Two-stage range and velocity estimation method and device for frequency agile radar

Abstract

This invention discloses a two-stage range and velocity estimation method for frequency-agile radar, comprising: performing a first-stage linear processing on the echo signal, calculating and summing the squared amplitude magnitudes of each pulse, using a constant false alarm rate (CFAR) detector to determine whether a target is valid at the peak of the summation result, successively detecting new targets and iteratively correcting the parameters of detected targets until a detection termination condition is met, obtaining the final set of ranges and amplitudes of all targets; performing a second-stage linear processing on the echo, estimating the precise range and velocity of each target based on the range units of each target obtained in the first stage using matched filtering and Newton's method. This invention also provides a two-stage range and velocity estimation device. The method provided by this invention addresses the problems of poor accuracy, high complexity, and detection difficulties in detecting target range and velocity using frequency-agile radar.

Description

Technical Field

[0001] This invention belongs to the field of radar signal processing technology, and in particular relates to a two-stage range and velocity estimation method and apparatus for frequency agile radar. Background Technology

[0002] With the rapid development of modern electronic technology, radar, as an important piece of electronic equipment, faces an increasingly complex electromagnetic environment and continuous advancements in electronic countermeasures technology. Frequency-agile radar, as a typical pulse-based radar, is characterized by its ability to rapidly switch between pulse carrier frequencies within its radar bandwidth. Due to the randomly changing carrier frequency of frequency-agile radar, it exhibits superior performance in low intercept, electronic countermeasures, and electromagnetic compatibility. Therefore, in-depth research on signal processing algorithms for frequency-agile radar has significant theoretical and practical value.

[0003] For long-range radar detection scenarios, frequency-agile radar systems are widely used for the joint estimation of target range and velocity parameters. Existing research on multi-target detection generally employs detection strategies based on multiple range cells, but this method suffers from high computational complexity and limited detection performance. From a signal processing perspective, traditional matched filtering (MF) techniques mainly involve two key steps: first, pulse compression (PC) processing is performed in the fast time domain, followed by a one-dimensional discrete Fourier transform (DFT) in the slow time domain. It is worth noting that this method suffers from a significant "discrete grid" effect; that is, when the target position deviates from the preset range grid or DFT grid, it leads to significant spectral leakage of target energy in both the range and frequency domains.

[0004] Linear frequency-agile radar is widely used in the military field due to its high range and velocity resolution and good anti-jamming performance. In complex jamming environments, fixed-frequency radar is susceptible to enemy interference, affecting target estimation and detection. Frequency-agile radar operates over a wide bandwidth using random frequency hopping, which is expected to evade malicious enemy interference and achieve target detection and estimation. Traditional methods for handling multi-target detection problems with frequency-agile radar often involve detection at multiple range cells, which has high computational complexity and makes detection difficult.

[0005] Patent CN118859129A discloses a segmented pulse compression cancellation method for inter-pulse and intra-pulse frequency agile radar to combat intermittent sampling interference. This method effectively addresses the challenges posed by intermittent sampling and forwarding interference to radar target detection, tracking, and identification. The invention employs inter-pulse frequency agile-intra-pulse frequency encoded signals and achieves interference suppression and target parameter reconstruction through steps such as segmented pulse compression and interference decision-making. Compared to traditional pulse compression radar, this method possesses stronger electronic countermeasures capabilities and can effectively reduce the interception capability of jammers on radar signals. However, this method still has some limitations in practical applications, such as high requirements for signal processing and computing resources, and the need to improve its adaptability to complex electromagnetic environments.

[0006] Patent CN116430386A discloses a high-resolution forward-looking imaging method and apparatus for broadband frequency-agile radar. It aims to solve the grid mismatch problem inherent in traditional sparse recovery algorithms when processing sparse scenes in discrete space, thus hindering the achievement of high-resolution forward-looking imaging and limiting the feasibility of broadband frequency-agile radar forward-looking imaging systems. This method directly models the acquired broadband signal, performs coarse parameter estimation and precise local parameter estimation, ultimately achieving parameter estimation in continuous space, avoiding the grid mismatch problem of traditional methods. Simultaneously, it utilizes the high-resolution range characteristic of broadband signals to improve the accuracy of azimuth measurement, thereby achieving high-resolution forward-looking imaging of both range and azimuth. However, this invention requires multiple iterations and parameter estimations, resulting in a large computational load, especially when the number of targets is large or the scene is complex. Furthermore, broadband signals are more sensitive to noise, requiring a higher signal-to-noise ratio to ensure imaging quality. Summary of the Invention

[0007] The purpose of this invention is to provide a two-stage range and velocity estimation method and apparatus for frequency agile radar. This method is used to solve the problems of poor accuracy, high complexity and difficulty in detecting the range and velocity of targets by frequency agile radar.

[0008] To achieve the first objective of this invention, the following technical solution is provided: a two-stage range-velocity estimation method for frequency-agile radar, comprising the following steps:

[0009] Step 1: Acquire the echo signal from the frequency agile radar and perform the first stage of linear processing on the echo signal to obtain the first stage target data for each target;

[0010] Step 2: Calculate the square of the amplitude magnitude corresponding to multiple pulses in the first stage target data, and use a constant false alarm rate detector to determine the effective targets based on the square of the amplitude magnitude to obtain the first result and construct the effective target set;

[0011] Step 3: Remove the targets stored in the valid target set from the echo signal, and perform the first stage linear processing on the removed echo signal to obtain the first stage target data of the remaining targets. Then execute the process of step 2 to obtain the second result and store it in the valid target set.

[0012] Meanwhile, the second result is fixed, the first result is corrected using Newton's method, and the corrected first result is put back into the set of valid targets;

[0013] Step 4: Repeat steps 2-3 until the termination condition is met to obtain the final valid target set containing distance and amplitude information of all targets;

[0014] Step 5: Perform a second-stage linear processing on the echo signal to obtain the second-stage target data for each target;

[0015] Step 6: Based on the final effective target set obtained in Step 4, select the second-stage target data of the targets and calculate the fine-resolution distance and velocity estimates for each target.

[0016] This invention employs a two-stage method to detect the range and velocity of multiple targets. In the first stage, after oversampling matched filtering of the frequency-agile echo signal, Newton's method is introduced to perform Newton correction on the range and amplitude values ​​in the frequency-agile radar echo signal. By fixing a portion of the detected results, the remaining detection results are iteratively corrected, solving the problem of enemy interference affecting target estimation and detection. In the second stage, pulse compression is used to process the frequency-agile echo signal. Based on the range estimation value from the first stage, the target's range cell is obtained, thus avoiding the problem of traditional methods requiring detection of each range cell and reducing computational complexity. Then, the Newtonized orthogonal matching pursuit (NOMP) algorithm is introduced to perform alternating minimization of the amplitude value, range-related value, and velocity-related value in the pulse compression signal at the fixed range cell using Newton correction, and by fixing a portion of the detected results, the remaining detection results are iteratively corrected.

[0017] Specifically, in step 2, the squares of the amplitude magnitudes corresponding to multiple pulses are summed, and the highest value of the summation result is used as the first detection unit to be input into the pre-constructed constant false alarm detector for judgment. If it is determined to be a valid target, the distance and amplitude values ​​corresponding to the first detection unit are corrected using Newton's method to obtain the corresponding first result, and the first result is stored in the set of valid targets.

[0018] Specifically, in step 3, the square of the amplitude magnitude in the first-stage target data of the remaining targets is calculated and summed based on the impulse dimension;

[0019] The highest value of the summation result is used as the second detection unit and input into the constant false alarm rate detector for judgment:

[0020] If a target is determined to be a valid target, the distance and amplitude values ​​corresponding to the second detection unit are corrected using Newton's method to obtain the corresponding second result, and the second result is stored in the set of valid targets.

[0021] Specifically, in step 6, based on the distance information of the targets in the final effective target set obtained in step 4, the distance unit corresponding to each target is obtained, and the distance unit corresponding to the target in the target data obtained in step 5 is selected. The fine-resolution distance and velocity estimate of each target is obtained through matched filtering and Newton's method.

[0022] Specifically, in the first stage of linear processing, an oversampling matched filtering method is used to pulse compress the echo signal to obtain pulse-compressed data.

[0023] Specifically, the detection process of the constant false alarm rate detector is as follows:

[0024] Remove guard cells near the target detection unit from all training units;

[0025] After removing the data, calculate the average of all data except for the maximum squared value of the amplitude magnitude to obtain the noise floor level;

[0026] The threshold for an effective target is determined based on the noise floor magnitude. The ratio of the maximum squared amplitude of the noise floor magnitude to the threshold is then compared with the threshold. If the ratio is greater than the threshold, an effective target is considered to exist; otherwise, an ineffective target is considered to exist.

[0027] Specifically, the modified expression for Newton's method is as follows:

[0028] ;

[0029] in, This indicates the corrected distance unit. This indicates the distance before correction by the first detection unit. Represents the cost function At the iteration point Regarding distance The first derivative, Represents the cost function At the iteration point Regarding distance The second derivative, ;

[0030] Based on the corrected distance units, the complex amplitude of the target is obtained using the least squares method, expressed by the formula:

[0031] ;

[0032] in, This indicates the corrected complex amplitude.

[0033] Specifically, the second-stage linear processing is based on pulse compression using fast Fourier transform. The input signal is subjected to fast Fourier transform, then multiplied with the frequency domain representation of the matched filter, and finally the result is converted back to the time domain through inverse fast Fourier transform. The slow time dimension data of the target distance cell in the second stage is extracted based on the target distance estimate in the first stage.

[0034] Specifically, in step 6, the expressions for the fine-resolution range and velocity estimates of the target are as follows:

[0035] ;

[0036] ;

[0037] in, Indicates the distance to the target. Represents the speed of light. Indicates the carrier frequency. Indicates the pulse repetition period. Indicates the speed of the target. It is an integer, which makes the digital frequency , , , Variables related to distance estimates This represents variables related to the speed estimate.

[0038] To achieve the second objective of this invention, the following technical solution is provided: a two-stage range-velocity estimation device, used to execute the steps of the above-described two-stage range-velocity estimation method for frequency-agile radar, comprising:

[0039] The first-stage target data acquisition unit is used to acquire the echo signal from the frequency agile radar, perform first-stage linear processing, and obtain the first-stage target data.

[0040] The first-stage Newton correction unit is used to calculate the square of the magnitude of the matched filter result and estimate the target distance by summing the impulses. The point with the highest value of the square of the magnitude is used as the first detection unit and input into the pre-constructed constant false alarm detector for detection. The distance and magnitude value corresponding to the first detection unit that is detected as a valid target are corrected by Newton's method, and the corrected first result is added to the set of valid targets.

[0041] The first-stage cyclic correction unit is used to remove the first result from the echo signal and perform linear processing on the echo signal after removing the first result to obtain the target data after removal. The point with the highest squared amplitude magnitude in the target data after removal is selected as the second detection unit and input into the pre-trained constant false alarm rate detector for detection. The distance and amplitude values ​​corresponding to the second detection unit that are detected as valid targets are corrected using Newton's method, and the corrected second result is added to the set of valid targets. At the same time, the second result is fixed, and the first result is cyclically corrected using Newton's method. After correction, it is stored in the set of valid targets again.

[0042] The first-stage target output unit is used to repeat the first-stage Newton correction unit and the first-stage cyclic correction unit until the judgment termination condition of the constant false alarm detector is met, and output the final set of valid targets to obtain the distance and amplitude information of all targets in the echo signal, and obtain the distance information therein.

[0043] The second-stage target data acquisition unit is used to acquire the echo signal from the frequency-agile radar and perform second-stage linear processing to obtain the target data in the second stage.

[0044] The second-stage Newton correction unit, based on the final effective target set obtained from the first-stage target output unit, uses the distances of the targets obtained therein to calculate the distance unit corresponding to each target. It then calculates the squared magnitude of the two-dimensional data of distance-related values ​​and velocity-related values ​​from the fixed distance units in the second-stage target data. Taking the highest value of the squared magnitude as the initial point, it uses the Newton orthogonal matching pursuit algorithm to correct the magnitude value, distance-related value, and velocity-related value corresponding to the first detection unit that is a valid target, and adds the corrected first result to the effective target set.

[0045] The second-stage cyclic correction unit is used to remove the first result from the echo signal in the second stage, and to perform second-stage linear processing on the echo signal after removing the first result to obtain the target data after removal. The point with the highest squared amplitude magnitude in the target data after removal is selected as the second detection unit, which is input into the pre-trained constant false alarm rate detector for detection. Newton's method is used to correct the amplitude value, distance-related value, and velocity-related value corresponding to the second detection unit that is detected as a valid target, and the corrected second result is added to the set of valid targets. At the same time, the second result is fixed, and Newton's method is used to perform cyclic correction on the first result. After correction, it is stored in the set of valid targets again.

[0046] The second stage outputs all amplitude values, distance-related values, and speed-related values ​​from the echo signal, and then converts these distance-related and speed-related values ​​into distance and speed information to obtain the speed information.

[0047] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0048] (1) Compared with the traditional pulse compression method based on matched filtering, the method is modified by Newton's method based on grid points, thus overcoming the grid mismatch problem and improving the accuracy of distance and velocity estimation;

[0049] (2) In complex interference environments, fixed-frequency radar is susceptible to local interference, which affects target estimation and detection. Frequency-agile radar operates in a wide bandwidth using random frequency hopping, which can effectively avoid malicious interference from the enemy and achieve target detection and estimation;

[0050] (3) After multi-pulse target distance detection, the target is fixed on a distance unit for target distance and velocity detection. By reducing the number of distance units that need to be processed, the computational complexity is significantly reduced and the detection efficiency is improved, thus effectively solving the problems of high complexity and detection difficulty in traditional methods. Attached Figure Description

[0051] Figure 1 This is a flowchart of the two-stage range and velocity estimation method for frequency-agile radar provided in this embodiment;

[0052] Figure 2 This is a schematic diagram of the two-stage distance-velocity estimation device provided in this embodiment;

[0053] Figure 3 This is a schematic diagram of the distance and velocity detection results for multiple targets provided in this embodiment;

[0054] Figure 4 This is a schematic diagram illustrating the detection results of multiple targets with similar velocities provided in this embodiment;

[0055] Figure 5 This is a schematic diagram of the detection and target recognition results of multi-target pulse interference provided in this embodiment. Detailed Implementation

[0056] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. The components of the embodiments of the present invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations. Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected embodiments of the invention. All other embodiments obtained by those skilled in the art based on the embodiments of the present invention without inventive effort are within the scope of protection of the present invention.

[0057] like Figure 1 As shown, the two-stage range-velocity estimation method for frequency-agile radar provided in this embodiment includes the following steps:

[0058] S110: Acquire the echo signal from the frequency agile radar and perform the first stage of linear processing on the echo signal to obtain the first stage target data.

[0059] The actual operation process in this embodiment is as follows:

[0060] First, initialize the constant false alarm rate (CFAR) detector by inputting the false alarm probability, maximum number of iterations, oversampling factor, correction count, number of guard units, and number of training units. Then, receive the echo signal from the frequency-agile radar. This can be expressed as a formula:

[0061] ;

[0062] in Indicates pulse index, Indicates the target number. Indicates the first Distance to each target It is noise.

[0063] ;

[0064] in, Indicates the carrier frequency, satisfying And each frequency point From the set Randomly selected from, Indicates rounding down. Indicates the frequency modulation interval, where It is the first Random frequency-modulated codewords of pulses, Indicates the pulse repetition period. Indicates the first The speed of each target.

[0065] Indicates the distance of the target. The atomic vector at that location is expressed by the formula:

[0066] ;

[0067] in, Indicates the sampling time. Indicates distance, Represents the speed of light. Indicates a pulse signal. This represents the number of samples in the distance dimension.

[0068] In the first stage of target distance estimation, an oversampling matched filter method is used. Linear pulse compression, followed by Doppler oversampling FFT, is expressed by the following formula:

[0069] ;

[0070] in, Indicates distance, Let n represent a finite discrete set, and n represent a fast-time index. Represents the speed of light. Indicates the fast sampling frequency. Represents the number of samples in the distance dimension. Sampling rate The number of samples corresponding to the lower pulse width. This represents the oversampling factor relative to the traditional pulse compression grid.

[0071] The target data for the first phase is expressed using the following formula:

[0072] ;

[0073] ;

[0074] in, Indicates the distance to the target. This indicates taking the maximum value; Indicates the target number The amplitude value of each pulse. Indicates the first The echo signal of each pulse, This indicates the conjugate transpose.

[0075] S120. For the target data obtained in S110, calculate the square of the amplitude magnitude of the target data, and sum the square of the amplitude magnitude corresponding to multiple pulses to estimate the distance of the target. The highest value of the summation result is used as the first detection unit and input into the pre-constructed constant false alarm detector for detection. If the target is determined to exist, Newton's method is used to correct the distance and amplitude value corresponding to the first detection unit that is detected as a valid target, and the corrected first result is added to the set of valid targets.

[0076] The amplitude corresponding to the first detection unit ,distance Maximum likelihood estimation This can be obtained by minimizing the energy of the residual:

[0077] ;

[0078] It is equivalent to maximizing the following cost function :

[0079] ;

[0080] in, Indicates the optimal , After finding the minimum value of the entire expression, return the corresponding variable value. Indicates taking the real part, , Describing the L2 norm, , That is:

[0081] ;

[0082] For a given Use the least squares method to find the th amplitude of each pulse The estimated value is:

[0083] ;

[0084] in .

[0085] Will exist By replacing it with the formula above, we can obtain... The generalized likelihood ratio detection (GLRT) estimation is essentially solving the following optimization problem:

[0086] ;

[0087] in, This is the cost function for GLRT.

[0088] In this embodiment, the determination of whether a target exists at the detection unit is expressed by the following formula:

[0089] ;

[0090] in, This represents the threshold used to determine a valid target. If the condition is met, it is determined that there is a valid target at the detection unit, and the next step of correction can be performed; otherwise, the algorithm stops.

[0091] In this embodiment, if the detection result of the first detection unit is a valid target, then Newton's method is used to correct the distance and amplitude values ​​of the unit. The correction process is as follows:

[0092] The distance of the first detection unit is corrected using the cost function and Newton's method. :

[0093]

[0094] Then the complex amplitude can be updated:

[0095]

[0096] The cost function is solved with respect to distance. The first and second derivatives are crucial, and the process is as follows:

[0097] ;

[0098] in, Indicates distance The first derivative, Indicates distance The second derivative, further, gives:

[0099] ;

[0100] In the formula The element The expression is:

[0101] ;

[0102] Will Abbreviated as Regarding The first derivative is:

[0103]

[0104] Will Abbreviated as Regarding The second derivative is:

[0105] ;

[0106] Finally, the distance and amplitude values ​​of the corrected first detection unit are added to the set of valid targets.

[0107] S130. After completing the first stage of data processing, the first result is removed from the echo signal, and the removed echo signal is subjected to the first stage of linear processing to obtain the removed target data; the square of the amplitude modulus of the removed target data is calculated and summed based on the pulse dimension; the highest value of the summation result is used as the second detection unit and input to the constant false alarm detector for judgment: if it is determined to be a valid target, the distance and amplitude values ​​corresponding to the second detection unit are corrected using Newton's method, and the corrected second result is stored in the set of valid targets; the second result is fixed, the first result is cyclically corrected using Newton's method, and the corrected result is stored in the set of valid targets again.

[0108] In this embodiment, it is assumed that the currently estimated set of effective targets is:

[0109] ;

[0110] The residual of the received signal is denoted as Then each time the first [number] is selected... Each objective is revised, and its corresponding residuals are... for:

[0111] ;

[0112] Then based on the residuals For the first The objectives will be revised and revised. Revised to So, residual It will also be updated to :

[0113] ;

[0114] use replace Then proceed to the next step. The process of revising each objective continues until all revisions are completed.

[0115] S140, repeat S120 and S130 until the judgment termination condition of the constant false alarm detector is met, so as to construct the final effective target set, which contains the distance information and amplitude information of all targets.

[0116] Assume the corrected set of valid targets is:

[0117] ;

[0118] The residual of the signal is then:

[0119] ;

[0120] residual signal The data is fed into the constant false alarm rate (CFAR) detection module to determine if any additional targets exist. If none exist, the algorithm stops. Otherwise, Newton's correction and iterative correction steps are executed.

[0121] The process continues until all target information in the echo signal has been detected or the maximum number of iterations of the constant false alarm rate detector has been reached. Then, the distance and amplitude information of all targets in the echo signal are output, and the distance information is selected.

[0122] S150: Obtain the echo signal from the frequency agile radar and perform linear processing in the second stage to obtain the target data for the second stage.

[0123] In the second stage of estimating the target velocity, the... The pulse number In a scenario with multiple targets, the echo signal, after pulse compression, can be represented as:

[0124] ;

[0125] in, Indicates the first The scattering intensity of each target, This indicates that the echo signal from the frequency agile radar has undergone pulse compression processing. It is noise.

[0126] It can be written as:

[0127] ;

[0128] The Newton-Orthogonal Matching Pursuit algorithm for each target can be written as:

[0129] ;

[0130] ;

[0131] ;

[0132] ;

[0133] in, Indicates the distance to the target. Represents the speed of light. Indicates the carrier frequency. Indicates the pulse repetition period. Indicates the speed of the target. It is an integer, which makes the digital frequency , , .

[0134] for Pulse compression is performed, and slow-time dimension data of the target location cell in the second stage is extracted based on the target distance estimate in the first stage.

[0135] Performing maximum likelihood estimation in the single-objective case yields the least squares problem:

[0136] ;

[0137] For a given and Optimal solution for:

[0138] ;

[0139] By eliminating It can be simplified to:

[0140] ;

[0141] in,

[0142] ;

[0143] ;

[0144] in, , To finely distinguish the number of samples in the distance and velocity dimensions, and It is the oversampling factor relative to the Nyquist grid. This represents the estimation results related to the distance-velocity dimension. This indicates taking the maximum value.

[0145] S160. In the second stage, Newton's method is used to obtain the fine-resolution range and velocity estimates for each target.

[0146] Using the frequency-corrected Newton coordinate descent method, under the condition of local convexity of the function, this update rule is employed to maximize... ,in , and with This indicates the number of Newton corrections.

[0147] ;

[0148] in, and This represents the corrected distance-related values ​​and speed-related values. Because... The gradient and the elements of the Hessian matrix are calculated as follows:

[0149] ;

[0150] ;

[0151] ;

[0152] ;

[0153] ;

[0154] Among them are:

[0155] ; ;

[0156] ;

[0157] ;

[0158] ;

[0159] Finally, the first detection unit that has been corrected will be... , and The value is added to the set of valid targets, and the final distance and velocity information of the target are calculated by combining the distance cell number.

[0160] This embodiment also provides a two-stage range and velocity estimation device, including a memory and a processor. The memory is used to store a computer program, and the processor is used to implement the two-stage range and velocity estimation method for frequency-agile radar described in the above embodiment when the computer program is executed.

[0161] like Figure 2 As shown, it includes:

[0162] 510. First-stage target data acquisition unit, used to acquire the echo signal of the frequency agile radar, perform first-stage linear processing, and obtain first-stage target data;

[0163] 520. The first stage Newton correction unit is used to calculate the square of the amplitude magnitude of the matched filter result and estimate the target distance by summing the impulses. The point with the highest value of the square of the amplitude magnitude is used as the first detection unit and input into the pre-constructed constant false alarm detector for detection. The distance and amplitude value corresponding to the first detection unit that is detected as a valid target are corrected by Newton's method, and the corrected first result is added to the set of valid targets.

[0164] 530. The first-stage cyclic correction unit is used to remove the first result from the echo signal and perform first-stage linear processing on the removed echo signal to obtain the first-stage target data of the remaining targets. It calculates the square of the amplitude magnitude in the first-stage target data of the remaining targets and sums it based on the pulse dimension. The highest value of the summation result is used as the second detection unit to be input into the constant false alarm detector for judgment. If it is determined to be a valid target, the distance and amplitude value corresponding to the second detection unit are corrected using Newton's method to obtain the corresponding second result. The second result is stored in the set of valid targets. The second result is fixed, and Newton's method is used to perform cyclic correction on the first result. The corrected result is stored in the set of valid targets again.

[0165] 540. The first-stage target output unit is used to repeat the first-stage Newton correction unit and the first-stage cyclic correction unit until the judgment termination condition of the constant false alarm detector is met, so as to construct the final effective target set, which contains the distance information and amplitude information of all targets.

[0166] 550. The second-stage target data acquisition unit performs second-stage linear processing on the echo signals obtained by the first-stage target data acquisition unit to obtain the second-stage target data for each target.

[0167] 560. The second-stage Newton correction unit, based on the final effective target set obtained from the first-stage target output unit, uses the distances of the targets obtained therein to calculate the distance unit corresponding to each target. It calculates the squared amplitude of the two-dimensional data of the distance-related values ​​and velocity-related values ​​of the fixed distance units in the second-stage target data, and takes the highest value of the squared amplitude magnitude as the initial point. It uses the Newton orthogonal matching pursuit algorithm to correct the amplitude value, distance-related value, and velocity-related value corresponding to the first detection unit that is an effective target, and adds the corrected first result to the effective target set.

[0168] 570. The second-stage cyclic correction unit is used to remove the first result from the echo signal in the second stage, and to perform second-stage linear processing on the echo signal after removing the first result to obtain the target data after removal. The point with the highest squared amplitude magnitude in the target data after removal is selected as the second detection unit, which is input into the pre-trained constant false alarm rate detector for detection. The amplitude value, distance-related value, and velocity-related value corresponding to the second detection unit that is detected as a valid target are corrected using Newton's method, and the corrected second result is added to the set of valid targets. At the same time, the second result is fixed, and the first result is cyclically corrected using Newton's method. After correction, it is stored in the set of valid targets again.

[0169] 580. The second stage outputs information on all amplitude values, distance-related values, and speed-related values ​​in the echo signal, and then converts the distance-related values ​​and speed-related values ​​into distance and speed to obtain the speed information.

[0170] To better illustrate the effectiveness of the solution provided in this embodiment, the following example is provided:

[0171] The radar parameters were set as follows in the experiment: carrier frequency =3GHz, pulse width =20us, pulse repetition period =200us, bandwidth =4MHz, sampling frequency =5MHz, pulse count =32, frequency modulation codeword exist Random value selection within the range, frequency hopping interval =10MHz.

[0172] Experimental Example 1: Set the actual target distances to 1000m and 1025m, and the speeds to 25m / s and 30m / s, respectively.

[0173] like Figure 3 The above describes the target detection results in a multi-target scenario, where... Figure 3 (1) is a diagram showing the signal reconstruction amplitude and target distance estimation under different pulses in the first stage of frequency agile switching provided in this embodiment. Figure 3 (2) is the detection result of the two-stage range and velocity estimation method for frequency-agile radar provided in this embodiment: the range error is less than 2m and the velocity error is less than 1m / s.

[0174] Experiment Example 2: Set the real target distances to 1000m and 1250m, and their velocities to 28m / s and 30m / s respectively. In a multi-target scenario, the minimum difference in radial velocity between the real targets is 2m / s. Figure 4The image shows a schematic diagram of the detection results for multiple targets with similar velocities. Figure 4 (1) is a diagram showing the signal reconstruction amplitude and target distance estimation under different pulses in the first stage of frequency agile switching provided in this embodiment. Figure 4 (2) is the detection result of the two-stage range and velocity estimation method for frequency-agile radar provided in this embodiment: the range error is less than 2m and the velocity error is less than 1m / s.

[0175] Experiment Example 3: In this experiment, the algorithm's ability to detect and suppress strong interference was evaluated. The real target distances were set to 1000m and 1250m, and the speeds to be 25m / s and 30m / s, respectively.

[0176] In frequency modulation code Strong interference was artificially injected into the corresponding 3rd, 4th, 7th, 11th, 19th, 20th, 27th, 29th, and 32nd pulses. Specifically, this was done by increasing the original noise variance. On top of that, an additional variance of 1 / 2 was added. The complex Gaussian noise, while other impulses contain only complex Gaussian noise with variance of 1. Thermal noise. In the first stage of range-amplitude detection, pulse compression was performed on all 32 pulses and peak amplitude was extracted. An abnormal increase in peak amplitude was observed at the nine interfered pulses, indicating the presence of external interference signals. By calculating the effective signal-to-noise ratio (SNR), we evaluated the SNR of each pulse. The results showed that the effective SNR of pulses 3, 4, 7, 11, 19, 20, 27, 29, and 32 was significantly lower than that of other pulses, thus accurately locating the interfered pulses. After removing these interfered pulses, the two-stage range-velocity estimation method for frequency-agile radar proposed in this embodiment was re-executed on the remaining interference-free pulses.

[0177] Figure 5 (1) is the effective signal-to-noise ratio diagram, which shows the effective signal-to-noise ratio of each pulse and clearly distinguishes the interfered pulses. Figure 5 Figure (2) shows a three-dimensional comparison (radial velocity-distance-target intensity) of the detection results of the real target, the target before interference removal, and the target after interference removal. It can be seen that the interference removal strategy significantly improves the accuracy of distance-velocity estimation. This verifies that the interference identification method based on the effective signal-to-noise ratio curve after pulse compression can accurately locate and remove high-power pulse interference under complex frequency agile systems, thereby improving the robustness and accuracy of target detection.

Claims

1. A two-stage range and velocity estimation method for agile frequency radar, characterized in that, Includes the following steps: Step 1: Acquire the echo signal from the frequency agile radar and perform the first stage of linear processing on the echo signal to obtain the first stage target data for each target; The first stage of linear processing uses an oversampling matched filtering method to perform pulse compression on the echo signal, resulting in pulse-compressed data. Step 2: Calculate the square of the amplitude magnitude corresponding to multiple pulses in the first stage target data, and use a constant false alarm rate detector to determine the effective targets based on the square of the amplitude magnitude to obtain the first result and construct the effective target set; Step 3: Remove the targets stored in the valid target set from the echo signal, and perform the first stage linear processing on the removed echo signal to obtain the first stage target data of the remaining targets. Then execute the process of step 2 to obtain the second result and store it in the valid target set. Meanwhile, the second result is fixed, the first result is corrected using Newton's method, and the corrected first result is put back into the set of valid targets; Step 4: Repeat steps 2-3 until the termination condition is met to obtain the final valid target set containing distance and amplitude information of all targets; Step 5: Perform a second-stage linear processing on the echo signal to obtain the second-stage target data for each target; The second-stage linear processing is based on pulse compression using fast Fourier transform. The input signal is subjected to fast Fourier transform, then multiplied with the frequency domain representation of the matched filter, and finally the result is converted back to the time domain by inverse fast Fourier transform. The slow time dimension data of the target distance unit in the second stage is extracted based on the target distance estimate in the first stage. Step 6: Based on the final effective target set obtained in Step 4, select the second-stage target data of the targets and calculate the fine-resolution distance and velocity estimates for each target.

2. The two-stage range and velocity estimation method for frequency-agile radar according to claim 1, characterized in that, In step 2, the squares of the amplitude magnitudes corresponding to multiple pulses are summed, and the highest value of the summation result is used as the first detection unit to be input into the pre-constructed constant false alarm detector for judgment. If it is determined to be a valid target, the distance and amplitude values ​​corresponding to the first detection unit are corrected using Newton's method to obtain the corresponding first result, and the first result is stored in the set of valid targets.

3. The two-stage range and velocity estimation method for frequency-agile radar according to claim 1, characterized in that, In step 3, the squared magnitude of the first-stage target data of the remaining targets is calculated and summed based on the impulse dimension; The highest value of the summation result is used as the second detection unit and input into the constant false alarm rate detector for judgment: If a target is determined to be a valid target, the distance and amplitude values ​​corresponding to the second detection unit are corrected using Newton's method to obtain the corresponding second result, and the second result is stored in the set of valid targets.

4. The two-stage range and velocity estimation method for frequency-agile radar according to claim 1, characterized in that, In step 6, based on the distance information of the targets in the final effective target set obtained in step 4, the distance unit corresponding to each target is obtained, and the distance unit corresponding to the target in the target data obtained in step 5 is selected. The fine-resolution distance and velocity estimate of each target is obtained by matched filtering and Newton's method.

5. The two-stage range and velocity estimation method for frequency-agile radar according to claim 1, characterized in that, The detection process of the constant false alarm rate detector is as follows: Remove guard cells near the target detection unit from all training units; After removing the data, calculate the average of all data except for the maximum squared value of the amplitude magnitude to obtain the noise floor level; The threshold for an effective target is determined based on the noise floor magnitude. The ratio of the maximum squared amplitude of the noise floor magnitude to the threshold is then compared with the threshold. If the ratio is greater than the threshold, an effective target is considered to exist; otherwise, an ineffective target is considered to exist.

6. The two-stage range-velocity estimation method for frequency-agile radar according to claim 1, 2, or 3, characterized in that, The corrected expression for Newton's method is as follows: ;in, Indicates the corrected distance unit. This indicates the distance before correction by the first detection unit. Represents the cost function At the iteration point Regarding distance The first derivative, Represents the cost function At the iteration point Regarding distance The second derivative, Based on the corrected distance unit, the complex amplitude of the target is obtained using the least squares method, expressed by the formula: ;in, This indicates the corrected complex amplitude.

7. The two-stage range and velocity estimation method for frequency-agile radar according to claim 1, characterized in that, In step 6, the expressions for the fine-resolution range and velocity estimates of the target are as follows: ; ;in, Indicates the distance to the target. Represents the speed of light. Indicates the carrier frequency. Indicates the pulse repetition period. Indicates the speed of the target. It is an integer, which makes the digital frequency , , , Variables related to distance estimates This represents variables related to the speed estimate.

8. A two-stage distance-velocity estimation device, characterized in that, The steps are for performing the two-stage range and velocity estimation method for frequency agile radar as described in any one of claims 1 to 7.

Images