precision value. This ultimately generates a clear and intuitive detection report, for example, directly listing detected targets at distances of 25 meters, 91 meters, and so on. This presents the results of the entire complex computational process in a way that operators and subsequent systems can directly understand and use.

[0103] In this application, the final radar detection result includes not only the most basic target position (distance), but also signal amplitude information. Signal amplitude information is crucial for subsequent target classification and identification. In this application, after obtaining a list of the serial numbers of the detection units that are determined to have targets, the results obtained in steps 4a to 4d are obtained according to the list. When obtaining the target distance, convert the sequence number recorded in the list into the physical distance, for example hour, is the distance information (position information) of the target to be measured; when the signal amplitude is obtained, Take out the value of the element corresponding to the sequence number, for example When taking out , the value is a complex number (e.g. ), and then calculate its modulus , this modulus value is the required, corrected, and most accurate signal amplitude.

[0104] In summary, after acquiring radar echo data, this application constructs the relevant CHIRP observation matrix according to the parameters of the radar transmission signal, and by constructing a compressed sensing model, uses the convex optimization method to obtain the convex optimization intermediate result of the compressed sensing signal, and then optimizes this result to obtain the reconstructed image of the target scene. At the same time, the noise level in the scene can be calculated. Then, based on the obtained target scene reconstructed image, the test statistic of the target detection problem is calculated, and the detection threshold is calculated according to the set false alarm rate. The false alarm rate is controlled while the target is detected. Finally, MATLAB is used as the operating platform to output the target detection result. The above whole process is as follows: Figure 2 shown. Figure 2 This is an overall flow chart of a radar target detection method based on compressed sensing shown in an embodiment of the present application. Figure 2 In the example, the radar observation data is the radar echo data, and the target scene reconstruction image is the result obtained from step 4a to step 4d. , the scene noise level is calculated in step 5a , the rest of the information is detailed in the above embodiments.

[0105] The present application will be further described below in conjunction with simulation experiments.

[0106] 1. Simulation conditions: The simulation experiments of this application were carried out on a computer configured with a 13th Gen Intel(R) Core(TM) i9-13900HX, 2200 MHz, 24-core processor, Microsoft Windows 11 operating system, and an NVIDIA GeForceRTX 4060 Laptop GPU; the configured simulation environment was MATLAB R2023b.

[0107] 2. Simulation content and result analysis: There are two simulation experiments in this application.

[0108] Simulation Experiment 1: When using CHIRP transmission signals, there are two targets in the scene, and the signal-to-noise ratio after accumulation is 20dB and 30dB, three methods are used, namely the present application and two existing technologies (traditional matched filtering method and convex optimization method for solving the compressed sensing model optimization problem), to process the radar single pulse echo respectively and recover the following: Figure 3 The target scene is shown in (a). The specific simulation parameters are as follows: the radar transmission wave waveform is a linear frequency modulation signal, the baseband signal center frequency is 0, the carrier frequency is 1GHz, the bandwidth is 5MHz, the transmission wave pulse width is 200μs, the radar pulse interval is 2ms, the baseband signal sampling rate is 5MHz, and the standard deviation of the Gaussian white noise in the scene is 0.1. The simulation results are shown in Figure 2. Figure 3 (b) Figure 3 (c) Figure 3 (d) shown. Figure 3 This is a schematic diagram of target scene simulation results reconstructed using different methods shown in an embodiment of the present application.

[0109] Figure 3 (b) is the restoration result of the target scene using the traditional matched filtering method, where Figure 2 (b) shows the range unit on the abscissa, and the signal amplitude on the ordinate, i.e., the estimated target scattering intensity. The cursors in the figure mark the specific range unit numbers of the two targets and the amplitude modulo the recovered scattering intensity.

[0110] Figure 3 (c) is the convex optimization result recovered using the general compressed sensing method, where Figure 2 The horizontal axis of (c) represents the range unit, and the vertical axis represents the signal amplitude, which is the estimated target scattering intensity. The cursors in the figure mark the specific range unit numbers of the two targets and the amplitude of the recovered scattering intensity modulo.

[0111] Figure 3 (d) is the restoration result using the improved target scene reconstruction technology proposed in this application, wherein, Figure 3 The horizontal axis of (d) represents the range unit, and the vertical axis represents the signal amplitude, i.e., the estimated target scattering intensity. The cursors in the figure mark the specific range unit numbers of the two targets and the amplitude of the recovered scattering intensity modulo.

[0112] Depend on Figure 3 (b) It can be seen that the target scene recovery result using the traditional matched filtering method is compared with the original target scene in Figure 3(a). The two have the same range unit number and the amplitude of the scattering intensity is basically the same, but it can be seen that there are large sidelobes near the target point. The presence of sidelobes may mask weak target signals and seriously affect the detection and resolution performance of the radar.

[0113] Depend on Figure 3 (c) It can be seen that the convex optimization result restored using the general compressed sensing method is compared with the original target scene Figure 3(a). The two have the same distance unit number. Although there are no sidelobes near the target point, the amplitude of the scattering intensity of the convex optimization result is significantly smaller than that of the original target. There is a certain deviation in the estimation of the target scattering intensity, which will significantly reduce the target detection performance and affect the radar performance.

[0114] Depend on Figure 3 (d) It can be seen that the restored result obtained using the technology proposed in this application is compared with the original target scene Figure 3(a). The two have the same distance unit number, the amplitude of the target scattering intensity is basically the same, and there are no side lobes near the target point. The target in the reconstructed target scene still maintains good sparsity. Compared with other methods, this application can effectively suppress the side lobes generated by traditional methods and have a more accurate estimate of the target amplitude.

[0115] Simulation Experiment 2: In the same scenario as Simulation Experiment 1, the method of this application is used to process the radar single pulse echo and then detect it. Through 10,000 Monte Carlo experiments, the actual false alarm rate of target detection under different false alarm rate settings is statistically calculated. The specific radar simulation parameters are consistent with those of Simulation Experiment 1, and the false alarm rate is set from 10-5 to 1. The simulation results are as follows: Figure 4 shown.

[0116] Figure 4 This is a schematic diagram of the false alarm rate when the method provided in one embodiment of the present application detects the target scene. Figure 4 The horizontal axis represents the false alarm rate set during detection, and the vertical axis represents the actual false alarm rate calculated in the Monte Carlo experiment. Figure 4 Ref represents a reference straight line passing through the origin and with a slope of 1.

[0117] Depend on Figure 4 It can be seen that when the improved detection method proposed in this application is used for detection, the relationship curve between the actual false alarm rate and the set false alarm rate is basically a straight line passing through the origin and with a slope of 1, indicating that the value of the actual false alarm rate is basically consistent with the set false alarm rate, which further reflects that the proposed detection method can effectively control the false alarm rate of detection.

[0118] In summary, the results obtained from the two simulation experiments demonstrate that this application not only effectively suppresses the sidelobes generated by traditional matched filtering methods during signal processing, but also accurately reconstructs the target's scattering intensity. Furthermore, the detector exhibits excellent flexibility, allowing precise control of the actual false alarm rate according to settings. These characteristics demonstrate that this application has significant application value in improving the detection performance and reliability of radar systems, particularly in weak target detection and high-precision signal reconstruction in complex environments, demonstrating significant technical advantages and broad application prospects.

[0119] This application aims at the processing requirements of Chirp signals, combines compressed sensing technology, and designs a detector with controllable false alarm rate by analyzing the statistical distribution of compressed sensing estimators. It can achieve accurate detection of weak targets while effectively suppressing clutter and sidelobes. This application deeply combines the convex optimization idea with radar signal processing in compressed sensing technology, solving the problem that traditional methods cannot effectively control the false alarm rate, that is, solving the second problem in the prior art mentioned above. In addition, this application proposes an improvement step for the convex optimization results suitable for Chirp signal scenarios, and further optimizes the characteristics of the radar observation matrix. Compared with the traditional matched filtering algorithm, this application has significantly improved the target detection capability, sidelobe suppression performance and computational efficiency, providing important technical support for the application of linear frequency modulation Chirp signals in radar systems, and has broad practical application prospects.

[0120] In summary, the solution provided by this application has the following beneficial effects compared with the prior art: First, this application is based on compressed sensing technology, and the processed recovery results have a certain degree of sparsity. Compared with traditional matched filtering methods, the recovery of the original radar echo signal retains the advantages of compressed sensing methods, namely, it effectively suppresses the high sidelobe problem that is common in traditional processing methods. This not only improves the radar's target resolution ability, but also reduces the impact of clutter interference on detection performance, providing a guarantee for the stable operation of the radar system.

[0121] Second, for CHIRP emission signals, this application adds a computational step to compensate for the convex optimization results, achieving more accurate target scene reconstruction and effectively reducing the amplitude deviation in target parameter estimation, resulting in a recovery result closer to the true one. This improvement significantly improves the accuracy of sparse reconstruction of the original signal in the case of Chirp signals, laying the foundation for achieving higher detection performance.

[0122] Third, for CHIRP emission signals, this application improves the convex optimization intermediate results restored by the compressed sensing convex optimization method, and can analytically deduce the noise level in the target scene, so that the detection threshold can be adjusted according to the noise situation to achieve target detection with controllable false alarm rate. Compared with traditional convex optimization result detection, the improved detection method proposed in this application has the advantages of high detection rate and controllable false alarm rate.

[0123] The following describes a radar target detection device based on compressed sensing provided by the present application. The radar target detection device based on compressed sensing described below and the radar target detection method based on compressed sensing described above can refer to each other.

[0124] Figure 5 This is a block diagram of a radar target detection device based on compressed sensing shown in an embodiment of the present application. Figure 5 The radar target detection device 500 based on compressed sensing of the present application may include: The first acquisition module 501 is used to acquire the radar echo signal returned by the radar transmission signal after reaching the target to be measured; A determination module 502 is configured to determine standard echo information of the target to be detected at a plurality of preset distance units according to the characteristics of the radar transmission signal; A second acquisition module 503 is configured to solve a convex optimization problem based on compressed sensing theory according to the radar echo signal and the standard echo information to obtain a convex optimization intermediate result, wherein the convex optimization intermediate result includes the position information of the target to be measured and the initial signal amplitude information; A third acquisition module 504 is configured to obtain a compensation coefficient corresponding to the convex optimization intermediate result according to the convex optimization intermediate result and the standard echo information; A correction module 505 is configured to correct the initial signal amplitude information according to the compensation coefficient to obtain corrected signal amplitude information; The fourth acquisition module 506 is configured to obtain a radar detection result of the target to be detected based on the position information of the target to be detected and the corrected signal amplitude information.

[0125] According to a radar target detection device 500 based on compressed sensing provided by this application, the fourth acquisition module 506 includes: A first determination submodule is configured to determine the noise intensity of the target to be measured in the radar space according to the compensation coefficient, the standard echo information, the convex optimization intermediate result, and the radar echo signal; A second determination submodule is used to determine a detection threshold according to the noise intensity and a preset false alarm rate; The first screening submodule is configured to screen the position information of the target to be detected and the corrected signal amplitude information according to the detection threshold to obtain a radar detection result of the target to be detected.

[0126] According to a radar target detection device 500 based on compressed sensing provided by the present application, the first screening submodule includes: A third determining submodule is configured to determine a test statistic corresponding to the target to be measured at each of the distance units according to the position information of the target to be measured and the corrected signal amplitude information; A second screening submodule is configured to screen the detection statistics according to the detection threshold to obtain a target detection statistic, wherein the probability that the target to be detected exists in the distance unit corresponding to the target detection statistic is greater than a preset probability; The fourth determination submodule is configured to determine, based on the target detection statistic, the radar detection result of the target to be detected and the corrected signal amplitude information.

[0127] According to a radar target detection device 500 based on compressed sensing provided by the present application, the radar transmission signal is represented by the following formula (1), the standard echo information is recorded using an observation matrix, and each element in the observation matrix is ​​determined by the following formula (2): (1) (2) in, is the carrier frequency, is the linear modulation frequency, , is the pulse width, is the bandwidth, is the row number of the element in the observation matrix, is the column number of the element in the observation matrix, is the pulse repetition interval, is the sampling rate, is the time interval between two samples.

[0128] According to a radar target detection device 500 based on compressed sensing provided by the present application, the radar echo signal includes a plurality of to-be-processed segments, which are obtained by pre-dividing the radar echo signal into different processing units based on a preset coherent processing time, and then dividing each processing unit into different processing units based on a pulse repetition interval of the radar transmit signal. One to-be-processed segment corresponds to radar echo information of one pulse, and the convex optimization intermediate result is calculated by the following formulas (3) and (4): (3) (4) in, is the intermediate result of convex optimization, is the radar echo information corresponding to a single segment to be processed, is the observation matrix, for The corresponding solution vector, is the total number of distance units, is the regularization parameter of the sparse penalty term in the convex optimization problem, |·| is the modulus of the complex number, and arg represents the optimal solution vector.

[0129] According to a radar target detection device 500 based on compressed sensing provided by this application, the third acquisition module 504 includes: An acquisition submodule is used to obtain a compensation coefficient corresponding to the convex optimization intermediate result according to the convex optimization intermediate result and the standard echo information through the following formulas (5) and (6); (5) (6) in, is the compensation coefficient, is the number of columns of the normalized observation matrix, is the number of rows of the normalized observation matrix, is the intermediate result of convex optimization, is the regularization parameter of the sparse penalty term in the convex optimization problem, is the normalized observation matrix, eig( ) represents the eigenvalue of the Gram matrix of the normalized observation matrix A. It is the first intermediate variable in the calculation process.

[0130] According to a radar target detection device 500 based on compressed sensing provided by this application, the correction module 505 includes: A correction submodule, configured to correct the initial signal amplitude information according to the compensation coefficient by using the following formula (7) to obtain corrected signal amplitude information; (7) in, is the intermediate result of the modified convex optimization, contains the corrected signal amplitude information, is the compensation coefficient, is the normalized observation matrix, is the intermediate result of convex optimization, is the normalized radar echo signal corresponding to a single segment to be processed.

[0131] According to a radar target detection device 500 based on compressed sensing provided by this application, the first determination submodule includes: A fifth determination submodule is configured to determine a detection threshold according to the intensity of the noise and a preset false alarm rate using the following formulas (10) and (11); (10) (11) in, is the detection threshold, is the preset false alarm rate, is the inverse cumulative distribution function of the standard normal distribution, is the noise intensity of the target to be measured in the radar space.

[0132] Figure 6 FIG. 1 is a schematic diagram of the physical structure of an electronic device according to an embodiment of the present application. Figure 6 As shown, the electronic device may include: a processor 610 , a communications interface 620 , a memory 630 and a communication bus 640 , wherein the processor 610 , the communications interface 620 and the memory 630 communicate with each other via the communication bus 640 . The processor 610 can call the logic instructions in the memory 630 to execute a radar target detection method based on compressed sensing, which includes: obtaining a radar echo signal returned by a radar transmission signal after reaching a target to be measured; determining the standard echo information of the target to be measured at multiple preset distance units based on the characteristics of the radar transmission signal; solving a convex optimization problem based on compressed sensing theory based on the radar echo signal and the standard echo information to obtain a convex optimization intermediate result, wherein the convex optimization intermediate result includes the position information and initial signal amplitude information of the target to be measured; obtaining a compensation coefficient corresponding to the convex optimization intermediate result based on the convex optimization intermediate result and the standard echo information; correcting the initial signal amplitude information based on the compensation coefficient to obtain corrected signal amplitude information; and obtaining a radar detection result of the target to be measured based on the position information of the target to be measured and the corrected signal amplitude information.

[0133] In addition, the logical instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when sold or used as an independent product. Based on this understanding, the technical solution of the present application, or the part that contributes to the existing technology, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the method described in each embodiment of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, and other media that can store program code.

[0134] On the other hand, the present application also provides a computer program product, which includes a computer program, which can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute a radar target detection method based on compressed sensing provided by the above methods, the method including: obtaining a radar echo signal returned by a radar transmission signal after reaching a target to be measured; determining the standard echo information of the target to be measured at multiple preset distance units based on the characteristics of the radar transmission signal; solving a convex optimization problem based on compressed sensing theory based on the radar echo signal and the standard echo information to obtain a convex optimization intermediate result, the convex optimization intermediate result including the position information and initial signal amplitude information of the target to be measured; obtaining a compensation coefficient corresponding to the convex optimization intermediate result based on the convex optimization intermediate result and the standard echo information; correcting the initial signal amplitude information based on the compensation coefficient to obtain the corrected signal amplitude information; and obtaining a radar detection result of the target to be measured based on the position information of the target to be measured and the corrected signal amplitude information.

[0135] On the other hand, the present application also provides a non-transitory computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, is implemented to execute a radar target detection method based on compressed sensing provided by the above-mentioned methods, the method comprising: obtaining a radar echo signal returned by a radar transmission signal after reaching a target to be measured; determining, based on the characteristics of the radar transmission signal, standard echo information of the target to be measured at multiple preset distance units; solving a convex optimization problem based on compressed sensing theory based on the radar echo signal and the standard echo information to obtain a convex optimization intermediate result, the convex optimization intermediate result including the position information and initial signal amplitude information of the target to be measured; obtaining a compensation coefficient corresponding to the convex optimization intermediate result based on the convex optimization intermediate result and the standard echo information; correcting the initial signal amplitude information based on the compensation coefficient to obtain corrected signal amplitude information; and obtaining a radar detection result of the target to be measured based on the position information of the target to be measured and the corrected signal amplitude information.

[0136] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate, and the components shown as units may or may not be physical units, i.e., they may be located in one location or distributed across multiple network units. Some or all of the modules may be selected based on actual needs to achieve the objectives of the present embodiment. Persons of ordinary skill in the art will be able to understand and implement the present invention without inventive effort.

[0137] Through the above description of the embodiments, those skilled in the art will clearly understand that each embodiment can be implemented using software plus a necessary general-purpose hardware platform, or of course, hardware. Based on this understanding, the essence of the above technical solution, or the portion 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, a magnetic disk, or an optical disk, and includes a number of instructions for causing a computer device (such as a personal computer, server, or network device) to execute the methods described in each embodiment or certain portions of the embodiments.

[0138] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, rather than to limit them. Although the present application has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present application.

Claims

1. A radar target detection method based on compressed sensing, characterized in that: include: Obtain the radar echo signal returned by the radar transmission signal after reaching the target to be measured; Determining standard echo information of the target to be measured at a plurality of preset distance units according to the characteristics of the radar transmission signal; Solving a convex optimization problem based on compressed sensing theory according to the radar echo signal and the standard echo information to obtain a convex optimization intermediate result, wherein the convex optimization intermediate result includes position information of the target to be measured and initial signal amplitude information; Obtaining a compensation coefficient corresponding to the convex optimization intermediate result according to the convex optimization intermediate result and the standard echo information; Correcting the initial signal amplitude information according to the compensation coefficient to obtain corrected signal amplitude information; A radar detection result of the target to be detected is obtained according to the position information of the target to be detected and the corrected signal amplitude information.

2. The radar target detection method based on compressed sensing according to claim 1, characterized in that: Obtaining a radar detection result of the target to be detected based on the position information of the target to be detected and the corrected signal amplitude information includes: Determining the noise intensity of the target to be measured in the radar space according to the compensation coefficient, the standard echo information, the convex optimization intermediate result, and the radar echo signal; Determining a detection threshold based on the noise intensity and a preset false alarm rate; The position information of the target to be detected and the corrected signal amplitude information are screened according to the detection threshold to obtain a radar detection result of the target to be detected.

3. The radar target detection method based on compressed sensing according to claim 2, characterized in that: The step of screening the position information of the target to be detected and the corrected signal amplitude information according to the detection threshold to obtain a radar detection result of the target to be detected includes: Determining a test statistic corresponding to the target to be measured at each of the distance units according to the position information of the target to be measured and the corrected signal amplitude information; The detection statistics are screened according to the detection threshold to obtain a target detection statistic, wherein the probability that the target to be detected exists in the distance unit corresponding to the target detection statistic is greater than a preset probability; According to the target detection statistic, a radar detection result of the target to be detected is determined from the position information of the target to be detected and the corrected signal amplitude information.

4. The radar target detection method based on compressed sensing according to claim 3, characterized in that: The radar transmission signal is expressed by the following formula (1), the standard echo information is recorded by an observation matrix, and each element in the observation matrix is ​​determined by the following formula (2): （1） （2） in, is the carrier frequency, is the linear modulation frequency, , is the pulse width, is the bandwidth, is the row number of the element in the observation matrix, is the column number of the element in the observation matrix, is the pulse repetition interval, is the sampling rate, is the time interval between two samples.

5. The radar target detection method based on compressed sensing according to claim 4, characterized in that: The radar echo signal includes a plurality of to-be-processed segments, which are obtained by dividing the radar echo signal into different processing units in advance using a preset coherent processing time as a unit, and then dividing each processing unit separately using a pulse repetition interval of the radar transmission signal as a unit. One to-be-processed segment corresponds to radar echo information of one pulse, and the convex optimization intermediate result is calculated by the following formula (3) and formula (4): （3） （4） in, is the intermediate result of convex optimization, is the radar echo information corresponding to a single segment to be processed, is the observation matrix, for The corresponding solution vector, is the total number of distance units, is the regularization parameter of the sparse penalty term in the convex optimization problem, |·| is the modulus of the complex number, Indicates taking the optimal solution vector.

6. The radar target detection method based on compressed sensing according to claim 5, characterized in that: The obtaining, according to the convex optimization intermediate result and the standard echo information, a compensation coefficient corresponding to the convex optimization intermediate result includes: According to the convex optimization intermediate result and the standard echo information, the compensation coefficient corresponding to the convex optimization intermediate result is obtained by the following formula (5) and formula (6); （5） （6） in, is the compensation coefficient, is the number of columns of the normalized observation matrix, is the number of rows of the normalized observation matrix, is the intermediate result of convex optimization, is the regularization parameter of the sparse penalty term in the convex optimization problem, is the normalized observation matrix, Represents solving the normalized observation matrix The eigenvalues ​​of the Gram matrix of It is the first intermediate variable in the calculation process.

7. The radar target detection method based on compressed sensing according to claim 6, characterized in that: The step of correcting the initial signal amplitude information according to the compensation coefficient to obtain corrected signal amplitude information includes: According to the compensation coefficient, the initial signal amplitude information is corrected by the following formula (7) to obtain the corrected signal amplitude information; （7） in, is the intermediate result of the modified convex optimization, contains the corrected signal amplitude information, is the compensation coefficient, is the normalized observation matrix, is the intermediate result of convex optimization, is the normalized radar echo information corresponding to a single segment to be processed.

8. The radar target detection method based on compressed sensing according to claim 7, characterized in that: The determining, according to the compensation coefficient, the standard echo information, the convex optimization intermediate result, and the radar echo signal, of the noise intensity of the target to be measured in the radar space includes: According to the compensation coefficient, the standard echo information, the convex optimization intermediate result and the radar echo signal, the noise intensity of the target to be measured in the radar space is determined by the following formula (8) and formula (9); （8） （9） in, is the noise intensity of the target to be measured in the radar space, is the compensation coefficient, is the normalized observation matrix, is the intermediate result of convex optimization, is the normalized radar echo information corresponding to a single segment to be processed, is the number of rows of the normalized observation matrix, is the number of columns of the normalized observation matrix, is the standard deviation of the noise of the radar echo signal, is the first intermediate variable in the calculation process, is the second intermediate variable in the calculation process, It represents the eigenvalues ​​of the Gram matrix of the normalized observation matrix A.

9. The radar target detection method based on compressed sensing according to claim 8, characterized in that: The determining of the detection threshold according to the noise intensity and the preset false alarm rate includes: According to the noise intensity and the preset false alarm rate, the detection threshold is determined by the following formula (10) and formula (11); （10） （11） in, is the detection threshold, is the preset false alarm rate, is the inverse cumulative distribution function of the standard normal distribution, is the noise intensity of the target to be measured in the radar space.

10. A radar target detection device based on compressed sensing, characterized in that: include: The first acquisition module is used to acquire the radar echo signal returned by the radar transmission signal after reaching the target to be measured; A determination module, configured to determine standard echo information of the target to be measured at a plurality of preset distance units according to the characteristics of the radar transmission signal; A second acquisition module is configured to solve a convex optimization problem based on compressed sensing theory according to the radar echo signal and the standard echo information to obtain a convex optimization intermediate result, wherein the convex optimization intermediate result includes position information of the target to be measured and initial signal amplitude information; A third acquisition module is used to obtain a compensation coefficient corresponding to the convex optimization intermediate result according to the convex optimization intermediate result and the standard echo information; a correction module, configured to correct the initial signal amplitude information according to the compensation coefficient to obtain corrected signal amplitude information; The fourth acquisition module is used to obtain the radar detection result of the target to be detected based on the position information of the target to be detected and the corrected signal amplitude information.

Images