Distributed networking radar strong and weak target joint coherent processing method

Abstract

The invention discloses a joint coherent processing method for strong and weak targets of a distributed networking radar, and aims to solve the problem of poor coherent effect caused by low parameter estimation precision of an existing coherent processing method adopted by a distributed networking radar system. A coherent parameter estimation problem is converted into a cosine similarity maximization problem, so that decoherence introduced by radar splitting is compensated, a'strong target-weak target 'interactive optimization criterion is constructed, a coherent parameter roughly estimated by a strong target is utilized to compensate the phase of a weak target signal, and the phase of the weak target signal is optimized. And the signal phase is fed back and adjusted by using the synthesis effect of the weak target, so that relatively accurate parameter estimation under the condition of a low signal-to-noise ratio is realized, coherent synthesis of distance-Doppler domain signals is completed, and the detection capability of a distributed networking radar system is improved.

Description

Technical Field

[0001] This invention belongs to the field of radar signal processing technology, specifically relating to a method for joint coherent processing of strong and weak targets in distributed networked radar. Background Technology

[0002] For traditional single-radar systems, pursuing higher performance inevitably involves improving amplifier power limits or efficiency. However, this method is limited by factors such as equipment technology and platform size, making it increasingly challenging to improve radar performance. Addressing the inherent performance bottlenecks of single radars, radar detection systems are shifting from current single-radar detection to distributed, multi-radar collaborative detection. Compared to monostatic radars, distributed networked radars can collect target scattering information from multiple dimensions such as space, frequency, and polarization, possessing the potential to expand the functionality and performance of radar systems.

[0003] Distributed networked radar cooperative detection refers to multiple radars distributed in space jointly undertaking detection tasks and processing all data received by each radar in the detection scenario. Spatial distribution and joint data processing are the prominent features of this detection method. If all cooperating radars are considered as a whole, they constitute a networked radar system. The individual radar units in the network can be placed in different locations and used together in a coherent and cooperative manner, thereby significantly enhancing target energy. The key technology for this is coherent signal synthesis. Distributed networked radar coherent synthesis aims to achieve enhanced received signal strength through a "small accumulations add up" approach. Networked radar coherent synthesis is equivalent to adjusting the transmit and receive times and phases of each radar to calibrate the decoherence introduced by the radar distribution.

[0004] Distributed networked radar systems, as an extension of single radar systems, possess more significant technological advantages, but still face technical challenges in coherent signal synthesis. Extensive research has analyzed coherent synthesis methods under high signal-to-noise ratio (SNR) conditions. However, parameter estimation is more difficult under low SNR conditions, necessitating the development of new analytical methods. Existing methods for coherent synthesis of networked radar signals fall into two categories: one involves estimation by detecting the peak value of the target itself in the echo, such as peak extraction methods; the other involves reconstructing the signal and then using the reconstructed signal for parameter estimation.

[0005] However, the two existing methods mentioned above have the following drawbacks: Disadvantage 1: Existing coherent parameter estimation has high estimation accuracy under high signal-to-noise ratio conditions, but when the target scattering point is undetectable, the parameter estimation accuracy is low or even fails.

[0006] Disadvantage 2: In existing coherent parameter estimation, the criterion for judging whether the cross-correlation coefficient has been estimated is the cross-correlation coefficient of the signal. However, the cross-correlation coefficient of the signal is greatly affected by the signal amplitude. When performing coherent parameter estimation, more attention should be paid to the matching degree of the signal phase to reduce the influence of the signal amplitude. Summary of the Invention

[0007] To address the aforementioned problems in the existing technology, this invention provides a method for joint coherent processing of strong and weak targets in distributed networked radar. The technical problem to be solved by this invention is achieved through the following technical solution: This invention provides a method for joint coherent processing of strong and weak targets in distributed networked radar systems, the method comprising: The strong target echo signals are acquired by the reference radar and each radar to be registered, and the strong target echo signals are sequentially processed by carrier frequency removal, pulse compression and coherent accumulation to obtain the corresponding strong processed signals. For each radar to be registered, the registration process includes: constructing a range-dimensional registration model and a Doppler registration model to compensate for the spatial differences between the reference radar and the radar to be registered based on the strong target coherent synthesis parameters; constructing a range-Doppler registration problem between the reference radar and the radar to be registered based on the strong processed signals corresponding to the reference radar and the radar to be registered, and converting the range-Doppler registration problem into a cosine similarity maximization problem; solving the cosine similarity maximization problem based on the range-dimensional registration model and the Doppler registration model to obtain the strong target coherent synthesis parameters, and registering the radar to be registered based on the strong target coherent synthesis parameters; Each registered radar acquires the echo signal of the corresponding weak target, and sequentially performs carrier frequency removal, pulse compression, and coherent accumulation processing on the echo signal of the weak target to obtain the corresponding weak processed signal. All weak signals are vectorized, and the fused signal is obtained by weighted fusion based on all vectorization results. For each registered radar, the discrete phase compensation process includes: calculating the maximum acceptable phase difference of the registered radar based on the fused signal; calculating the signal-to-noise ratio (SNR) gain based on the fused signal corresponding to the registered radar; calculating the acceptable SNR gain based on the fused signal corresponding to the maximum acceptable phase difference; determining whether the SNR gain exceeds the acceptable SNR gain; if it does, outputting the fused signal; if it does not, compensating the corresponding weak processing signal based on the maximum acceptable phase difference of the registered radar; and returning to the step of vectorizing all weak processing signals until the maximum number of cyclic compensations is met.

[0008] In one embodiment of the present invention, the strong target coherent synthesis parameters include an angle rotation matrix and a translation vector; The range-dimensional registration model constructed based on the strong target coherent synthesis parameters is expressed as follows: ; in, This represents the range dimension index of the strong target signal in the radar to be registered. This represents the range resolution unit of the radar to be registered. , These represent the elevation and azimuth angles of the strong target relative to the radar to be registered, respectively. Indicates the range resolution unit of the reference radar. , These represent the elevation and azimuth angles of the strong target relative to the reference radar, respectively. This represents the range dimension index of a strong target signal in the reference radar. Represents the angle rotation matrix. Represents the translation vector; The Doppler registration model constructed based on the strong target coherent synthesis parameters is expressed by the following formula: ; in, The Doppler index represents the strong target signal of the radar to be registered. The Doppler index representing the strong target signal of the reference radar. , , These represent the radial velocities of the strong target relative to the reference radar and the radar to be registered, respectively. and These represent the velocity resolution units of the reference radar and the radar to be registered, respectively.

[0009] In one embodiment of the present invention, the range-Doppler registration problem between the reference radar and the radar to be registered, constructed based on the strong processed signals corresponding to the reference radar and the radar to be registered, is expressed by the following formula: ; in, This indicates the strong processing signal corresponding to the reference radar. This indicates the strong processing signal corresponding to the radar to be registered. This represents the ratio of the amplitude values ​​of the echo signals received by different radars. This indicates the error caused by noise.

[0010] In one embodiment of the present invention, the distance-Doppler registration problem is transformed into a cosine similarity maximization problem, expressed by the following formula: ; in, Indicates parameters related to coherent synthesis The function, , , , The angles by which the rotation matrix rotates about the X-axis, Y-axis, and Z-axis are represented. , , The distances of the translation vector along the X, Y, and Z axes are represented. Represents the cosine similarity function. , This indicates the operation of finding the 2-norm.

[0011] In one embodiment of the present invention, solving the cosine similarity maximization problem to obtain strong objective coherent synthesis parameters includes: The strong objective coherent synthesis parameters are obtained by solving the cosine similarity maximization problem using a genetic algorithm.

[0012] In one embodiment of the present invention, all weakly processed signals are vectorized, and a weighted fusion process is performed based on all vectorization results to obtain a fused signal, including: All weak processing signals are vectorized and combined based on all vectorization results to form a new processing signal; Based on the new processed signal and fusion weights, construct a fusion signal optimization problem; The fusion signal optimization problem is transformed into a power signal optimization problem; Solve the power signal optimization problem to obtain the optimal fusion weights; The fusion calculation is performed based on the optimal fusion weights and the new processed signal, and the fusion calculation result is then vectorized and inversely processed to obtain the fused signal.

[0013] In one embodiment of the present invention, a fusion signal optimization problem is constructed based on the new processed signal and fusion weights, expressed by the following formula: ; in, Indicates the initial fusion signal. Indicates the fusion weight. Expressing the request conjugate, This indicates a new processing signal. In one embodiment of the present invention, the fusion signal optimization problem is transformed into a power signal optimization problem, expressed by the following formula: ; in, Indicates the first digit in the fused signal The fusion weights corresponding to each component Expressing the request conjugate, This represents the correlation matrix between all registered radars. Indicates the first digit in the fused signal The new processing signal corresponds to each component. Indicates the number of radars to be registered. This indicates the transpose operation.

[0014] In one embodiment of the present invention, a fusion calculation is performed based on the optimal fusion weights and the new processed signal, and the fusion calculation result is then subjected to vectorized inverse processing to obtain the fused signal, expressed by the following formula: ; in, The first digit of the fused signal One portion, Indicates the first digit in the fused signal The optimal fusion weights for each component Expressing the request conjugate, This indicates a new processing signal. This indicates the inverse process of vectorization. This indicates the transpose operation.

[0015] In one embodiment of the present invention, the formula for calculating the maximum acceptable phase difference is: ; in, Indicates the maximum acceptable phase difference. This indicates an acceptable loss of synthetic gain. Calculated based on the fused signal, This indicates the number of radars to be registered. The beneficial effects of this invention are: This invention proposes a joint strong and weak target coherent processing method for distributed networked radar. Addressing the problem of low parameter estimation accuracy leading to poor coherent synthesis results in existing coherent processing methods for distributed networked radar systems, this method transforms the coherent parameter estimation problem into a cosine similarity maximization problem to compensate for the decoherence introduced by radar splitting. Furthermore, an interactive optimization criterion of "strong target-weak target" is constructed. The coherent parameters roughly estimated by the strong target are used to compensate for the phase of the weak target signal, and the synthesis effect of the weak target is used to feedback and adjust the signal phase. This achieves relatively accurate parameter estimation under low signal-to-noise ratio conditions, completing the coherent synthesis of range-Doppler domain signals and improving the detection capability of the distributed networked radar system.

[0016] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Attached Figure Description

[0017] Figure 1 This is a flowchart illustrating a distributed networked radar joint coherent processing method for strong and weak targets provided in an embodiment of the present invention; Figure 2 This is a schematic diagram of the signal processing process corresponding to strong and weak targets provided in an embodiment of the present invention; Figure 3 This is a registration diagram of a reference radar and a radar to be registered, provided in an embodiment of the present invention. Figure 4 This is a more detailed schematic diagram of the processing procedure of the distributed networked radar joint coherent processing method for strong and weak targets provided in the embodiments of the present invention. Detailed Implementation

[0018] The present invention will be further described in detail below with reference to specific embodiments, but the implementation of the present invention is not limited thereto.

[0019] Please see Figure 1 This invention provides a method for joint coherent processing of strong and weak targets in a distributed networked radar system, specifically including the following steps: S10. The reference radar and each radar to be registered acquire the corresponding strong target echo signal, and perform carrier frequency removal, pulse compression and coherent accumulation processing on the strong target echo signal in sequence to obtain the corresponding strong processed signal.

[0020] In this embodiment of the invention, the reference radar and each radar to be registered employ the same signal processing procedure, as follows: Figure 2 As shown, taking the signal processing of a reference radar as an example: Assume the transmitted signal of the reference radar in the system is a linear frequency modulated signal, where the first... The first strong target echo signal corresponding to each pulse signal is: (1); in, Indicates the echo signal. For pulse width, For the reference radar center frequency, For frequency modulation slope, , For signal bandwidth, To save time, For slow time, subscript Indicates the pulse sequence number. For the initial phase, The amplitude of the pulse. Represents the imaginary unit. This is a rectangle function.

[0021] No. The echo signal corresponding to each pulse signal, after carrier frequency removal (down-conversion processing), is as follows: (2); in, The signal amplitude of the echo signal. For reference, the time delay of the radar receiving the echo signal, At the speed of light, The radial velocity of the strong target in the direction of the reference radar line of sight. In order to be in Constantly monitor the distance between the radar and strong targets. For wavelength, It is complex Gaussian white noise. This is the Doppler frequency.

[0022] No. Each pulse signal corresponds to the signal after carrier frequency removal, and after pulse compression, it can be represented as: (3); in, represents the impulse response of the impulse compression filter, and * represents the convolution operator. It can be represented as: (4); In radar signal processing, coherent accumulation (CPI) is achieved by performing Fast Fourier Transform (FFT) on M pulses at the same range gate. Within each Coherent Processing Interval (CPI), FFT is performed on each column of the signal matrix to perform the coherent accumulation operation. (5); in, Indicates the first The coherent accumulation result corresponding to the column signal is a discrete signal. Represents the first element in the signal matrix. The corresponding pulse signal of the first The signal matrix consists of M pulses, each corresponding to a compressed pulse signal, forming a two-dimensional range-Doppler data matrix composed of the range dimension and the Doppler dimension. Each column of range-Doppler signals in the signal matrix is ​​processed using formula (5) to obtain the strong processed signal corresponding to the reference radar.

[0023] Compared to a single pulse echo signal, performing a fast Fourier transform on each column of data and coherently accumulating the signal improves the signal-to-noise ratio and makes it easier to obtain target information under conditions of low signal-to-noise ratio.

[0024] For each radar to be registered, a reference radar similar to the one described above is used. Figure 2 The signal processing process shown is different in that the processing objects are the radar to be registered and strong targets, in order to obtain the strong processing signal corresponding to each radar to be registered.

[0025] S20. For each radar to be registered, the registration process includes: constructing a range-dimensional registration model and a Doppler registration model to compensate for the differences in radar spatial position between the reference radar and the radar to be registered based on the strong target coherent synthesis parameters; constructing a range-Doppler registration problem between the reference radar and the radar to be registered based on the strong processed signals corresponding to the reference radar and the radar to be registered, and converting the range-Doppler registration problem into a cosine similarity maximization problem; solving the cosine similarity maximization problem based on the range-dimensional registration model and the Doppler registration model to obtain the strong target coherent synthesis parameters, and registering the radar to be registered based on the strong target coherent synthesis parameters.

[0026] In this embodiment of the invention, the registration process performed by each radar to be registered is the same. Assuming the reference radar is A and the radar to be registered is B, let's take the registration of reference radar A and radar B to be registered as an example: Given a strong target with a high signal-to-noise ratio, both reference radar A and radar B to be registered will receive echo signals from this strong target. The received echoes are processed using steps S10 and S20 respectively to obtain the corresponding strong target processed signals. Since the radar positions are unknown, assume the strong target's position coordinates in reference radar A are... The position coordinates in radar B to be registered are Then, the position coordinates of the strong target in reference radar A and the radar to be registered B will have the following relationship: (6); Where H represents the angular rotation matrix, This represents the translation vector.

[0027] (7); for The translation vector. This indicates the angle of rotation of the coordinate axis around the X-axis. This indicates the angle of rotation of the coordinate axis around the Y-axis. This indicates the angle of rotation of the coordinate axis around the Z-axis.

[0028] and This represents the position coordinates in a Cartesian coordinate system, while radar measurements are typically position vectors in a spherical coordinate system. Therefore, it needs to be converted to spherical coordinates: (8); in, , , These represent the range, elevation angle, and azimuth angle of the strong target relative to reference radar A, respectively. Similarly, , , These are the range, elevation angle, and azimuth angle of the strong target relative to the radar B to be registered.

[0029] The target's range and radial velocity relative to reference radar A and the radar to be registered B are as follows: (9); (10); in, , These represent the radial velocities of the strong target relative to reference radar A and the radar B to be registered, respectively. , These are the range resolution units for reference radar A and radar B to be registered, respectively. , These are the range dimension indices of strong target signals in reference radar A and the radar B to be registered, respectively. and These are the velocity resolution units for reference radar A and radar B to be registered, respectively. and Let A and B be the Doppler indices of the strong target signals in reference radar A and the radar B to be registered, respectively. Then the range-dimensional registration model can be expressed as: (11); The Doppler registration model can be expressed as: (12); in, Based on the relationship between the range dimension and Doppler dimension of a strong target relative to reference radar A and the radar B to be registered, the registration equation between reference radar A and the radar B to be registered is as follows: (13); This is done to compensate for differences in radar spatial position, ensuring that the target signal obtained by the same target from different radars is within the same range and Doppler cell. The parameters to be estimated are the values ​​of the angle rotation matrix H and the translation vector r; these parameters are the coherent synthesis parameters for strong targets.

[0030] Because the distance and RCS (Radar Cross Section) of the same target are different relative to different radars, this will lead to different amplitudes of the echo signals received by each radar. This represents the ratio of the amplitude values ​​of the echo signals received by different radars. This indicates the strongly processed signal obtained by reference radar A after passing through S10. If the strong processed signal obtained by radar B to be registered after passing through S10 is given, then the following relationship exists: (14); Furthermore, the range-Doppler registration problem between the reference radar A and the radar B to be registered, constructed based on the strong processed signals corresponding to the reference radar A and the radar B to be registered, is expressed by the following formula: (15); in, This indicates the strong processing signal corresponding to the reference radar. This indicates the strong processing signal corresponding to the radar to be registered. This represents the ratio of the amplitude values ​​of the echo signals received by different radars. This indicates the error caused by noise.

[0031] Cosine similarity can be used to characterize the phase similarity of signals; the greater the cosine similarity, the closer the phases of the two signals. The parameters that need to be compensated for in the coherent synthesis of this embodiment of the invention are as follows: (16); The registration problem of distance dimension-Doppler dimension is transformed into a cosine similarity maximization problem, which can be expressed by the formula: (17); in, Indicates parameters related to coherent synthesis The function, , , , The angles by which the rotation matrix rotates about the X-axis, Y-axis, and Z-axis are represented. , , The distances of the translation vector along the X, Y, and Z axes are represented. Represents the cosine similarity function. , This represents the operation of calculating the 2-norm. It can be seen that by using cosine similarity processing, it becomes... To reach the maximum, determine the required coherent synthesis parameters. The coherent synthesis parameters for strong targets can be determined using these coherent synthesis parameters. , In this embodiment of the invention, a genetic algorithm can be used to solve formula (17) to obtain the coherent synthesis parameters. Thus, the parameters for strong target coherent synthesis are obtained. , .

[0032] Finally, based on the obtained strong target coherent synthesis parameters... , Coherent alignment is performed on the radars to be registered, so that they can be registered as follows: Figure 3 As shown. All radars to be registered follow a registration process similar to that between reference radar A and radar B to be registered.

[0033] S30. Each registered radar acquires the echo signal of the corresponding weak target, and sequentially performs carrier frequency removal, pulse compression, and coherent accumulation processing on the echo signal of the weak target to obtain the corresponding weak processed signal.

[0034] Each registered radar obtained after S20 uses a method similar to the reference radar in S10, such as... Figure 2 The signal processing procedure shown is different in that the processing objects are the registered radar and weak targets. For the specific implementation, please refer to S10, which will not be repeated here, to obtain the corresponding weak processing signal.

[0035] S40. Vectorize all weak signals and perform weighted fusion processing based on all vectorization results to obtain the fused signal.

[0036] This invention embodiment vectorizes all weakly processed signals and weights and fuses them according to all vectorization results to obtain a fused signal. The process includes: vectorizing all weakly processed signals and combining all vectorization results to form a new processed signal; constructing a fusion signal optimization problem based on the new processed signal and fusion weights; converting the fusion signal optimization problem into a power signal optimization problem; solving the power signal optimization problem to obtain the optimal fusion weights; performing fusion calculation based on the optimal fusion weights and the new processed signal, and performing inverse vectorization processing on the fusion calculation result to obtain the fused signal. More specifically: After N registrations, the radar performs S30 processing to obtain the corresponding weak target processed signals, which are represented as follows: , , ,…, Vectorizing the signal yields: (18); in, This is a vectorization operator, which rearranges the signal column by column into a long column vector. Assume the original signal matrix has a size of... Then the vectorized processing , , ,…, The vector length is .

[0037] At this point, the phase difference in the signal differs when the target appears in different range Doppler cells. Assume the target's fusion weight in the fused signal is... Then, based on the new processed signal and fusion weights, a fusion signal optimization problem is constructed, expressed by the formula: (19); in, Indicates the initial fusion signal. This represents the fusion weights, and is an N×1 weight fusion vector. Expressing the request conjugate, This indicates a new processing signal. =[ , , ,…, ] T The purpose of achieving cooperative detection fusion through formula (19) is to improve the signal-to-noise ratio of the fused target. The core idea is to keep the signal power of the current range-Doppler unit fixed and then minimize the power of all signals.

[0038] The target's position in the fused signal can be obtained using formula (19). Each component corresponds to The signal power after weighted fusion is: (20); in, Let N be the correlation matrix between the registered radars. Let N be the second-order noise statistics of the registered radars. Indicates the first digit in the fused signal The new processing signal corresponds to each component. express The Middle One portion, express The Middle One portion, express The Middle One portion, express The Middle One portion, The value range is 1~ , This represents the number of radars to be registered. Therefore, the first component in formula (20) is fixed, i.e. To keep the value fixed, and to minimize the variance of the weighted fused signal power, we need to find... Minimum value, equivalent to finding The minimum value of is ultimately transformed into a power signal optimization problem, which is expressed by the formula: (twenty one); in, Indicates the first digit in the fused signal The fusion weights corresponding to each component Expressing the request conjugate, This represents the correlation matrix between all registered radars. Indicates the first digit in the fused signal The new processing signal corresponds to each component. This indicates the transpose operation.

[0039] Based on the matrix inversion rule, the optimal solution for fusion weights can be expressed as: (twenty two); in, Expressing the request The reverse, It is a complex conjugate.

[0040] Finally, fusion calculations are performed based on the optimal fusion weights and the new processed signal. The fusion calculation results are then vectorized and inversely processed to obtain the fused signal, expressed by the formula: (twenty three); in, The first digit of the fused signal One portion, Indicates the first digit in the fused signal The optimal fusion weights for each component Expressing the request conjugate, This indicates a new processing signal. This indicates the inverse process of vectorization. This indicates the transpose operation. Formula (23) can be used to achieve the coherent fusion result of N registered radars for each component in the fused signal under the corresponding optimal fusion weight, thus obtaining the final fused signal.

[0041] S50. For each registered radar, the discrete phase compensation process includes: calculating the maximum acceptable phase difference of the registered radar based on the fused signal; calculating the signal-to-noise ratio (SNR) gain based on the fused signal corresponding to the registered radar; calculating the acceptable SNR gain based on the fused signal corresponding to the maximum acceptable phase difference; determining whether the SNR gain exceeds the acceptable SNR gain; if it does, outputting the fused signal; if it does not, compensating the corresponding weak processing signal based on the maximum acceptable phase difference of the registered radar; returning to the step of vectorizing all weak processing signals until the maximum number of cyclic compensations is met.

[0042] After parameter estimation using the S20 strong target and compensation of the echo signals, the envelopes of all echo signals are aligned, and the phases are initially aligned. However, since the performance of the DCAR (Distributed Cooperative Active Reflector) depends on high-precision CPs (Channel Parameters) estimation, phase adjustment is necessary. First, the relative relationship between phase deviations in multi-radar strong-weak target compensation is derived. Second, this embodiment defines an acceptable synthetic gain loss and constructs an inequality equation from this loss. Finally, the phase tolerance is derived by solving the inequality equation. The phase tolerance is a discrete phase value, which can be defined as the maximum acceptable phase difference between any two radars. Based on this phase tolerance, the phases of each radar can be adaptively adjusted according to the relative phase relationship. The specific derivation process of this method is as follows: After parameter compensation for each radar to be registered with the S20 strong target, for example, the phase difference between the echo signal received by the Nth registered radar and the echo signal received by the reference radar is: (twenty four); in, Let the phase difference between the Nth registered radar and the reference radar A be roughly compensated. The signal power after coherent synthesis is analyzed and derived. It is assumed that the signals of each registered radar, after compensation using strong target parameters, have the same amplitude and only exhibit phase deviation. For a radar system consisting of N registered radars, the signal power of the N registered radars after coherent synthesis is... It can be represented as: (25); Accordingly, the synthesized theoretical signal power Each channel is aligned with a reference radar, and can be represented as: (26); Due to phase deviation, there is a certain gain loss. The acceptable combined gain loss is defined as... , can be represented as: (27); According to Euler's formula, it can be rewritten as: (28); By solving the inequality equations, a specific phase tolerance can be obtained. This phase tolerance is defined as the maximum acceptable phase difference between any two registered radars, and can be derived as follows: (29); This phase tolerance is a discrete phase value used to guide the compensation of each radar. For any registered radar, a phase value can be cyclically compensated to satisfy the obtained phase tolerance; this cyclically compensated phase value is the calculated phase tolerance. In the phase region... Within this range, the maximum cyclic compensation number can be expressed as: (30); After processing by S50, the phase of weak targets can be aligned in a set of compensations, realizing the correction of the phase deviation of each weak target based on the phase relationship between multiple registered radars, and significantly reducing the number of compensation cycles. In each compensation cycle: the signal-to-noise ratio (SNR) gain is calculated based on the fused signal corresponding to the registered radar, and the acceptable SNR gain is calculated based on the fused signal corresponding to the maximum acceptable phase difference. It is determined whether the SNR gain of the registered radar exceeds the acceptable SNR gain. If it does, the fused signal is output, indicating that the coherent gain loss is within the acceptable gain loss range and the parameter estimation has reached the corresponding accuracy, which means that coherent synthesis is completed. If it does not exceed the acceptable gain loss range, the weak processing signal obtained by S40 is compensated using the maximum acceptable phase difference. Based on the compensated weak processing signal, the step of vectorizing all weak processing signals is returned, and the subsequent discrete phase compensation process continues until the maximum number of compensation cycles is met.

[0043] A more complete and detailed flowchart of the embodiments of the present invention is shown below, for example... Figure 4 As shown.

[0044] In summary, the distributed network radar joint strong and weak target coherent processing method proposed in this invention addresses the problem of low parameter estimation accuracy leading to poor coherent synthesis results in existing coherent processing methods for distributed network radar systems. To achieve more accurate parameter estimation, the coherent parameter estimation problem is transformed into a cosine similarity maximization problem, thereby compensating for the decoherence introduced by radar splitting. A "strong target-weak target" interactive optimization criterion is constructed, using the coherent parameters roughly estimated by the strong target to compensate for the phase of the weak target signal, and then using the synthesis effect of the weak target to feedback and adjust the signal phase, thereby achieving more accurate parameter estimation under low signal-to-noise ratio conditions, completing the coherent synthesis of range-Doppler domain signals, and improving the detection capability of the distributed network radar system.

[0045] In the description of this invention, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include one or more of that feature. In the description of this invention, "a plurality of" means two or more, unless otherwise explicitly specified.

[0046] Although the invention has been described herein in conjunction with various embodiments, those skilled in the art, by reviewing the specification and accompanying drawings, will understand and implement other variations of the disclosed embodiments in carrying out the claimed invention. In the specification, the word "comprising" does not exclude other components or steps, and "a" or "an" does not exclude a plurality. While certain measures are described in different embodiments, this does not mean that these measures cannot be combined to produce good results.

[0047] The above description, in conjunction with specific preferred embodiments, provides a further detailed explanation of the present invention. It should not be construed that the specific implementation of the present invention is limited to these descriptions. For those skilled in the art, various simple deductions or substitutions can be made without departing from the concept of the present invention, and all such modifications and substitutions should be considered within the scope of protection of the present invention.

Claims

1. A method for joint coherent processing of strong and weak targets in a distributed networked radar system, characterized in that, The method includes: The strong target echo signals are acquired by the reference radar and each radar to be registered, and the strong target echo signals are sequentially processed by carrier frequency removal, pulse compression and coherent accumulation to obtain the corresponding strong processed signals. For each radar to be registered, the registration process includes: constructing a range-dimensional registration model and a Doppler registration model to compensate for the spatial differences between the reference radar and the radar to be registered based on the strong target coherent synthesis parameters; constructing a range-Doppler registration problem between the reference radar and the radar to be registered based on the strong processed signals corresponding to the reference radar and the radar to be registered, and converting the range-Doppler registration problem into a cosine similarity maximization problem; solving the cosine similarity maximization problem based on the range-dimensional registration model and the Doppler registration model to obtain the strong target coherent synthesis parameters, and registering the radar to be registered based on the strong target coherent synthesis parameters; Each registered radar acquires the echo signal of the corresponding weak target, and sequentially performs carrier frequency removal, pulse compression, and coherent accumulation processing on the echo signal of the weak target to obtain the corresponding weak processed signal. All weak signals are vectorized, and the fused signal is obtained by weighted fusion based on all vectorization results. For each registered radar, the discrete phase compensation process includes: calculating the maximum acceptable phase difference of the registered radar based on the fused signal; calculating the signal-to-noise ratio (SNR) gain based on the fused signal corresponding to the registered radar; calculating the acceptable SNR gain based on the fused signal corresponding to the maximum acceptable phase difference; determining whether the SNR gain exceeds the acceptable SNR gain; if it does, outputting the fused signal; if it does not, compensating the corresponding weak processing signal based on the maximum acceptable phase difference of the registered radar; and returning to the step of vectorizing all weak processing signals until the maximum number of cyclic compensations is met.

2. The distributed network radar joint coherent processing method for strong and weak targets according to claim 1, characterized in that, The strong target coherent synthesis parameters include the angle rotation matrix and translation vector; The range-dimensional registration model constructed based on the strong target coherent synthesis parameters is expressed as follows: ； in, This represents the range dimension index of the strong target signal in the radar to be registered. This represents the range resolution unit of the radar to be registered. , These represent the elevation and azimuth angles of the strong target relative to the radar to be registered, respectively. Indicates the range resolution unit of the reference radar. , These represent the elevation and azimuth angles of the strong target relative to the reference radar, respectively. This represents the range dimension index of a strong target signal in the reference radar. Represents the angle rotation matrix. Represents the translation vector; The Doppler registration model constructed based on the strong target coherent synthesis parameters is expressed by the following formula: ； in, The Doppler index represents the strong target signal of the radar to be registered. The Doppler index representing the strong target signal of the reference radar. , , These represent the radial velocities of the strong target relative to the reference radar and the radar to be registered, respectively. and These represent the velocity resolution units of the reference radar and the radar to be registered, respectively.

3. The distributed network radar joint coherent processing method for strong and weak targets according to claim 1, characterized in that, The range-Doppler registration problem between the reference radar and the radar to be registered, constructed based on the strong processed signals corresponding to the reference radar and the radar to be registered, is expressed by the formula: ； in, This indicates the strong processing signal corresponding to the reference radar. This indicates the strong processing signal corresponding to the radar to be registered. This represents the ratio of the amplitude values ​​of the echo signals received by different radars. This indicates the error caused by noise.

4. The distributed network radar joint coherent processing method for strong and weak targets according to claim 3, characterized in that, The registration problem in the distance dimension-Doppler dimension is transformed into a cosine similarity maximization problem, expressed by the formula: ； in, Indicates parameters related to coherent synthesis The function, , , , The angles by which the rotation matrix rotates about the X-axis, Y-axis, and Z-axis are represented. , , The distances of the translation vector along the X, Y, and Z axes are represented. Represents the cosine similarity function. , This indicates the operation of finding the 2-norm.

5. The distributed network radar joint coherent processing method for strong and weak targets according to claim 1, characterized in that, Solving the cosine similarity maximization problem yields the strongly objective coherent synthesis parameters, including: The strong objective coherent synthesis parameters are obtained by solving the cosine similarity maximization problem using a genetic algorithm.

6. The distributed network radar joint coherent processing method for strong and weak targets according to claim 1, characterized in that, All weakly processed signals are vectorized, and a weighted fusion process is performed based on all vectorization results to obtain the fused signal, including: All weak processing signals are vectorized and combined based on all vectorization results to form a new processing signal; Based on the new processed signal and fusion weights, construct a fusion signal optimization problem; The fusion signal optimization problem is transformed into a power signal optimization problem; Solve the power signal optimization problem to obtain the optimal fusion weights; The fusion calculation is performed based on the optimal fusion weights and the new processed signal, and the fusion calculation result is then vectorized and inversely processed to obtain the fused signal.

7. The distributed network radar joint coherent processing method for strong and weak targets according to claim 6, characterized in that, Based on the new processed signal and fusion weights, a fusion signal optimization problem is constructed, expressed by the following formula: ； in, Indicates the initial fusion signal. Indicates the fusion weight. Expressing the request conjugate, This indicates a new processing signal.

8. The distributed network radar joint coherent processing method for strong and weak targets according to claim 6, characterized in that, The fused signal optimization problem is transformed into a power signal optimization problem, expressed by the following formula: ； in, Indicates the first digit in the fused signal The fusion weights corresponding to each component Expressing the request conjugate, This represents the correlation matrix between all registered radars. Indicates the first digit in the fused signal The new processing signal corresponds to each component. Indicates the number of radars to be registered. This indicates the transpose operation.

9. The distributed network radar joint coherent processing method for strong and weak targets according to claim 6, characterized in that, The fusion calculation is performed based on the optimal fusion weights and the new processed signal, and the fusion calculation result is then vectorized and inversely processed to obtain the fused signal, expressed by the following formula: ； in, The first digit of the fused signal One portion, Indicates the first digit in the fused signal The optimal fusion weights for each component Expressing the request conjugate, This indicates a new processing signal. This indicates the inverse process of vectorization. This indicates the transpose operation.

10. The distributed network radar joint coherent processing method for strong and weak targets according to claim 1, characterized in that, The formula for calculating the maximum acceptable phase difference is: ； in, Indicates the maximum acceptable phase difference. This indicates an acceptable loss in synthesized gain. Calculated based on the fused signal, This indicates the number of radars to be registered.

Images