An ISAR wideband direct sampling receiver amplitude-phase distortion correction method

Abstract

This invention discloses a method for amplitude and phase distortion correction in ISAR broadband direct sampling receivers. First, an amplitude and phase distortion echo signal model for the ISAR broadband direct sampling receiver is constructed. Matched filtering is applied to the amplitude and phase distortion echo signal at a single scattering point, and a fitness function is constructed based on the matched filtering results. Then, the PSO algorithm is used to iteratively optimize the amplitude and phase distortion coefficient vector of the echo signal at each scattering point, obtaining an estimated vector of amplitude and phase distortion coefficients, thus obtaining the amplitude and phase distortion estimates for the echo signal at each discrete point. Finally, an amplitude and phase compensation factor is calculated based on the estimated amplitude and phase distortion values ​​of the echo signal at the discrete points. This factor is used to correct the amplitude and phase distortion echo signal at each discrete point in real time, obtaining the corrected echo signal at each discrete point. The corrected echo signals at each discrete point are then superimposed to obtain the final corrected echo signal. This method solves the problems of insufficient targeting and easy getting trapped in local optima in high-dimensional solutions, thus improving ISAR imaging performance.

Description

Technical Field

[0001] This invention belongs to the field of radar signal processing technology, specifically relating to a method for amplitude and phase distortion correction of an ISAR broadband direct sampling receiver. Background Technology

[0002] Wideband radar imaging technology includes two main imaging methods: inverse synthetic aperture radar (ISAR) and synthetic aperture radar (SAR). Unlike SAR, ISAR images non-cooperative targets. These targets do not actively provide information such as their position and motion parameters, and are usually highly concealed, threatening, and destructive. Wideband direct sampling receivers rely on high-speed ADCs and digital signal processing technology, abandoning the analog deskewing and mixing preprocessing of traditional radar receivers, and directly sampling the radar intermediate frequency signal at high speed. This provides advantages such as high resolution and strong anti-interference capabilities, laying the foundation for high-resolution ISAR imaging. However, due to factors such as ADC nonlinearity and link coupling, the echo is prone to amplitude and phase distortion, leading to undesirable results such as main lobe broadening and side lobe increase in matched filtering. Therefore, amplitude and phase correction of the radar echo signal is required.

[0003] Existing radar amplitude and phase error estimation methods focus on array-type radars, with limited research on single-base radar scenarios. Single-base radar research mainly falls into two categories. The first is pre-distortion compensation (also known as online compensation), which achieves real-time correction by embedding a pre-distortion module with opposite distortion characteristics at the transmitter, requiring no external calibration source. However, it often only addresses transmit link errors, failing to cover receive channel distortion, and some schemes suffer from high computational complexity and limited adaptability. The second is ground processing (also known as offline compensation), which relies on external calibration sources or post-processing of echo data, reducing computational complexity but lacking real-time performance and susceptible to target motion and environmental interference, making it difficult to isolate inherent system errors. Both methods suffer from insufficient adaptability and limited correction accuracy, failing to meet the amplitude and phase distortion correction requirements of ISAR broadband direct sampling receivers. Summary of the Invention

[0004] To address the shortcomings of existing technologies, the technical problem this invention aims to solve is to propose a method for amplitude and phase distortion correction in ISAR broadband direct sampling receivers.

[0005] The present invention solves the aforementioned technical problem by adopting the following technical solution: A method for amplitude and phase distortion correction in an ISAR broadband direct sampling receiver, characterized by the following steps: Step 1: Construct an amplitude-phase distortion echo signal model for an ISAR broadband direct sampling receiver; (1) In the formula, This is the column vector of the amplitude and phase distortion echo signal. For the first A column vector of echo signal amplitude distortion values ​​at each scattering point For the first The ideal echo signal column vector of each scattering point For the first A column vector of echo signal phase distortion values ​​from each scattering point. For element-wise Hadamard product, For exponential operations, Indicates the imaginary part. This represents the total number of scattering points; The amplitude-phase distortion echo signal model at a single scattering point is as follows: (2) (3) (4) (5) (6) In the formula, For the first Amplitude and phase distortion echo signals at each scattering point The amplitude loss-of-bass matrix, For fast time vectors, For the first The amplitude distortion coefficient vector of the echo signal at each scattering point. For transpose operation, For the first The first scattering point echo signal Amplitude distortion coefficient, here , For the first The amplitude coefficient of the echo signal at the scattering point. For discrete rectangular window functions, For the first Ideal phase column vectors of scattering points For the first The radial distance from each scattering point to the radar. At the speed of light, The signal pulse width, For ISAR carrier frequencies, It is a diagonal matrix. The phase distortion basis matrix, For frequency modulation parameter vectors, For the first Step frequency, here , For the first The vector of phase distortion coefficients of the echo signal at each scattering point. For the first The first scattering point echo signal Phase distortion coefficient, here ; Step 2: Perform matched filtering on the amplitude and phase distortion echo signal of the single scattering point to obtain the amplitude of each time sampling point; construct the fitness function based on the matched filtering results as follows: (9) (10) (11) (12) In the formula, For the first The variable attention weight vector for each iteration To optimize the target vector, For the penalty factor vector, To penalize the regularized vector, The peak pulse pressure is the sidelobe ratio. Entropy is the energy concentration. To increase attention to amplitude distortion, For phase distortion attention, The amplitude of the first side lobe. Main lobe amplitude, For the first The amplitude of each fast time sampling point Indicates taking the absolute value; No. The variable attention weights for each optimization objective in the next iteration are: (14) In the formula, For the first During the nth iteration Variable attention weights for each optimization objective. For the first During the nth iteration The relative change of each optimization objective It is a constant with a very small value; The penalty regularization vector is represented as: (15) (16) (18) (20) In the formula, The penalty amount for the main lobe width. A 20dB width penalty is applied. This is the penalty amount for peak reversal detection. The relative change rate of the main lobe width. The relative change rate of width is 20dB. The main lobe peak value, The peak value of the side lobe; Step 3: The PSO algorithm is used to iteratively optimize the amplitude and phase distortion coefficient vector of the echo signal at each scattering point to obtain the amplitude and phase distortion coefficient estimate vector, thereby obtaining the amplitude and phase distortion estimate of the echo signal at each discrete point. Step 4: Calculate the amplitude and phase compensation factor based on the amplitude and phase distortion estimates of the discrete point echo signals, and use the amplitude and phase compensation factor to correct the amplitude and phase distortion echo signals of each discrete point in real time to obtain the corrected echo signals of each discrete point; superimpose the corrected echo signals of each discrete point to obtain the final corrected echo signal. The formula for calculating the amplitude compensation factor at each discrete point is: (32) In the formula, For the first Amplitude compensation factor for each scattering point For the first Estimated amplitude distortion of echo signal at each scattering point; The formula for calculating the phase compensation factor at each discrete point is: (34) In the formula, For the first Phase compensation factor for each scattering point For the first Estimated phase distortion of echo signal at each scattering point; The echo signals after correction at each discrete point are: (36) In the formula, For the first Echo signal after correction of each scattering point; The final corrected echo signal is: (37) In the formula, This is the final corrected echo signal.

[0006] Furthermore, the iterative process of the PSO algorithm includes: S3.1 Population initialization; The amplitude and phase distortion coefficient vectors of the scattering points are treated as particles, each containing 8 dimensions; the initial positions of each particle are represented as follows: (twenty three) In the formula, For the first The first particle The initial position of the dimension. , The first Negative and positive boundaries of the dimensional position. The number of particles; For the first The orientation identifier of the dimensional position. Indicates the first The position of the dimension takes values ​​along the positive boundary direction. Indicates the first The position is taken along the negative boundary direction; S3.2 Calculate the particle fitness, the mean and minimum population fitness, and adaptively adjust the particle inertia weight according to equation (26); (26) In the formula, For the first The second iteration The inertial weight of each particle, , These represent the maximum and minimum values ​​of the particle inertia weight. For the first The second iteration The fitness of each particle , For the first The mean and minimum fitness of the population in each iteration; S3.3, Start amplitude distortion coefficient optimization, pause phase distortion coefficient optimization, keep the global optimal position of phase dimension unchanged, and update the velocity and position of particle amplitude dimension through equations (27) and (28); (27) (28) In the formula, , For the first The second iteration The velocity and position of each particle in the amplitude dimension , For the first The second iteration The velocity and position of each particle in the amplitude dimension For the first The optimal position in the amplitude dimension of each particle. For the global optimal position in the magnitude dimension, , As a learning factor, , A random number with a value between 0 and 1; Based on the updated position of the particle amplitude dimension, calculate the particle fitness and update the individual optimal position and the global optimal position in the amplitude dimension; S3.4 Start phase distortion coefficient optimization, pause amplitude distortion coefficient optimization, keep the global optimal position of amplitude dimension unchanged, and update the velocity and position of particle phase dimension through equations (29) and (30); (29) (30) In the formula, , For the first After the nth iteration The velocity and position of each particle in the phase dimension , For the first After the nth iteration The velocity and position of each particle in the phase dimension For the first The optimal position of each particle's phase dimension. This represents the globally optimal position in the phase dimension. Calculate particle fitness based on the updated particle phase dimension position, and update the individual optimal position and global optimal position of the phase dimension based on the particle fitness. S3.5. Perform particle mutation after each iteration; S3.6 Repeat steps S3.2 to S3.5 to optimize the amplitude and phase distortion coefficients step by step. When the number of iterations reaches the maximum number of iterations, terminate the iteration and output the global optimal position as the estimated value vector of amplitude and phase distortion coefficients.

[0007] Furthermore, the positive and negative boundary vectors of the particle position , Represented as: (twenty one) (twenty two) In the formula, For the first The first scattering point echo signal Amplitude distortion coefficient error, For the first The first scattering point echo signal Phase distortion coefficient error, .

[0008] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. This invention employs a penalized regularized variable attention mechanism to construct a fitness function. Attention weights are dynamically allocated based on the real-time relative changes of each optimization objective (including peak-to-sidelobe ratio, entropy energy concentration, amplitude distortion attention, and phase distortion attention), achieving adaptive targeted focusing on core distortion features. Furthermore, multi-dimensional penalized regularization constraints, including main lobe width, 20dB width, and peak inversion detection, are proposed to avoid signal distortion during optimization, significantly improving the accuracy and stability of distortion estimation. Simultaneously, the traditional PSO algorithm is improved in multiple dimensions. An 8-dimensional optimization particle is constructed to adapt to amplitude and phase distortion features. An adaptive inertial weight adjustment strategy balances global search and local fine-grained optimization. The step-by-step iterative optimization strategy for amplitude and phase distortion decomposes the high-dimensional coupled solution into low-dimensional independent solutions. A regularized population initialization and particle mutation mechanism are introduced to effectively break the local optimum trap and overcome the limitation of the traditional PSO algorithm in easily getting trapped in local optima in high-dimensional solutions of multi-order amplitude and phase distortion. Therefore, by combining the penalty regularization variable attention mechanism with the improved particle swarm optimization algorithm, the problems of insufficient targeting and easy getting trapped in local optima in high-dimensional solutions of traditional correction methods are solved, thereby improving the target feature extraction accuracy and ISAR imaging performance of ISAR broadband direct sampling receiver.

[0009] 2. Addressing the characteristics of ISAR broadband direct sampling receiver links, a model for amplitude and phase distortion echo signals driven by multi-factor coupling distortion in the radar system link is constructed. This model fully considers the amplitude and phase distortion problems caused by ADC nonlinearity and link coupling, specifically incorporating 0-3 order amplitude distortion and 0, 2-4 order phase distortion characteristics. It also integrates core ISAR parameters such as fast time dimension and target equivalent independent scattering point characteristics to achieve a refined and comprehensive characterization of broadband direct sampling echo signal distortion. Dimensional unification and normalization are introduced during the modeling process to eliminate the influence of frequency modulation slope and pulse width on the distortion coefficients, solving the problem of insufficient fitting ability of traditional models for broadband high-speed sampling signal distortion. Furthermore, the model supports matrix superposition of multiple scattering points, better meeting the imaging requirements of ISAR for non-cooperative targets.

[0010] 3. To address the issues of local optima and imbalance between global and local searches that arise in the traditional PSO algorithm for high-dimensional amplitude and phase distortion solutions in ISAR broadband direct sampling receivers, multi-dimensional improvements are made based on the coupling characteristics of amplitude and phase distortion to adapt to the distortion correction requirements of ISAR broadband direct sampling receivers. Firstly, the population initialization method is optimized, abandoning random initialization and constructing an 8-dimensional optimized particle with 4 amplitude and 4 phase paths. Orientation identifiers are introduced, and positive and negative symmetric boundaries are set to achieve full-coverage regularized initialization, balancing convergence and comprehensive distortion feature coverage. Secondly, an adaptive inertial weight adjustment strategy is proposed. Combining the mean and minimum of the population fitness during iteration, large weights are used in the early stages of iteration to focus on global search, while small weights focus on local fine-tuning in the later stages. Simultaneously, the weights are dynamically adjusted based on the relationship between the fitness of individual particles and the population mean to balance the breadth and accuracy of the search. Third, the amplitude-phase step-by-step iterative optimization strategy first fixes the globally optimal phase and optimizes only the amplitude distortion, then fixes the globally optimal amplitude and optimizes only the phase distortion, decomposing the high-dimensional coupled solution into low-dimensional independent solutions, significantly reducing computational complexity. Fourth, a particle mutation mechanism is added to avoid getting trapped in local optima during the iteration process, further improving the accuracy and stability of amplitude-phase distortion estimation. Attached Figure Description

[0011] Figure 1 This is an overall flowchart of the present invention; Figure 2 These are the matched filtering results for first- to third-order amplitude distortion signals according to an embodiment of the present invention. Figure 3 The results of matched filtering for 2nd to 4th order phase distortion signals in this embodiment of the invention; Figure 4 This is the amplitude and phase distortion signal matched filtering result of an embodiment of the present invention; Figure 5 This represents the ideal matched filtering result for radar signals in an embodiment of the present invention. Figure 6 This is the matched filtering result of the radar signal after amplitude and phase distortion correction according to an embodiment of the present invention. Detailed Implementation

[0012] Specific embodiments are given below with reference to the accompanying drawings. These specific embodiments are only used to describe the technical solution of the present invention in detail and are not intended to limit the scope of protection of this application.

[0013] like Figure 1 As shown, this invention provides a method for amplitude and phase distortion correction in an ISAR broadband direct sampling receiver, comprising the following steps: Step 1: Construct an amplitude-phase distortion echo signal model for an ISAR broadband direct sampling receiver; (1) In the formula, This is the column vector of the amplitude and phase distortion echo signal. To increase the number of sampling points in a fast time, For the first A column vector of echo signal amplitude distortion values ​​at each scattering point For the first The ideal echo signal column vector of each scattering point For the first A column vector of echo signal phase distortion values ​​from each scattering point. For element-wise Hadamard product, For exponential operations, Indicates the imaginary part. This represents the total number of scattering points; Considering the multi-factor coupling distortion of the radar system link, the amplitude and phase distortion echo signal model of a single scattering point is as follows: (2) (3) (4) (5) (6) In the formula, For the first Amplitude and phase distortion echo signals at each scattering point The amplitude loss-of-bass matrix, For fast time vectors, For the first The amplitude distortion coefficient vector of the echo signal at each scattering point. For transpose operation, For the first The first scattering point echo signal Amplitude distortion coefficient, , For the first The amplitude coefficient of the echo signal at the scattering point. For discrete rectangular window functions, For the first Ideal phase column vectors of scattering points For the first The radial distance from each scattering point to the radar. At the speed of light, The signal pulse width, For ISAR carrier frequencies, It is a diagonal matrix. The phase distortion basis matrix, For frequency modulation parameter vectors, For the first Step frequency, For the first The vector of phase distortion coefficients of the echo signal at each scattering point. For the first The first scattering point echo signal Phase distortion coefficient.

[0014] Step 2: Perform matched filtering on the amplitude and phase distortion echo signal of the single scattering point. Based on the matched filtering results, construct the fitness function based on the multi-objective optimization strategy of penalized regularized variable attention mechanism (PR-VAMO). Construct a matched filter template signal based on the ideal echo signal: (7) In the formula, To match the filter template signal, This is a conjugate operation. Invert the fast-time vector; The amplitude-phase distortion echo signal is matched and filtered using a matched filter template signal to obtain the amplitude at each fast time sampling point. (8) The fitness function is: (9) (10) (11) (12) In the formula, For the first The variable attention weight vector for each iteration To optimize the target vector, The peak pulse pressure is the sidelobe ratio. Entropy is the energy concentration. To increase attention to amplitude distortion, For phase distortion attention, For the penalty factor vector, To penalize the regularized vector, The amplitude of the first side lobe. Main lobe amplitude, For the first The amplitude of each fast time sampling point Indicates taking the absolute value; The variable attention weight allocation is based on the real-time relative change magnitude of each optimization objective. The formula for calculating the relative change of each optimization objective in the next iteration is: (13) In the formula, For the first During the nth iteration The relative change of each optimization objective , The first , During the nth iteration The values ​​of the optimization objective are: The length of the sliding window. It is a constant with a very small value to avoid the denominator being zero; No. The variable attention weights for each optimization objective in the next iteration are represented as follows: (14) In the formula, For the first During the nth iteration Variable attention weights for each optimization objective. It is a constant with a very small value; The penalty regularization vector is represented as: (15) In the formula, The penalty amount for the main lobe width. A 20dB width penalty is applied. This is the penalty amount for peak reversal detection; The main lobe width penalty rule is as follows: (16) (17) In the formula, The relative change rate of the main lobe width. For the first The main lobe width at the next iteration. This is the initial main lobe width; The 20dB width penalty rule is as follows: (18) (19) In the formula, The relative change rate of width is 20dB. For the first The width was 20dB in the next iteration. The initial width is 20dB. The peak inversion detection penalty rule is as follows: (20) In the formula, The main lobe peak value, This represents the peak value of the sidelobe.

[0015] Step 3: The improved PSO algorithm is used to iteratively optimize the amplitude and phase distortion coefficient vector of the echo signal at each scattering point to obtain the amplitude and phase distortion coefficient estimate vector, thereby obtaining the amplitude and phase distortion estimate of the echo signal at each discrete point. S3.1 Population initialization; Using the amplitude and phase distortion coefficient vectors of the scattering points as particles in the population, an 8-dimensional particle is constructed; The position of each particle is denoted as , , The first The amplitude and phase dimension position of each particle; The positive and negative boundaries of the particle position are set based on the amplitude and phase distortion coefficient error. The negative and positive boundary vectors are expressed as follows: (twenty one) (twenty two) In the formula, For the first The first scattering point echo signal Amplitude distortion coefficient error, For the first The first scattering point echo signal Phase distortion coefficient error; Introducing orientation identifiers allows each particle's dimension to extend in different boundary directions, generating the initial position of each particle, represented as: (twenty three) In the formula, For the first The first particle The initial position of the dimension. , The first Negative and positive boundaries of the dimensional position. The number of particles; For the first The orientation identifier of the dimensional position. Indicates the first The position of the dimension takes values ​​along the positive boundary direction. Indicates the first The position is taken along the negative boundary direction; Based on the initial position of the particle, the fitness of the particle is calculated using the fitness function, and the position of the particle with the maximum fitness is used as the initial optimal position of the individual and the global optimal position. S3.2 During the iteration process, calculate the mean and minimum fitness of the population, and adaptively adjust the particle inertia weights based on the relationship between particle fitness and the mean and minimum fitness of the population. (twenty four) (25) In the formula, , For the first The mean and minimum fitness of the population in the next iteration. For the first The second iteration The fitness of each particle; The formula for updating particle inertia weights is: (26) In the formula, For the first The second iteration The inertial weight of each particle, , These represent the maximum and minimum values ​​of the particle inertia weight; S3.3, Start amplitude distortion coefficient optimization, pause phase distortion coefficient optimization, keep the global optimal position of the phase dimension unchanged, and update the velocity and position of the particle amplitude dimension by the following formula; (27) (28) In the formula, , For the first The second iteration The velocity and position of each particle in the amplitude dimension , For the first The second iteration The velocity and position of each particle in the amplitude dimension For the first The optimal position in the amplitude dimension of each particle. For the global optimal position in the magnitude dimension, , As a learning factor, , A random number with a value between 0 and 1; Based on the updated position of the particle amplitude dimension, calculate the particle fitness and update the individual optimal position and the global optimal position in the amplitude dimension; S3.4 Start phase distortion coefficient optimization, pause amplitude distortion coefficient optimization, keep the global optimal position of amplitude dimension unchanged, and update the velocity and position of particle phase dimension by the following formula; (29) (30) In the formula, , For the first The second iteration The velocity and position of each particle in the phase dimension , For the first The second iteration The velocity and position of each particle in the phase dimension For the first The optimal position of each particle's phase dimension. This represents the globally optimal position in the phase dimension. Calculate particle fitness based on the updated particle phase dimension position, and update the individual optimal position and global optimal position of the phase dimension based on the particle fitness. S3.5. Perform particle mutation after each iteration; Set particle mutation ratio Randomly select particles and reset the positions of each dimension of the selected particles according to the particle mutation formula to prevent them from getting trapped in local optima; (31) In the formula, For the first The first particle Wei's new position for Random numbers within the interval; S3.6 Repeat steps S3.2 to S3.5 to optimize the amplitude and phase distortion coefficients step by step. When the number of iterations reaches the maximum number of iterations, terminate the iteration and output the global optimal position as the amplitude and phase distortion coefficient estimation vector, including the amplitude distortion coefficient estimation vector. and phase distortion coefficient estimate vector Thus, the amplitude distortion estimate and phase distortion estimate of the discrete-point echo signal are obtained; where, , For the first The first scattering point echo signal Estimated values ​​of amplitude and phase distortion coefficients.

[0016] Step 4: Calculate the amplitude and phase compensation factor based on the amplitude and phase distortion estimates of the discrete point echo signals, and use the amplitude and phase compensation factor to correct the amplitude and phase distortion echo signals of each discrete point in real time to obtain the corrected echo signals of each discrete point; superimpose the corrected echo signals of each discrete point to obtain the final corrected echo signal. Based on the amplitude distortion estimate of the echo signal at discrete points, an amplitude compensation factor is constructed for the discrete points. To achieve cancellation of amplitude distortion across all orders, the formula for calculating the amplitude compensation factor is as follows: (32) In the formula, For the first Amplitude compensation factor for each scattering point For the first The amplitude distortion estimate of the echo signal at each scattering point is constructed based on the vector of amplitude distortion coefficient estimates, and the calculation formula is as follows: (33) Based on the phase distortion estimate of the echo signal at discrete points, a phase compensation factor for discrete points is constructed. To achieve full-order phase distortion cancellation, the formula for calculating the phase compensation factor is as follows: (34) In the formula, For the first Phase compensation factor for each scattering point For the first The estimated phase distortion value of the echo signal at each scattering point is constructed based on the vector of estimated phase distortion coefficients, and the calculation formula is as follows: (35) Based on amplitude and phase compensation factors, the amplitude and phase distortion echo signal is corrected to cancel out amplitude and phase distortion. The corrected echo signal is as follows: (36) In the formula, For the first Echo signal after correction of each scattering point; The corrected echo signals from each scattering point are superimposed to obtain the final corrected echo signal: (37) In the formula, This is the final corrected echo signal.

[0017] Example This embodiment uses an X-band analog broadband radar echo signal with a carrier frequency of 10 GHz, a signal bandwidth of 400 MHz, a radar pulse width of 0.4 μs, a sampling rate of 2 GHz, and 8000 sampling points. Since zero-order amplitude distortion and first-order phase distortion have relatively small impacts on the quality of the radar echo signal, only 1-3 order amplitude distortion and 2-4 order phase distortion are used for modeling. Amplitude distortion of each order is randomly generated between [0 dB, 15 dB], and phase distortion of each order is randomly generated between [0 rad, 10 rad]. Figure 2 The results are from matched filtering of 1st to 3rd order amplitude distortion signals. Figure 3 The graph shows the results of matched filtering for phase-distorted signals of orders 1-4. Figure 4 The result of matched filtering for amplitude and phase distortion signals. Figure 5 This represents the ideal matched filtering result for the radar signal in this embodiment. Figure 6 This is the matched filtering result of the radar signal after amplitude and phase distortion correction in this embodiment.

[0018] Table 1. Parameters of the Improved Particle Swarm Optimization Algorithm

[0019] Table 2 Amplitude and phase distortion correction results

[0020] As shown in Table 2, this invention provides more accurate estimations of 1st-3rd order amplitude distortion and 2nd-4th order phase distortion. The error for 1st order amplitude distortion is approximately 18.7%, for 2nd order it is only 0.9%, and for 3rd order it is approximately 2.9%, showing a consistent overall trend. The error for 2nd order phase distortion is approximately 7.5%, while the errors for 3rd and 4th order phase distortion are both controlled within 3%. Figure 6 It can be seen that the corrected echo signal improves the main lobe broadening coefficient and peak sidelobe ratio, and enhances the range alignment accuracy, providing strong support for ISAR imaging.

[0021] The amplitude and phase distortion echo signal model of the ISAR broadband direct sampling receiver constructed in this invention fully covers the nonlinear coupling characteristics of 1st-3rd order amplitude distortion and 2nd-4th order phase distortion, providing a theoretical basis for accurate estimation. Based on the PR-VAMO fusion multi-objective fitness function, the core distortion features are focused through dynamic attention weights, and penalized regularization is introduced to constrain signal distortion, significantly improving the estimation stability and targeting. An improved particle swarm optimization algorithm is adopted, which achieves global optimization through adaptive weights, step-by-step iteration and particle mutation mechanism, breaking through the local optimum bottleneck of traditional algorithms and greatly improving the accuracy of high-dimensional distortion solution.

[0022] Any aspects not covered in this invention are applicable to existing technologies.

Claims

1. A method for amplitude and phase distortion correction in an ISAR broadband direct sampling receiver, characterized in that, Includes the following steps: Step 1: Construct an amplitude-phase distortion echo signal model for an ISAR broadband direct sampling receiver; (1) In the formula, This is the column vector of the amplitude and phase distortion echo signal. For the first A column vector of echo signal amplitude distortion values ​​at each scattering point For the first The ideal echo signal column vector of each scattering point For the first A column vector of echo signal phase distortion values ​​from each scattering point. For element-wise Hadamard product, For exponential operations, Indicates the imaginary part. This represents the total number of scattering points; The amplitude-phase distortion echo signal model at a single scattering point is as follows: (2) (3) (4) (5) (6) In the formula, For the first Amplitude and phase distortion echo signals at each scattering point The amplitude loss-of-bass matrix, For fast time vectors, For the first The amplitude distortion coefficient vector of the echo signal at each scattering point. For transpose operation, For the first The first scattering point echo signal Amplitude distortion coefficient, here , For the first The amplitude coefficient of the echo signal at the scattering point. For discrete rectangular window functions, For the first Ideal phase column vectors of scattering points For the first The radial distance from each scattering point to the radar. At the speed of light, The signal pulse width, For ISAR carrier frequencies, It is a diagonal matrix. The phase distortion basis matrix, For frequency modulation parameter vectors, For the first Step frequency, here , For the first The vector of phase distortion coefficients of the echo signal at each scattering point. For the first The first scattering point echo signal Phase distortion coefficient, here ; Step 2: Perform matched filtering on the amplitude and phase distortion echo signal of the single scattering point to obtain the amplitude of each time sampling point; construct the fitness function based on the matched filtering results as follows: (9) (10) (11) (12) In the formula, For the first The variable attention weight vector for each iteration To optimize the target vector, For the penalty factor vector, To penalize the regularized vector, The peak pulse pressure is the sidelobe ratio. Entropy is the energy concentration. To increase attention to amplitude distortion, For phase distortion attention, The amplitude of the first side lobe. Main lobe amplitude, For the first The amplitude of each fast time sampling point Indicates taking the absolute value; No. The variable attention weights for each optimization objective in the next iteration are: (14) In the formula, For the first During the nth iteration Variable attention weights for each optimization objective. For the first During the nth iteration The relative change of each optimization objective It is a constant with a very small value; The penalty regularization vector is represented as: (15) (16) (18) (20) In the formula, The penalty amount for the main lobe width. A 20dB width penalty is applied. This is the penalty amount for peak reversal detection. The relative change rate of the main lobe width. The relative change rate of width is 20dB. The main lobe peak value, The peak value of the side lobe; Step 3: The PSO algorithm is used to iteratively optimize the amplitude and phase distortion coefficient vector of the echo signal at each scattering point to obtain the amplitude and phase distortion coefficient estimate vector, thereby obtaining the amplitude and phase distortion estimate of the echo signal at each discrete point. Step 4: Calculate the amplitude and phase compensation factor based on the amplitude and phase distortion estimates of the discrete point echo signals, and use the amplitude and phase compensation factor to correct the amplitude and phase distortion echo signals of each discrete point in real time to obtain the corrected echo signals of each discrete point; superimpose the corrected echo signals of each discrete point to obtain the final corrected echo signal. The formula for calculating the amplitude compensation factor at each discrete point is: (32) In the formula, For the first Amplitude compensation factor for each scattering point For the first Estimated amplitude distortion of echo signal at each scattering point; The formula for calculating the phase compensation factor at each discrete point is: (34) In the formula, For the first Phase compensation factor for each scattering point For the first Estimated phase distortion of echo signal at each scattering point; The echo signals after correction at each discrete point are: (36) In the formula, For the first Echo signal after correction of each scattering point; The final corrected echo signal is: (37) In the formula, This is the final corrected echo signal.

2. The ISAR broadband direct sampling receiver amplitude and phase distortion correction method according to claim 1, characterized in that, The iterative process of the PSO algorithm includes: S3.1 Population initialization; The amplitude and phase distortion coefficient vectors of the scattering points are treated as particles, each containing 8 dimensions; the initial positions of each particle are represented as follows: (23) In the formula, For the first The first particle The initial position of the dimension. , The first Negative and positive boundaries of the dimensional position. The number of particles; For the first The orientation identifier of the dimensional position. Indicates the first The position of the dimension takes values ​​along the positive boundary direction. Indicates the first The position is taken along the negative boundary direction; S3.2 Calculate the particle fitness, the mean and minimum population fitness, and adaptively adjust the particle inertia weight according to equation (26); (26) In the formula, For the first The second iteration The inertial weight of each particle, , These represent the maximum and minimum values ​​of the particle inertia weight. For the first The second iteration The fitness of each particle , For the first The mean and minimum fitness of the population in each iteration; S3.3, Start amplitude distortion coefficient optimization, pause phase distortion coefficient optimization, keep the global optimal position of phase dimension unchanged, and update the velocity and position of particle amplitude dimension through equations (27) and (28); (27) (28) In the formula, , For the first The second iteration The velocity and position of each particle in the amplitude dimension , For the first The second iteration The velocity and position of each particle in the amplitude dimension For the first The optimal position in the amplitude dimension of each particle. For the global optimal position in the magnitude dimension, , As a learning factor, , A random number with a value between 0 and 1; Based on the updated position of the particle amplitude dimension, calculate the particle fitness and update the individual optimal position and the global optimal position in the amplitude dimension; S3.4 Start phase distortion coefficient optimization, pause amplitude distortion coefficient optimization, keep the global optimal position of amplitude dimension unchanged, and update the velocity and position of particle phase dimension through equations (29) and (30); (29) (30) In the formula, , For the first After the nth iteration The velocity and position of each particle in the phase dimension , For the first After the nth iteration The velocity and position of each particle in the phase dimension For the first The optimal position of each particle's phase dimension. This represents the globally optimal position in the phase dimension. Calculate particle fitness based on the updated particle phase dimension position, and update the individual optimal position and global optimal position of the phase dimension based on the particle fitness. S3.

5. Perform particle mutation after each iteration; S3.6 Repeat steps S3.2 to S3.5 to optimize the amplitude and phase distortion coefficients step by step. When the number of iterations reaches the maximum number of iterations, terminate the iteration and output the global optimal position as the estimated value vector of amplitude and phase distortion coefficients.

3. The ISAR broadband direct sampling receiver amplitude and phase distortion correction method according to claim 2, characterized in that, Positive and negative boundary vectors of particle position , Represented as: (21) (22) In the formula, For the first The first scattering point echo signal Amplitude distortion coefficient error, For the first The first scattering point echo signal Phase distortion coefficient error, .

Images