Low complexity MIMO-SAR echo separation method based on multi-sub-pulse separation and weighted synthesis

Abstract

The application discloses a low-complexity MIMO-SAR echo separation method based on multi-sub-pulse separation and weighted synthesis, which comprises the following steps: S1: designing a MIMO-SAR transmitting waveform based on a segmented phase encoding mechanism, and recording echo signals of multiple transmitters; S2: performing matched filtering processing on mixed signals in a receiver; S3: separating multiple sub-pulse echo signals in the elevation dimension by using a beam forming technology; S4: performing time shift on the separated sub-pulse signals, and ensuring that each sub-pulse signal of the same transmitter occupies the same time delay range; S5: inversing a decoding vector, namely a sub-pulse weighting coefficient, according to a phase encoding matrix in the step S1; and S6: performing weighted synthesis on the sub-pulse signals in the step S4, and realizing final MIMO-SAR echo separation. The method can reduce the number of interference components in echo separation, simplify the array configuration required by the beam forming while ensuring the separation performance, effectively reduce the complexity of the MIMO-SAR system, and has a wide application prospect in practical application.

Description

Technical Field

[0001] This invention relates to a low-complexity MIMO-SAR echo separation method based on multi-subpulse separation and weighted synthesis, belonging to the field of MIMO-SAR signal processing technology. Background Technology

[0002] Synthetic Aperture Radar (SAR) systems possess comprehensive global environmental monitoring capabilities in both military and civilian fields, attracting significant interest from researchers. However, with increasingly stringent technical requirements and expanding applications, current SAR systems need to possess multiple capabilities, such as: 1) high azimuth resolution and unambiguous imaging over a wide mapping band; 2) downward-looking 3D imaging; and 3) Ground Moving Target Indication (GMTI). To provide these imaging capabilities, SAR systems must utilize multiple receiving channels to increase their degrees of freedom. In practice, traditional Single-Input Multiple-Output (SIMO) SAR systems offer limited degrees of freedom (DOF), driving the development of next-generation Multiple-Input Multiple-Output (MIMO) SAR systems.

[0003] According to the deployment characteristics of the transmitting and receiving antennas, MIMO-SAR can be divided into two categories: 1) distributed MIMO-SAR (referring to bistatic / multistatic SAR) and 2) centralized MIMO-SAR. Considering the coherence of the data acquired by the radar system, this scheme mainly studies the centralized MIMO-SAR system. As we all know, MIMO-SAR needs to satisfy the orthogonality between each transmitted waveform. However, there is no waveform that is simultaneously at the same frequency and completely orthogonal. The cross-correlation energy (i.e., interference energy) caused by the non-orthogonality of the waveforms will seriously degrade the imaging performance of the SAR system

[12] . Therefore, it is crucial to develop advanced waveform schemes or cross-correlation energy suppression techniques for the realization of the MIMO concept.

[0004] In recent years, researchers both domestically and internationally have conducted extensive research on the aforementioned issues. Typically, MIMO-SAR echo separation can be achieved in the time, frequency, Doppler, and spatial domains. For example, the Space-Time Coding (STC) mechanism proposed by Kim and Wiesbeck. However, STC technology reduces the number of azimuth sampling samples or requires a high Pulse Repetition Frequency (PRF), thus affecting the imaging performance of SARGMTI or High-Resolution Wide-Swath (HRWS). Therefore, in 2014, Krieger proposed an orthogonal waveform beamforming scheme, namely the Short-Term Shift-Orthogonal (STSO) scheme. This scheme achieves MIMO-SAR echo separation through waveform orthogonality and elevation-dimensional digital beamforming (DBF). While the STSO waveform transmission mechanism can largely unlock the superior performance of MIMO-SAR, it significantly increases the complexity of the MIMO-SAR system. The current challenge lies in simplifying array configuration and saving system resources while achieving MIMO-SAR echo separation. Summary of the Invention

[0005] To address the aforementioned issues, this invention proposes a low-complexity MIMO-SAR echo separation method based on multi-subpulse separation and weighted synthesis. Compared to traditional beamforming echo separation schemes, this method effectively reduces the number of interference components that need to be suppressed, thereby saving elevation dimension system resources required for beamforming. This also greatly simplifies the array configuration of the MIMO-SAR system.

[0006] The low-complexity MIMO-SAR echo separation method based on multi-sub-pulse separation and weighted synthesis disclosed in this invention includes the following steps:

[0007] A low-complexity MIMO-SAR echo separation method based on multi-subpulse separation and weighted synthesis includes the following steps:

[0008] S1: Design the transmitted waveform and use a receiver to record the echo signals of multiple transmitted waveforms;

[0009] S2: Perform matched filtering on the echo signals of multiple transmitted waveforms in the receiver;

[0010] S3: Using beamforming technology to separate multiple sub-pulse echo signals in the pitch dimension;

[0011] S4: Time-shift the separated sub-pulse signals to ensure that each sub-pulse signal from the same transmitter occupies the same time delay range;

[0012] S5: Retrieve the solution encoding vector, i.e., the sub-pulse weighting coefficients, from the phase encoding matrix in step S1;

[0013] S6: Weighted synthesis of the sub-pulse signals in S4 to achieve final MIMO-SAR echo separation.

[0014] Preferably, in step S1, a segmented phase coding (SPC) waveform is selected as the transmitted waveform of a multi-input multi-output synthetic aperture radar (MIMO-SAR). Assuming the MIMO-SAR has M transmitters and N receivers, the segmented phase coding waveform s... m The general form of (t) is expressed as:

[0015]

[0016]

[0017] s0(t)=[s 1 (t),s 2 (t),…,s M (t)] T

[0018] In the formula, t represents fast time, and k represents the k-th transmitter. Indicates the encoded phase, s k (t)=s(t+(Mk)T s ) represents the sub-pulse signal, T s Indicates the sub-pulse width, (·) T Indicates the transpose operation;

[0019] The echo signal received by receiver Rxn after the M segmented phase-coded waveforms are backscattered by ground targets is represented as follows:

[0020]

[0021] In the formula, η represents slow time, and h n,m (t,η) represents the channel response. This indicates a fast-time convolution operation.

[0022] Preferably, in step S2,

[0023] The result after matching filter processing of the echo signal Recorded as:

[0024]

[0025] In the formula, F t (·) represents the fast-time Fourier transform. For fast-time inverse Fourier transform, f r k is the distance frequency. r For frequency modulation slope, This is the kth sub-pulse signal.

[0026] Preferably, in step S3,

[0027] As can be seen from step S1, the time delay interval between adjacent sub-pulse signals is T. s In a distributed scenario, the echo signal is further represented as:

[0028]

[0029] In the formula, t0, W s η and t′ represent the time delay center, time delay length, azimuth slow time, and time delay corresponding to each distance unit in the scene, respectively. Let be the echo signal of the k-th sub-pulse transmitted by the m-th transmitter and received by the n-th receiver, and

[0030] When W s <T s At that time, the sub-pulse signals do not overlap, so they can be directly separated in the fast time domain. When W s >T s At that time, the sub-pulse signal will exhibit aliasing; furthermore, when W s >(M-1)T s At this time, the M sub-pulse signals will have the same time delay range, and the echo signal within this time delay range is represented as:

[0031]

[0032] The sub-pulse signal is separated using beamforming technology. The separated sub-pulse signal is represented as follows:

[0033]

[0034] Preferably, in step S4,

[0035] According to step S3, the time delay range corresponding to the kth sub-pulse signal is denoted as:

[0036]

[0037] The sub-pulse signal is time-shifted in the distance-frequency domain, i.e.:

[0038]

[0039] In the formula, Indicates the separated sub-pulse signal After time shift, the sub-pulse signal, p c denoted by , where ⊙ represents the time shift corresponding to each sub-pulse signal, and ⊙ represents the Hadamard product operation.

[0040] Preferably, in step S5,

[0041] Pair pulse signal Decoding is performed to further suppress interference signals: the decoded vector is calculated as follows:

[0042]

[0043] In the formula, C, c M and I m These represent the encoding matrix, encoding vector, and unit vector, respectively.

[0044] Preferably, in step S6,

[0045] The calculated decoding vector is used to weight and synthesize the sub-pulse signals to suppress residual interference signals.

[0046]

[0047] Compared with existing technologies, the beneficial effects of the invented low-complexity MIMO-SAR echo separation method based on multi-subpulse separation and weighted synthesis are: the method can halve the number of interference components in the unit to be processed, thereby saving the system resources required for MIMO-SAR echo separation and reducing system complexity. Attached Figure Description

[0048] Figure 1 It is a low-complexity MIMO-SAR echo separation method based on multi-subpulse separation and weighted synthesis.

[0049] Figure 2 This is a comparison of the matched filtering results between the STSO mechanism and the proposed method;

[0050] Figure 3 (a) is the truth map of the scene illuminated by transmitter 1;

[0051] Figure 3 (b) is the truth map of the scene illuminated by transmitter 2;

[0052] Figure 4 (a) shows the imaging results of the mixed signals;

[0053] Figure 4 (b) is the imaging result after echo separation using the traditional STSO mechanism;

[0054] Figure 4 (c) is the imaging result after echo separation using the proposed method; Detailed Implementation

[0055] The following detailed description, with reference to the accompanying drawings, illustrates the low-complexity MIMO-SAR echo separation method based on multi-subpulse separation and weighted synthesis proposed in this invention.

[0056] Compared with traditional interference methods, the low-complexity MIMO-SAR echo separation method based on multi-pulse separation and weighted synthesis proposed in this invention mainly explores how to reduce the number of interference components to be suppressed and reduce the complexity of the MIMO-SAR system while ensuring echo separation performance. Its main steps are as follows:

[0057] S1: Design the transmission waveform based on the segmented phase coding mechanism and record the echo signals from multiple transmitters;

[0058] This scheme uses a segmented phase-coded waveform as the transmit waveform for MIMO-SAR. Generally, the specific form of this waveform can be expressed as:

[0059]

[0060]

[0061] s0(t)=[s(t+(M-1)T s ),s(t+(M-2)T s ),…,s(t)] T

[0062] Typically, s(t) can be any basic radar signal. Without loss of generality, this scheme uses a linear frequency modulated signal to design the MIMO-SAR transmit waveform. Furthermore, C = [c1, c2, ..., c M The symbol ] represents the encoding matrix. For M types of segmented phase-coded waveforms, the general form of the encoding matrix C can be written as:

[0063]

[0064] Taking point target P as an example, the echo signal emitted by Txm and received by Rxn can be expressed as:

[0065]

[0066] h n,m (t,η)=σ p δ(tt n,m )exp(-j2πf c t n,m )

[0067] In the formula, σ p , t n,m Let δ(t) represent the point target amplitude, point target time delay, and impulse response function, respectively. Considering that a receiver in a MIMO-SAR system receives the echoes of all transmitted signals, the echo signal received by Rxn can be written as:

[0068]

[0069] S2: Perform matched filtering on the mixed signal in the receiver;

[0070] Before performing DBF processing on the echo signal, matched filtering is required to accurately estimate the DOA information of the signal. Echo data r n The matched filtering result of (t,η) can be expressed as:

[0071]

[0072] Observing the above formula, it can be seen that the mixed signal It can be composed of multiple sub-pulse signals The sub-pulse signal will be separated in the following steps.

[0073] S3: Using beamforming technology to separate multiple sub-pulse echo signals in the pitch dimension;

[0074] Under far-field conditions, it can be assumed that the sub-pulse signals from different transmitters have the same time delay. To simplify the analysis, this step will analyze the echo signal of transmitter Txm. From step S1, it can be seen that the time delay interval between adjacent sub-pulses is T. s In a distributed scenario, the mixed signal can be further represented as:

[0075]

[0076] In the formula, t0 and W s Let represent the time delay center and time delay extent of the illumination scene, respectively. Observing the above formula, we can see that when W s <T s At this time, the sub-pulse signals do not overlap, therefore they can be directly separated in the fast time domain. When W... s >T s At this time, aliasing will occur in the sub-pulse signals. Furthermore, when W... s >(M-1)T s At this time, the M sub-pulse signals will have the same time delay range. The mixed signal within this range can be represented as:

[0077]

[0078] Observing the above formula, it can be seen that within the same processing unit, each sub-pulse signal is obtained by backscattering from the target at different locations. For example, considering the time delay t... c Processing unit at the location, sub-pulse signal It is due to time delay The backscattering of the point target at that location is obtained, i.e.

[0079]

[0080] In other words, each sub-pulse signal in this unit has different DOA information, so the sub-pulse signals can be separated using DBF technology. This scheme assumes that there are no terrain undulations in the MIMO-SAR imaging region, i.e., the height of the imaging region is known a priori (H=0). In this case, the DOA of each sub-pulse signal in the processing unit can be calculated as follows:

[0081]

[0082] In the formula, arcsin(·) represents the arcsine function, and R is the instantaneous slant range corresponding to the processing unit. Based on the angle information calculated by the above formula, the guidance vector matrix can be further constructed, i.e.

[0083] V(θ)=[v(θ1),v(θ2),…,v(θ M )]

[0084]

[0085] Based on the deterministic guidance vector matrix described above, the optimal weight vector for DBF processing can be calculated as follows:

[0086] W = [ω1, ω2, ..., ω M ]=V(θ)(V H (θ)V(θ)) -1

[0087] In the formula, (·) H This represents the conjugate transpose operation. It involves transposing the weight vector ω... k Processing the pitch-dimensional spatial snapshots can enable the processing of sub-pulse signals. The separation of the sub-pulse signals. The separated sub-pulse signals can be represented as:

[0088]

[0089] S4: Time-shift the separated sub-pulse signals to ensure that each sub-pulse signal from the same transmitter occupies the same time delay range;

[0090] According to step S3, the time delay range corresponding to the kth sub-pulse signal can be denoted as:

[0091]

[0092] As shown in the above equation, different sub-pulse signals occupy different time delay ranges. To perform subsequent weighted synthesis operations, the sub-pulse signals need to be time-shifted. Generally, time-shifting can be performed in the distance-frequency domain, i.e.

[0093]

[0094] S5: Retrieve the solution encoding vector, i.e., the sub-pulse weighting coefficients, from the phase encoding matrix in step S1.

[0095] The separated sub-pulse signal still contains echo signals from other transmitters, therefore it needs to be decoded to further suppress interference signals. The decoding vector is defined as follows:

[0096]

[0097] To completely suppress residual interference energy, the weighted synthesis process of the sub-pulse signals can be expressed as:

[0098]

[0099] In the formula, h n (t,η)=[h n,1 (t,η),h n,2 (t,η),…,h n,M (t,η)] T Considering The term is independent of fast time t, and the above formula can be further written as:

[0100]

[0101] To separate the channel response h n,m (t,η), only needs to satisfy

[0102]

[0103] Therefore, the decoded vector can be calculated as:

[0104]

[0105] S6: Weighted synthesis of the sub-pulse signals in S4 to achieve final MIMO-SAR echo separation.

[0106] The sub-pulse signals are weighted and synthesized using the calculated decoding vector to suppress residual interference signals. This process can be described as follows:

[0107]

[0108] The overall technical process is as follows Figure 1 As shown.

[0109] Next, the performance of the proposed method will be analyzed through simulation experiments on point targets and area targets. The specific simulation parameters are shown in Table 1. The selected MIMO-SAR system has 2 transmitters and 2 receivers, i.e., M=2, N=2. Figure 2 This represents the matched filtering results of the mixed signal obtained by the STSO mechanism and the proposed method in the point target experiment. Observation Figure 2 It is known that the STSO mechanism contains two interference signal components, while the proposed method contains only one. The proposed method can halve the number of interference components in the echo.

[0110] Table 1. Main parameters involved in the simulation data.

[0111]

[0112]

[0113] Furthermore, the performance of the proposed method was verified through surface target simulation experiments. In this group of experiments, the imaging results of the Gaofen-3 (GF-3) satellite were used as the ground truth map, such as... Figure 3 As shown. Figure 3 (a) represents the illumination scene of transmitter 1. Figure 3 (b) represents the illumination scene of transmitter 2. The azimuth dimension of the two imaging scenes is 750m and the range dimension is 2100m. Figure 4 This representation shows echo separation results using different methods without loss of generality. Figure 4 Only the echo separation results for transmitter 1 are given. Figure 4 (a) shows the imaging result of the mixed signal obtained from the segmented phase-coded waveform. Note that region 1 contains only near-end interference (with the same angle of arrival as the desired signal), while region 2 contains both near-end and far-end interference (with different angles of arrival than the desired signal). Observation Figure 4 (a) It can be seen that if echo separation is not performed, the cross-correlation energy between the transmitted waveforms will seriously degrade the imaging results of the observed scene. Figure 4 (b) indicates the echo separation result under the STSO mechanism. Since the MIMO-SAR system only has two elevation dimension receiving channels, the limited elevation dimension resources result in residual interference energy in the imaging results, affecting the scene observation effect. Figure 4 (c) represents the echo separation result under the proposed method. It can be observed that the cross-correlation energy in the image has been completely suppressed, and the imaging quality is consistent with the ground truth image.

[0114] In summary, the proposed method can achieve echo separation even with limited elevation dimension resources in MIMO-SAR systems, effectively suppressing the cross-correlation energy between waveforms. This method has broad application prospects in practical situations.

Claims

1. A low-complexity MIMO-SAR echo separation method based on multi-sub-pulse separation and weighted synthesis, characterized in that: Includes the following steps: S1: Design the transmitted waveform and use a receiver to record the echo signals of multiple transmitted waveforms; S2: Perform matched filtering on the echo signals of multiple transmitted waveforms in the receiver; S3: Using beamforming technology to separate multiple sub-pulse echo signals in the pitch dimension; S4: Time-shift the separated sub-pulse signals to ensure that each sub-pulse signal from the same transmitter occupies the same time delay range; S5: Retrieve the solution encoding vector, i.e., the sub-pulse weighting coefficients, from the phase encoding matrix in step S1; S6: Weighted synthesis of the sub-pulse signals in S4 to achieve final MIMO-SAR echo separation; In step S1, a segmented phase-coded waveform is selected as the transmitted waveform for a multi-input multi-output synthetic aperture radar (MIMO-SAR). It is assumed that MIMO-SAR has... One transmitter, One receiver, segmented phase-coded waveform The form is expressed as: In the formula, This indicates the fast time, and k represents the k-th transmitter. Indicates the encoded phase. Indicates sub-pulse signal, Indicates the sub-pulse width. Indicates the transpose operation; The echo signal received by receiver Rxn after the M segmented phase-coded waveforms are backscattered by ground targets is represented as follows: In the formula, Indicates slow time. Indicates channel response, This indicates a fast-time convolution operation; In step S3, as can be seen from step S1, the time delay interval between adjacent sub-pulse signals is... In a distributed scenario, the echo signal is further represented as: In the formula, , , and These represent the time delay center, time delay length, azimuth slow time, and time delay corresponding to each distance unit in the scene, respectively. Let be the echo signal of the k-th sub-pulse transmitted by the m-th transmitter and received by the n-th receiver, and ; when At that time, the sub-pulse signals do not overlap, so separation can be performed directly in the fast time domain. At this time, aliasing will occur in the sub-pulse signals; furthermore, when hour, Each sub-pulse signal will have the same time delay range, and the echo signal within this time delay range is represented as: The sub-pulse signal is separated using beamforming technology. The separated sub-pulse signal is represented as follows: 。 2. The low-complexity MIMO-SAR echo separation method based on multi-sub-pulse separation and weighted synthesis as described in claim 1, characterized in that: In step S2, The result after matching filter processing of the echo signal Recorded as: In the formula, For Fast Time Fourier Transform, For fast-time inverse Fourier transform, For distance frequency, For frequency modulation slope, This is the kth sub-pulse signal.

3. The low-complexity MIMO-SAR echo separation method based on multi-sub-pulse separation and weighted synthesis as described in claim 1, characterized in that: In step S4, According to step S3, the first The time delay range corresponding to each sub-pulse signal is denoted as: The sub-pulse signal is time-shifted in the distance-frequency domain, i.e.: In the formula, Indicates the separated sub-pulse signal Sub-pulse signal after time shift This represents the time shift corresponding to each sub-pulse signal. To perform the Hadamard product operation.

4. The low-complexity MIMO-SAR echo separation method based on multi-sub-pulse separation and weighted synthesis as described in claim 1, characterized in that: In step S5, Pair pulse signal Decoding is performed to further suppress interference signals: the decoded vector is calculated as follows: In the formula, , and These represent the encoding matrix, encoding vector, and unit vector, respectively.

5. The low-complexity MIMO-SAR echo separation method based on multi-sub-pulse separation and weighted synthesis as described in claim 1, characterized in that: In step S6, The calculated decoding vector is used to weight and synthesize the sub-pulse signals to suppress residual interference signals. 。

Images