Segmented coherent accumulation detection method for high-speed maneuvering target based on sub-segment signal distance-azimuth coupling relationship

Abstract

The invention provides a high-speed maneuvering target segment coherent accumulation detection method based on a sub-segment signal distance-azimuth coupling relation. The method comprises the following steps: carrying out fast Fourier transform on echo signals after pulse compression in a fast time domain; performing segmentation processing on the echo signal after the fast Fourier transform in a slow time domain; searching a high-order motion parameter of the target, constructing a high-order phase correction function, and eliminating the influence of the high-order motion parameter; compensating Doppler ambiguity and inter-subsegment residual distance migration; and performing inverse fast Fourier transform on the corrected and compensated echo signal in a fast time domain to complete joint accumulation of a plurality of sub-segment signals so as to obtain a result after distance-Doppler domain coherent accumulation processing of the high-speed maneuvering target. The problems that an existing method is too high in operation complexity, serious in integral loss, limited in application range and the like when used for detecting the high-speed maneuvering target are solved, and the high-maneuvering target with complex arbitrary-order range migration and Doppler migration can be effectively detected.

Description

Technical Field

[0001] The present invention relates to the field of communications technology and, more specifically, to a method for detecting high-maneuvering targets using segmented coherent integration based on the range-azimuth coupling relationship between subsegments in radar signal processing. The present invention can be used in existing radars utilizing long-term coherent integration technology to achieve long-range detection of high-speed maneuvering targets in near-space. Background Art

[0002] With the rapid development of aerospace and remote sensing technologies, existing highly maneuverable targets (HMTs), such as jet fighters, hypersonic vehicles, near-space vehicles, and spacecraft, have made significant progress and are beginning to pose a significant threat to radar remote sensing and detection. Therefore, the detection of HMTs is a key function of modern radar systems, and the performance of radar systems related to HMT detection has a significant impact on radar applications such as surveillance, tracking, imaging, and identification. However, HMTs typically exhibit long range, high speed, high maneuverability, and low radar cross-section characteristics, resulting in a low signal-to-noise ratio (SNR) of target echoes. This low SNR makes HMTs difficult to detect. Furthermore, the high speed and high maneuverability of HMTs easily lead to range migration (RM), Doppler migration (DFM), and Doppler ambiguity, further complicating target detection. Consequently, the detection of HMTs has become a challenging topic and a problem for modern radar systems, attracting widespread global attention.

[0003] SNR is a key factor in improving the detection probability of HMTs. Multi-pulse integration is an important technology for improving the SNR of HMT signals. Multi-pulse integration methods can be divided into two types: incoherent integration and coherent integration algorithms. Radon transform and Hough transform are typical incoherent integration algorithms that concentrate the signal energy on the curved trajectory of the target. However, the incoherent integration method ignores the phase information of the target signal and only uses the amplitude of the echo to improve the SNR of the target, which can lead to severe integration loss under low SNR conditions. Therefore, in order to further enhance the SNR of the target signal and improve the detection probability of the target, coherent integration processing methods using the phase information of the target signal should be studied.

[0004] Traditional Doppler filter banks are a typical coherent integration method for moving target detection (MTD). However, during long integration times, the range and Doppler frequency variations associated with moving targets (HMTs) can exceed the corresponding range and Doppler resolution. In this situation, range and Doppler migration are prone to occur, significantly degrading the coherent integration performance associated with MTD methods. Therefore, the effects of range and Doppler migration should be effectively eliminated to achieve satisfactory coherent integration.

[0005] Several typical methods have been proposed to address range migration caused by constant velocity, such as the Radon Fourier transform (RFT) and its modifications, and the axis rotation (AR) transform. These methods employ a target trajectory search approach. To avoid this target trajectory search, the keystone transform (KT) and its modifications, methods based on range and velocity decoupling (RVDT), and methods based on the azimuth resampling technique (ART) have been developed. KT eliminates the coupling between range and azimuth by rescaling slow-time variables. However, the KT method suffers from Doppler ambiguity and high computational overhead due to interpolation. To avoid interpolation, RVDT-based methods employ matrix-complex multiplication and Fourier transform (FT) to remove range migration caused by constant velocity. The ART method combines a non-uniform fast Fourier transform with a sparse Fourier transform to achieve coherent integration. However, these methods can only remove range migration caused by constant velocity and are still affected by Doppler migration.

[0006] To address the Doppler migration problem, well-known Radon transform-based methods, such as the Radon-fractional Fourier transform (RFRFT) and its improved RFRFT method, have been introduced. However, these methods suffer from high computational costs and severe blind-speed sidelobes (BSSL) performance limitations. To reduce the computational burden, range-azimuth frequency-domain methods have been developed. However, these methods require the target signal to have a large time-bandwidth product because they apply the stationary phase principle to obtain the range-azimuth spectrum. To address this issue, improved KT-based methods have been proposed, such as the KT matched filter (KTMF) method, the exact echo signal-KT method, the generalized second-order KT method, and the second-order KT-modified time-scaling transform method. However, these methods are still affected by Doppler ambiguity and high-order range and Doppler migration (i.e., arbitrary-order range and Doppler migration). Therefore, to address Doppler ambiguity, the improved KTMF (MKTMF) algorithm, methods based on range-frequency cross-correlation (RFCC), and methods based on the one-dimensional scaled Fourier transform (SCFT) have been proposed. However, the MKTMF algorithm is limited by the complex Doppler ambiguity and high-order range and Doppler migration caused by the high-order motion of HMTs. Since these processing steps include nonlinear operations, the above method may not be applicable to the coherent integration of HMTs with complex motion. Since these processing steps include nonlinear operations, the above method may not be applicable to the coherent integration of HMTs with complex motion.

[0007] To address arbitrary-order range and Doppler migration, typical methods based on the generalized RFT (GRFT) have been proposed. However, these methods are computationally expensive and still suffer from the blind velocity sidelobe (BSSL) effect. To reduce computational cost, a subaperture-based KTMF method has been proposed. However, this method is limited by Doppler ambiguity. If used directly in the presence of Doppler spectrum ambiguity, the target trajectory will be split into several parts, resulting in severe integration loss. To address this issue, hybrid integration (HI)-based methods have been proposed, which take Doppler ambiguity into account. However, these methods utilize incoherent integration between different subapertures to focus on maneuvering targets, resulting in severe coherent integration loss. Subsequently, methods based on the adjacent cross-correlation function (ACCF) and the range-frequency inversion transform have been proposed due to their extremely low computational cost. However, their processing steps include nonlinear operations, which can lead to significant performance loss in low SNR conditions. The one-dimensional SCFT method cannot be used for low-speed targets.

[0008] In summary, the existing long-term coherent integration detection method has problems such as high computational complexity, serious integral loss, and limited application scope for the detection of high-speed maneuvering targets. Summary of the Invention

[0009] The purpose of the present invention is to address the shortcomings of the above-mentioned existing technologies and propose an efficient segmented coherent integration method. The method uses the range-azimuth coupling relationship between sub-segments to perform migration compensation and complete the coherent accumulation detection of high-speed maneuvering targets. This method takes into account the Doppler spectrum ambiguity, arbitrary-order range migration, and Doppler migration caused by the complex motion of high-speed maneuvering targets under low signal-to-noise ratio conditions.

[0010] The specific steps of the present invention include the following:

[0011] (1) Perform fast Fourier transform on the echo signal after pulse compression in the range domain to obtain the echo signal s r (f,t m ), t m Indicates slow time;

[0012] (2) The echo signal s after fast Fourier transformation r (f,t m ) is segmented in the slow time domain, and the entire coherent accumulation time T a Divided into N s sub-segments, each sub-segment has a length of T s =T a / N s ;

[0013] (3) Setting the high-order motion parameters to be searched The search scope, represents the ρth-order motion parameter of the first subsegment;

[0014] Construct a high-order phase correction function to eliminate the influence of high-order motion parameters in each sub-segment k=1,2,…,N s , t mk is the slow time variable of the k-segment;

[0015] According to the following distance-azimuth coupling relationship between sub-segments:

[0016]

[0017] Construct the residual migration compensation function H between sub-segments m f,f mk , where f mk represents the frequency variable corresponding to the slow time variable of the k-segment, μ′ 1,1 is the baseband radial velocity of the target, and λ is the signal wavelength;

[0018] (4) Multiple sampling within the search range Based on sampling Repeat the following steps to correct, compensate, and perform segmented coherent accumulation on the signal:

[0019] (4a) Multiply the range frequency domain-slow time domain high-order phase correction function of each sub-segment with the corresponding echo signal to obtain the corrected range frequency domain-slow time domain high-speed maneuvering target echo signal:

[0020]

[0021] (4b) Perform fast Fourier transform on the corrected signal in the slow time domain to obtain the corrected range-frequency-Doppler domain radar echo signal s′ k,s-m f,f mk ;

[0022] (4c) The inter-subsegment residual migration compensation function is used to compensate for Doppler ambiguity and inter-subsegment residual migration as follows:

[0023] s′ ks-mm f,f mk =s′ k,s-m f,f mk H m f,f mk ;

[0024] (4d) Perform inverse fast Fourier transform on the corrected and compensated echo signal in the range domain, and complete the joint accumulation of multiple sub-segment signals to obtain the result s′t,f after the range-Doppler domain coherent accumulation processing of the high-speed maneuvering target mk as follows:

[0025] s′ ks-mm t,f mk =IFFT[s′ ks-mm f,f mk ]

[0026]

[0027] (5) In all sampling-based The peak search is performed during the obtained coherent accumulation process. The parameters corresponding to the peak are the high-order motion parameters of the matching target. The peak also corresponds to the final coherent accumulation detection result, which is:

[0028]

[0029] Furthermore, the inter-subsegment residual migration compensation function H m f,f mk The build is as follows:

[0030]

[0031] Where PRF is the pulse repetition frequency of the radar and M is the Doppler ambiguity number.

[0032] Furthermore, the high-order phase correction function The build is as follows:

[0033]

[0034] Furthermore, the highest order of motion parameters ρ is taken to be equal to 3.

[0035] The principles for achieving the objectives of the present invention are: First, a segmented operation that facilitates parallel processing is proposed to remove range and Doppler migration within a subsegment. Second, by analyzing the characteristics of high-order inter-subsegment range and Doppler migration, a high-order phase correction function is proposed to effectively eliminate high-order inter-subsegment range and Doppler migration. Finally, by leveraging the range-azimuth coupling relationship between subsegments, a subsegment residual migration compensation function (NRMCF) is proposed to achieve coherent target accumulation. The proposed NRMCF corrects the inter-subsegment residual range migration caused by radial velocity using only one matrix multiplication step, avoiding complex interpolation operations.

[0036] Compared with the prior art, the present invention has the following specific advantages:

[0037] First, the proposed method can accurately and simultaneously compensate for the complex range and Doppler migration caused by the complex arbitrary-order motion of high-speed maneuvering targets, and achieve coherent integration within and between subsegments.

[0038] Second, the proposed method is robust to complex Doppler ambiguity caused by arbitrary-order Doppler migration of high-speed maneuvering targets and solves the blind speed sidelobe effect.

[0039] Third, the proposed method is a linear process and can therefore be used in low SNR environments and is not disturbed by cross terms in multi-target situations.

[0040] Fourth, compared with the optimal GRFT method, the proposed algorithm has lower computational cost while maintaining similar cumulative performance.

[0041] Fifth, the proposed method can be implemented through parallel processing at the same time, which further reduces the processing time and helps to improve the real-time processing capability. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] Figure 1 Schematic diagram of segmented operation in an embodiment of the present invention;

[0043] Figure 2 1 is a schematic diagram of simulation results of sub-segment segmentation obtained in simulation experiment 1 of an embodiment of the present invention;

[0044] Figure 3 1 is a schematic diagram of compensation for inter-subsegment range migration and Doppler migration obtained in simulation experiment 1 according to an embodiment of the present invention;

[0045] Figure 4 1 is a schematic diagram of the coherent accumulation result obtained in the simulation experiment 1 of the embodiment of the present invention;

[0046] Figure 5 2 is a schematic diagram of target detection results obtained in simulation experiment 2 of an embodiment of the present invention when Doppler center ambiguity exists.

[0047] Figure 6 3 is a diagram of the detection results of the method of the present invention under Doppler center ambiguity obtained in simulation experiment 3 of an embodiment of the present invention;

[0048] Figure 7 Schematic diagram of multi-target detection results obtained in simulation experiment 4 of an embodiment of the present invention;

[0049] Figure 8 1 is a schematic diagram of the detection results under low signal-to-noise ratio obtained in simulation experiment 5 of an embodiment of the present invention;

[0050] Figure 9 This is a comparison chart of the detection results of the method of the present invention and other methods obtained in simulation experiment 5 of an embodiment of the present invention;

[0051] Figure 10 This is a comparison chart of the success detection probability results of the method of the present invention and other methods obtained in simulation experiment 6 of an embodiment of the present invention;

[0052] Figure 11 2 is a schematic diagram of simulation results obtained by the method of the present invention using experimental data scenario A in simulation experiment 7 of an embodiment of the present invention;

[0053] Figure 12 2 is a schematic diagram of simulation results obtained by the method of the present invention using experimental data scenario B in simulation experiment 7 of an embodiment of the present invention;

[0054] Figure 13 1 is a schematic diagram of target detection results obtained by the method of the present invention under airborne synthetic aperture radar data in simulation experiment 8 of an embodiment of the present invention;

[0055] Figure 14 1 is a schematic diagram of target detection results obtained by the method of the present invention under Ka-band radar data in simulation experiment 9 of an embodiment of the present invention. DETAILED DESCRIPTION

[0056] The present invention will be described in further detail below with reference to the accompanying drawings.

[0057] Step 1: The radar transmits a linear frequency modulation signal and receives the corresponding radar echo signal.

[0058] The transmitted linear frequency modulation signal expression is as follows:

[0059]

[0060] Among them, rect(·) represents the window function, that is, T p is the pulse width of the transmitted signal, γ represents the frequency modulation rate, f c is the carrier frequency and t is the fast time.

[0061] The target baseband echo signal received by the radar is as follows:

[0062]

[0063] Where, σ is the backscatter coefficient of the target, t m is the slow time, c is the speed of light, λ is the signal wavelength, R(t m ) is the instantaneous slant range between the radar and the target during the accumulation process.

[0064] The uniform instantaneous range between the radar and a high-speed maneuvering target can be given by the following formula:

[0065]

[0066] Where R0 is the initial tilt range, i = 1, 2, ..., ρ is the order of target motion, μ irepresents the i-th order of the target radial motion parameters, namely radial velocity, acceleration and jerk...

[0067] Step 2: Pulse compression is performed on the echo signal of the high-speed maneuvering target at the distance. The expression after pulse compression is as follows, where: B represents the bandwidth of the radar transmission signal;

[0068] Step 3: Perform a fast Fourier transform on the demodulated and pulse compressed high-speed maneuvering target echo signal to obtain the radar echo signal in the range-frequency domain-slow-time domain. The expression is as follows, where f represents the range-frequency variable:

[0069]

[0070] Step 4: segment the echo signal after fast Fourier transform in the slow time domain and divide the whole coherent accumulation time T a Divided into N s sub-segments, each sub-segment has a length of T s =T a / N s , the operation diagram is as follows Figure 1 As shown:

[0071] After applying the segmentation operation, the expressions of the k-th sub-segment signal in the range frequency and azimuth slow time domains are as follows:

[0072]

[0073] k=1,2,…,N s , t mk is the slow time variable obtained after segmentation operation, R k t mk is the instantaneous distance of the kth subsegment.

[0074] According to R k t mk The definition of , after segmentation, is as follows:

[0075]

[0076] where R k,0 represents the initial radial range of the kth sub-segment signal, μ k,1 represents the initial radial velocity of the kth sub-segment signal.

[0077] Step 5: Construct a high-order phase correction function to eliminate the influence of high-order motion parameters.

[0078] Assuming that there is Doppler center ambiguity, taking the first sub-segment as an example, the initial radial velocity of a high-speed moving target with Doppler center ambiguity can be rewritten as follows:

[0079]

[0080] Among them, μ′ 1,1 ∈[-λPRF / 4,λPRF / 4] is the baseband radial velocity, M is the Doppler ambiguity number, and PRF represents the pulse repetition frequency of the radar.

[0081] The difference in initial radial distance and velocity between each sub-segment is the key factor affecting the cross-sub-segment coherent accumulation process. According to the instantaneous slant range R(t m ), the initial radial distance R of the kth sub-segment signal k，0 and speed μ k,1 The definition of is as follows:

[0082]

[0083] Among them, R 1，0 is the initial distance, i=1,2,…, is the order of target motion, μ k.i represents the i-th order of the target radial motion parameters of the k-th sub-segment signal, namely radial velocity, acceleration and jerk…;

[0084] According to the expressions of the initial radial distance and velocity of the sub-segment signal, the simplified echo signal expression in the range-frequency-slow time domain can be obtained as follows:

[0085]

[0086] The inter-segment phase changes caused by the high-order motion parameters in the above formula will affect the coherent integration processing of the target signal. Therefore, the range frequency domain-slow time domain high-order phase correction function is constructed according to the following formula:

[0087]

[0088] in, are the high-order motion parameters to be searched;

[0089] Multiply the range frequency domain-slow time domain high-order phase correction function with the segmented echo signal to obtain the corrected range frequency domain-slow time domain high-speed maneuvering target echo signal.

[0090]

[0091] because and The simplified expression is as follows:

[0092]

[0093] If the search parameters When matching the true parameters of the target, the variations caused by high-order motion parameters are effectively eliminated.

[0094] Step 6: Search for high-order motion parameters of the target to compensate for Doppler ambiguity and residual migration between sub-segments.

[0095] Perform fast Fourier transform on the corrected echo signal in the slow time domain to obtain the corrected range-frequency-Doppler domain radar echo signal, which is expressed as follows:

[0096]

[0097] The position of the target peak in the azimuth Doppler domain can be expressed as follows:

[0098]

[0099] k=1,2,…,N s , where f m,k Represents the frequency variable corresponding to the slow time variable after segmentation, μ′ 1,1 is the baseband radial velocity of the target;

[0100] Based on the range-azimuth coupling characteristics of the target signal in the above equation, the range position between target segments and the last exponential term of the corrected range-frequency-Doppler domain radar echo signal both change with the azimuth Doppler frequency. The inter-segment residual migration compensation function in the range-Doppler domain is constructed as follows:

[0101]

[0102] The inter-segment residual migration compensation function is multiplied by the corrected echo signal to obtain the compensated range-frequency-Doppler domain high-speed maneuvering target echo signal.

[0103] s′ ks-mm f,f mk =s′ k,s-m f,f mk H m f,f mk

[0104] Step 7: Perform inverse fast Fourier transform on the corrected and compensated echo signal in terms of range, complete the joint accumulation of multiple sub-segment signals, and obtain the result of range-Doppler domain coherent accumulation processing of the high-speed maneuvering target.

[0105] The expression of the echo signal after correction and compensation after inverse fast Fourier transform at distance is as follows:

[0106]

[0107] Perform joint accumulation of sub-segment signals and obtain the coherently integrated range-Doppler domain high-speed maneuvering target echo signal expression as follows:

[0108]

[0109] By searching for high-order motion parameters within the search range Repeat steps 5-7 to obtain the coherent accumulation results under different parameters. When the searched parameters match the actual parameters, the peak value of the coherent accumulation result will be obtained. The parameters corresponding to the peak value are is the high-order motion parameter of the target, and the peak value corresponds to the final coherent accumulation detection result:

[0110]

[0111] The present invention is further described below in conjunction with simulation experiments.

[0112] 1. Simulation conditions:

[0113] Assume that the received echo does not contain noise. The motion parameters of the high-speed maneuvering target are set as: initial velocity u1 = 50 m / s, acceleration u2 = 20 m / s 2 , acceleration u3=10m / s 3 The radar system parameters are set as: carrier frequency f c =1.2GHz, transmission signal bandwidth B = 20MHz, sampling frequency f s =40MHz, pulse repetition frequency PRF=1000Hz, number of sub-segments N s =34.

[0114] 2. Simulation content and result analysis:

[0115] There are nine simulation experiments of the present invention.

[0116] Simulation experiment 1:

[0117] The simulation experiment 1 of the present invention is to detect the high maneuverability target under the above simulation conditions using the high maneuverability target segmented coherent accumulation detection method based on the distance-azimuth coupling relationship between sub-segments of the present invention. The results are as follows: Figure 2 、 Figure 3 and Figure 4 shown.

[0118] Figure 2 This is the simulation result of sub-segment segmentation. Figure 2 (a) is the result of pulse compression of the signal echo received by the radar. Figure 2 (b) is the result of the azimuth Fourier transform of the signal echo. Figure 2 (a) and Figure 2As shown in (b), the target trajectory is distributed across multiple range and Doppler frequency bins. Consequently, the target's radial velocity, acceleration, and jerk induce significant range and Doppler migration, causing the target signal to be severely defocused. The segmentation results for subsegments 1, 10, and 20 are used as examples. Figure 2 (c), (d), and (e) are the corresponding segmentation results in the range-slow time domain for subsegments 1, 10, and 20. It can be seen that after the segmentation operation, the intra-subsegment range migration of each subsegment is removed. Figure 2 (f), (g), and (h) are the results for subsegments 10 and 20 in the range-Doppler domain. It can be seen that after segmentation, the Fourier transform operation can focus the signal of each subsegment. However, the range and azimuth Doppler bins of the target peak shift significantly between subsegments, resulting in significant inter-subsegment range and Doppler migration. The target energy is distributed across the range and azimuth Doppler bins of different subsegments.

[0119] Figure 3 is the result of compensation for range migration and Doppler migration between sub-segments. Figure 3 (a) shows the acceleration and jerk search results. The search range and search interval vary depending on the specific situation. In this simulation, the acceleration search range is approximately -200% to +200% of the target parameter settings, and the jerk search range is approximately -200% to +200% of the target parameter settings. The search interval is one unit. Based on the target's peak position, the target's radial acceleration and jerk can be obtained. Figure 3 (b), (c), and (d) show the results of applying the high-order motion phase correction function. The resulting radial acceleration and jerk eliminate the inter-segment range migration and Doppler frequency variation caused by the target's high-order motion. However, despite successfully eliminating the inter-segment range migration and Doppler frequency variation caused by high-order motion, the influence of inter-segment range migration caused by radial velocity still remains. Figure 7 Figures (m), (n), and (o) are the results of applying the inter-subsegment residual migration compensation function. It can be seen that after correction using the inter-subsegment residual migration compensation function, the inter-subsegment distance migration caused by radial velocity is effectively compensated, and the target energy in different subsegments is focused on the same peak position.

[0120] Figure 4 is the result of coherent accumulation. Figure 4 (a) is the result of coherent accumulation across subsegments. Figure 4 (b) shows the cross-subsegment coherent integration result without compensation for inter-subsegment range and Doppler migration. The comparison shows that the target energy is still not fully focused due to the presence of inter-subsegment range and Doppler migration. Figure 4(c) shows the cross-subsegment coherent integration result without compensating for inter-subsegment range migration caused by radial velocity. It can be seen that although the inter-subsegment range migration and Doppler migration caused by high-order motion have been removed, the focusing result is still not ideal due to the influence of the inter-subsegment range migration caused by radial velocity.

[0121] Simulation experiment 2:

[0122] Simulation Experiment 2 of the present invention verifies the effectiveness of the present invention in the presence of Doppler center ambiguity. The signal-to-noise ratio (after pulse compression) is set to 8 dB. The basic radar parameters of the simulation are the same as those in Experiment 1. In this simulation, a Doppler center ambiguity target is set, and the target parameters are set to initial velocity u1 = 225 m / s and acceleration u2 = 15 m / s. 2 , acceleration u3=5m / s 3 The experimental results are as follows. Figure 5 shown.

[0123] Figure 5 is the target detection result when Doppler center ambiguity exists. Figure 5 (a) is the result of target pulse compression, Figure 5 The curved trajectory corresponding to the target in (a) has obvious range migration phenomenon. Figure 5 (b) is the range-Doppler domain after pulse compression, from Figure 5 As can be seen from (a) and 5(b), the target suffers from a strong defocusing effect in the range-Doppler domain, and the Doppler spectrum of the target is completely located within one PRF band. Figure 5 (c) is the matching result of target acceleration and jerk. Figure 5 (d) is the Doppler ambiguity number obtained using the proposed method. Figure 5 (e) is the final coherent accumulation result. It can be seen that the residual inter-subsegment range migration is effectively eliminated by using the obtained Doppler ambiguity number. Figure 5 (f) is the focusing result of the proposed method without Doppler ambiguity compensation. Figure 5 (e) Comparison shows that after Doppler ambiguity compensation, the proposed method can obtain a good focusing result. Therefore, the proposed method can effectively solve the Doppler center ambiguity effect.

[0124] Simulation experiment 3:

[0125] The simulation experiment 3 of the present invention is to evaluate the performance of the proposed algorithm in the presence of Doppler spectrum ambiguity. The basic parameters and signal-to-noise ratio of the radar are the same as those in experiment 1. The target parameters are set as initial velocity u1 = 350 m / s, acceleration u2 = 100 m / s 2 , acceleration u3=80m / s 3 The experimental results are as follows. Figure 6 shown.

[0126] Figure 6 is the detection result of the proposed algorithm when Doppler spectrum ambiguity exists. Figure 6 (a) and (b) show the results in the range-slow-time domain and range-Doppler domain after pulse compression. As shown in Figures 6(a) and (b), the target exhibits significant range and Doppler migration, and the target's Doppler spectrum is split into several pulse repetition frequency (PRF) bands. This simulation uses a target with a split Doppler spectrum and applies a segmentation operation to the target. Figure 6 (c) is the matching result of acceleration and jerk. Figure 6 (d) is the matching Doppler ambiguity number. The range migration and Doppler migration between sub-segments are eliminated by obtaining the acceleration, jerk and Doppler ambiguity number parameters. Figure 6 (e) is the final good focusing result. Figure 6 (f) shows the range migration compensation result of the SKT method. It can be seen that because the SKT method directly uses KT to remove range migration, the target trajectory is split into multiple parts, resulting in severe coherence integration loss. Therefore, compared with the SKT method, the proposed method is more robust in dealing with Doppler spectrum ambiguity.

[0127] Simulation experiment 4:

[0128] The purpose of the simulation experiment 4 of the present invention is to verify the effectiveness of the proposed method in the case of multiple targets. The basic parameters and signal-to-noise ratio of the radar are the same as those in experiment 1. Three maneuvering targets with jerk are considered in the experiment, namely target 1 (Tar 1), target 2 (Tar 2) and target 3 (Tar 3). The simulation target parameters are shown in Table 1, and the simulation results are shown in Table 1. Figure 7 shown.

[0129] Table 1 Simulation parameters of the three targets

[0130] Velocity(m / s) <![CDATA[Acceleration(m / s 2 )]]> <![CDATA[Jerk(m / s 3 )]]> Tar 1 225 20 5 Tar 2 225 15 10 Tar 3 -200 10 7

[0131] Figure 7 It is the target detection result under multiple targets. Figure 7 (a) is the pulse compression result of target 1, target 2 and target 3. It can be seen that the curve trajectories of these targets show obvious range migration. Figure 7 (b) shows the range-Doppler domain results after pulse compression, and the target trajectory also shows obvious Doppler migration. Figure 7 (c), (d), (e), and (f) are the search results of Doppler ambiguity, acceleration, and jerk for targets 1, 2, and 3. A high-order phase correction function is applied using the obtained motion parameters. Figure 7(g), (h), and (i) are the results after applying the NRMCF operation to eliminate the residual migration. It can be seen that the proposed method eliminates the range migration and Doppler migration between sub-segments, and finally achieves good focusing results for targets 1, 2, and 3.

[0132] Simulation experiment 5:

[0133] The simulation experiment 5 of the present invention verifies the detection performance of the proposed method at low signal-to-noise ratio. The target initial velocity u1 in the simulation is 350m / s, and the acceleration u2 is 150m / s. 2 , acceleration u3=60m / s 3 The basic parameters of the radar are the same as those in Experiment 1. In order to verify the performance of this method, this simulation experiment considers the low signal-to-noise ratio situation and sets the signal-to-noise ratio after pulse compression to -3dB. The experimental results are shown in Figure 2. Figure 8 , Figure 9 shown.

[0134] Figure 8 It is the detection result of the target under low signal-to-noise ratio. Figure 8 (a) is the target pulse compression result. It can be seen that due to the interference of background noise, the target trajectory is submerged in the background. Figure 8 (b) is the result of Fourier transform in the slow time domain. It can be seen that due to the influence of range migration and Doppler migration, the energy of the target is still submerged by the background. Figure 8 (c) and (d) are Figure 8 The results in (a) and (b) remove the background noise. Figure 8 In (c) and (d), the target's range and Doppler migration are clearly visible, and the target energy is out of focus in the range-Doppler domain. In addition, the target's Doppler spectrum spans multiple PRF bands, showing significant Doppler center blurring. Figure 8 (e) is the matching result of acceleration and jerk. Figure 8 (f) is the matching Doppler ambiguity number. By obtaining the acceleration, jerk and Doppler ambiguity number, the proposed method effectively removes the range migration and Doppler migration between each sub-segment, and finally obtains a well-focused result.

[0135] Figure 9 This is the comparison result between the proposed method and the existing methods. Figure 9 (a) is the final focusing result of the proposed method. Figure 9 (b), (c), (d) and (e) are the processing results of standard RFT, SRFT, MKTMF and SKT methods. Figure 9(b), (c) and (d) show the out-of-focus results of the standard RFT, SRFT and MKTMF methods, as these methods ignore high-order range and Doppler migration. Figure 9 The SKT method shown in (e) cannot effectively solve the problem of Doppler spectrum splitting, so the performance of the SKT method is degraded.

[0136] Simulation experiment 6:

[0137] Simulation Experiment 6 of the present invention further verifies the detection performance of the proposed method through Monte Carlo experiments. The simulation parameters of the basic radar and target are consistent with those in Experiment 1. In the experiment, Gaussian noise is added to the target signal. For comparison, MTD, standard RFT, SRFT, GRFT, MKTMF, and SKT methods are used. The detection performance is measured by using a CFAR detector with the above six algorithms and the proposed method. The false alarm rate in the experiment is set to 10 -4 300 Monte Carlo experiments were performed for each signal-to-noise ratio (after pulse compression). The experimental results are shown in Figure 2. Figure 10 shown.

[0138] Figure 10 Figure 2 shows the relationship between target detection probability and SNR for six algorithms and the proposed method. It can be seen that the proposed method outperforms the MTD, standard RFT, SRFT, MKTMF, and SKT methods. The MTD, standard RFT, SRFT, and MKTMF algorithms ignore high-order range and Doppler migration, while the SKT method is affected by Doppler spectrum splitting. Compared to the optimal GRFT method, the proposed method achieves similar coherent integration performance, but with significantly lower computational cost.

[0139] Simulation experiment 7:

[0140] The simulation experiment 7 of the present invention is to verify the simulation results of the proposed method under real data. The experiment quoted a section of RADARSAT-1 Vancouver scene SAR data recorded by the C-band space radar, and the data used included the actual data scene A and the actual data scene B. The main radar parameters of the RADARSAT-1 Vancouver scene SAR data are as follows: the radar carrier frequency is 5.3GHz, the range bandwidth is 30.116MHz, and the pulse repetition frequency is 1256.98Hz. The experimental results are as follows Figure 11 , Figure 12 shown.

[0141] Figure 11 It is the target detection result using experimental data scene A. Figure 11 (a) is the selected actual data scene A. A single target is marked in the figure, denoted as Tar A. Figure 11(b) is the pulse compression result of Tar A. It can be seen that Tar A exhibits a significant range migration effect, causing the target energy to be defocused over distance. Figure 11 (c) is the pulse compression result of Tar A in the range-Doppler domain. Due to the Doppler migration effect, the target energy cannot be focused in the Doppler domain. Figure 11 (d) is the matching result of acceleration and Doppler fuzzy number. By using the obtained acceleration and Doppler fuzzy number, the range migration and Doppler migration between sub-segments are effectively removed. Figure 11 (e) is the focusing result. It can be seen that the proposed method achieves good focusing. Figure 11 (f) shows the processing result of the GRFT method. A comparison shows that the blind velocity sidelobes are clearly observed in the figure. Therefore, compared with the GRFT method, the proposed method can effectively avoid the influence of the blind velocity sidelobes.

[0142] Figure 12 It is the target detection result using experimental data scene B. Figure 12 (a) is the scene of actual data B. These data contain two targets, denoted as Tars B and C respectively. Figure 12 (b) is the pulse compression result in the range-Doppler domain for Tars B and C. It can be seen that Tars B and C exhibit obvious range migration and Doppler migration in the range-Doppler domain. Figure 12 (c) is the acceleration and Doppler fuzzy number matching result of Tars B and C. It can be seen that the acceleration and Doppler fuzzy number of Tar B and Tar C are the same. Figure 12 (d) shows the final focusing result for the target. As can be seen, by using the obtained acceleration and Doppler ambiguity to remove the range and Doppler migration between subsegments, good focusing results are achieved for Tars B and C. Therefore, the processing results for these two targets demonstrate the effectiveness of the proposed method in the multi-target scenario.

[0143] Simulation experiment 8:

[0144] The simulation experiment 8 of the present invention verifies the effectiveness of the proposed method on a section of measured airborne synthetic aperture radar data collected by an X-band three-channel airborne radar. The data acquisition radar operates in the side-view strip imaging mode and the flight speed is 120m / s. The basic parameters of the airborne radar are as follows: the radar carrier frequency is 8.85GHz, the PRF is 1000Hz, and the radar range bandwidth is 40MHz. The experimental results are shown in Figure 8. Figure 13 shown.

[0145] Figure 13 It is the target detection result under airborne synthetic aperture radar data. Figure 13(a) is the result of pulse compression of selected airborne SAR data in the range-Doppler domain. Figure 13 (b) is the result after clutter suppression. Due to the strong ground clutter effect, the moving target is submerged in the background. After using the expansion factor method, clutter suppression is successfully achieved. The target is marked as Tar D and Figure 13 (b) is marked. Figure 13 (c) and Figure 13 (d) shows the target trajectory of Tar D in the range-slow-time domain and the range-Doppler domain. Due to the influence of range migration and Doppler migration, the target trajectory of Tar D shows a typical defocusing phenomenon in the range-Doppler dimension. Figure 13 (e) is the matching result of acceleration and Doppler fuzzy number. The obtained acceleration and Doppler fuzzy numbers effectively remove the range migration and Doppler migration between sub-segments. Figure 13 (f) is the focusing result of Tar D obtained by the proposed method. It can be seen that a good focusing result is obtained.

[0146] Simulation experiment 9:

[0147] The simulation experiment 9 of the present invention verifies the effectiveness of the proposed method on a section of measured Ka-band radar data. These measured data were collected by setting up a Ka-band radar on a 90-meter-high tower, and the observation target was a small fixed-wing UAV flying at a constant radial velocity. The basic parameters of the Ka-band radar are as follows: the radar carrier frequency is 35 GHz, the PRF is 32 kHz, and the radar's range resolution is 1.875 meters. The experimental results are shown in Figure 9. Figure 14 shown.

[0148] Figure 14 It is the target detection result under Ka-band radar data. Figure 14 (a) and Figure 14 (b) is the pulse compression result of the target. The selected target shows obvious defocusing due to the range migration phenomenon. Figure 14 (c) is the result after processing using the proposed method. It can be seen that a good focusing result is obtained after using the proposed method.

[0149] In summary, the segmented coherent integration detection method for high-maneuvering targets based on the range-azimuth coupling relationship between subsegments provided by the present invention primarily addresses the problems of high computational complexity, severe integration loss, and limited application scope in existing long-term coherent integration detection methods for high-speed maneuvering targets. Its implementation steps include: transmitting a linear frequency modulation signal by a radar and receiving a corresponding radar echo signal; demodulating and pulse compressing the received high-speed maneuvering target echo signal in terms of range; performing a fast Fourier transform (FFT) on the pulse-compressed echo signal in terms of range; segmenting the FFT echo signal in the slow-time domain; searching for the target's high-order motion parameters and constructing a high-order phase correction function to eliminate the influence of the high-order motion parameters; compensating for Doppler ambiguity and residual range migration between subsegments; performing an inverse fast Fourier transform (IFFT) on the corrected and compensated echo signal in terms of range, completing the joint accumulation of multiple subsegment signals, and obtaining the result of range-Doppler domain coherent integration processing for high-speed maneuvering targets. The present invention is a coherent integration method for detecting high-maneuvering targets with complex arbitrary-order range and Doppler migration.

Claims

1. A segmented coherent integration detection method for high-speed maneuvering targets based on the sub-segment signal range-azimuth coupling relationship is characterized by: The following steps are included: (1) Perform fast Fourier transform on the echo signal after pulse compression in the range domain to obtain the echo signal s r (f,t m ), t m Indicates slow time; (2) The echo signal s after fast Fourier transformation r (f,t m ) is segmented in the slow time domain, and the entire coherent accumulation time T a Divided into N s sub-segments, each sub-segment has a length of T s =T a / N s ; (3) Setting the high-order motion parameters to be searched The search scope, represents the ρth-order motion parameter of the first subsegment; Construct a high-order phase correction function to eliminate the influence of high-order motion parameters in each sub-segment k=1,2,…,N s , t mk is the slow time variable of the k-segment; According to the following distance-azimuth coupling relationship between sub-segments: Construct the residual migration compensation function H between sub-segments m f,f mk , where f mk The frequency variable μ′ corresponding to the slow time variable of the k-segment 1,1 is the baseband radial velocity of the target, and λ is the signal wavelength; (4) Multiple sampling within the search range Based on sampling Repeat the following steps to correct, compensate, and perform segmented coherent accumulation on the signal: (4a) Multiply the range frequency domain-slow time domain high-order phase correction function of each sub-segment with the corresponding echo signal to obtain the corrected range frequency domain-slow time domain high-speed maneuvering target echo signal: (4b) Perform fast Fourier transform on the corrected signal in the slow time domain to obtain the corrected range-frequency-Doppler domain radar echo signal s′ k,s-m f,f mk ; (4c) The inter-subsegment residual migration compensation function is used to compensate for Doppler ambiguity and inter-subsegment residual migration as follows: s′ ks-mm f,f mk =s′ k,s-m f,f mk H m f,f mk ; (4d) Perform inverse fast Fourier transform on the corrected and compensated echo signal in the range domain, and complete the joint accumulation of multiple sub-segment signals to obtain the result s′t,f after the range-Doppler domain coherent accumulation processing of the high-speed maneuvering target mk as follows: s′ ks-mm t,f mk =IFFT[s′ ks-mm f,f mk ] (5) In all sampling-based The peak search is performed during the coherent accumulation process, and the parameters corresponding to the peak To match the high-order motion parameters of the target, the peak value corresponds to the final coherent accumulation detection result, which is:

2. The method according to claim 1, characterized in that The inter-subsegment residual migration compensation function H m f,f mk The build is as follows: Where PRF is the pulse repetition frequency of the radar and M is the Doppler ambiguity number.

3. The method according to claim 1, characterized in that The high-order phase correction function The build is as follows:

4. The method according to claim 1, wherein The highest order of motion parameters ρ is taken to be equal to 3.

Images