Small platform FMCW radar foresight high-resolution imaging method based on sparse sampling

Abstract

The invention discloses a small platform FMCW radar foresight high-resolution imaging method based on sparse sampling, and the method employs platform motion to synthesize a large-aperture virtual array so as to improve the radar foresight imaging azimuth resolution. According to the method, firstly, a uniform linear array is utilized, sampling is carried out at different moments through platform movement, and a preliminary imaging result of each channel is obtained through distance-angle estimation; and through adaptive wave beam forming processing, a null is formed at a false target (grating lobe and left and right blurring). And finally, performing coherent accumulation on imaging results of all channels to obtain high-resolution distance-angle imaging. Simulation and experiment results show that the method realizes high-resolution foresight imaging while significantly reducing the echo data volume.

Description

Technical Field

[0001] The present invention belongs to the field of radar imaging, and in particular relates to a small platform FMCW radar forward-looking high-resolution imaging method based on sparse sampling. Background Art

[0002] In recent years, forward-looking radar imaging technology has shown broad application prospects in scenarios such as unmanned aerial vehicle (UAV) environmental monitoring and autonomous driving. Real array-based forward-looking imaging, or DBS imaging, is a widely used method. However, the former is limited by the space of small platforms and struggles to achieve sufficiently high angular resolution. The latter often requires sampling a large amount of echo data, placing higher demands on the platform's signal processing capabilities and also suffers from Doppler ambiguity. Existing methods for small-platform FMCW radar forward-looking imaging primarily focus on improving angular resolution, resolving Doppler ambiguity, and reducing data volume.

[0003] Super-resolution algorithms such as MUSIC can improve the angular resolution of forward-looking imaging, but they require prior knowledge of the number of targets and multiple snapshots for optimal performance. Nested arrays, coprime arrays, or MIMO radars optimize the array structure, but the improvement in angular resolution is limited and the structure is complex.

[0004] In terms of Doppler ambiguity resolution, there is a method that uses bistatic SAR technology to obtain additional target information through another transmitting antenna in a different position to avoid the Doppler ambiguity problem, but it is not suitable for most scenarios; multi-channel forward-looking SAR technology can obtain high-resolution and ambiguity-free imaging results, but the amount of echo data is large, and the platform signal processing capabilities are required to be high.

[0005] In terms of reducing the amount of echo data, compressed sensing theory uses the sparsity of the signal to reconstruct the image under undersampling, avoiding the sampling of large amounts of data. However, the quality of the reconstructed image using this type of method depends on the performance of the observation matrix. Summary of the Invention

[0006] The present invention proposes a small platform FMCW radar forward high-resolution imaging method based on sparse sampling, which achieves high angular resolution while eliminating Doppler ambiguity in forward imaging and avoiding the problem of excessive echo data volume.

[0007] The technical solution to achieve this is: a forward-looking FMCW radar imaging method based on sparse sampling intervals. First, a uniform linear array is used to perform sampling at different times through platform motion, and preliminary imaging results for each channel are obtained through range-angle estimation. Then, adaptive beamforming is used to form nulls at false targets. Finally, the imaging results of each channel are coherently accumulated to obtain high-resolution range-angle imaging. The specific steps are as follows:

[0008] Step 1: A single-transmitter, multi-receiver uniform linear array radar is deployed at the front end of the platform, perpendicular to the platform's motion direction. It transmits an FMCW signal and uses the platform's motion to obtain sparsely sampled echo data. After mixing, a sparsely sampled difference frequency signal is obtained, and the process proceeds to step 2.

[0009] Step 2: Perform distance offset compensation, filter out the remaining video items, and perform distance movement correction. Build a sparsely sampled distance-angle difference frequency signal model, and then proceed to step 3.

[0010] Step 3: Perform range-dimensional FFT and angle-dimensional CBF processing on the data of each channel to obtain a preliminary range-angle image of the target. The preliminary range-angle image of the target includes grating lobes and blur, and then proceed to step 4.

[0011] Step 4: For targets in different range units, use the real array to roughly estimate their angles and calculate the angles of the grating lobes and blurred targets. Through adaptive beamforming, a null is generated at the position of the false target. The range and angle imaging results of each channel are weighted and coherently accumulated to obtain a range-angle image without ambiguity and grating lobes.

[0012] Compared with existing technologies, this invention offers significant advantages: It improves radar imaging angular resolution by synthesizing a large-aperture sparse virtual array through platform motion. Compared to traditional forward-looking imaging methods, this method significantly reduces the amount of echo data and lowers the signal processing requirements of the platform. Furthermore, it utilizes adaptive beamforming to suppress false targets, resulting in high-resolution range-angle images free of Doppler ambiguity and grating lobes. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 This is a flow chart of a small platform FMCW radar forward-looking high-resolution imaging method based on sparse sampling according to the present invention.

[0014] Figure 2 This is a diagram of the forward-looking imaging geometric model in an embodiment of the present invention.

[0015] Figure 3 These are the imaging results before algorithm processing (without false target suppression); (a) is the distance-angle diagram, and (b) is the angle-dimension slice of the distance-angle diagram.

[0016] Figure 4 These are the imaging results after algorithm processing (false target suppression); (a) is the distance-angle diagram, and (b) is the angle dimension slice of the distance-angle diagram.

[0017] Figure 5This figure compares the results of the proposed algorithm and other traditional algorithms in resolving targets in the angle dimension; (a) shows the result when the two targets are located on the same side of the front view area, and (b) shows the result when the two targets are located on opposite sides of the front view area. DETAILED DESCRIPTION

[0018] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0019] It should be noted that all directional indications in the embodiments of the present invention (such as up, down, left, right, front, back, etc.) are only used to explain the relative position relationship, movement status, etc. between the various components under a certain specific posture (as shown in the accompanying drawings). If the specific posture changes, the directional indication will also change accordingly.

[0020] In addition, the terms "first," "second," and so on, used in this disclosure are for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of technical features being referenced. Thus, features specified as "first" or "second" may explicitly or implicitly include at least one such feature. In the description of this disclosure, "plurality" means at least two, such as two or three, unless otherwise specifically defined.

[0021] In the present invention, unless otherwise specified or limited, the terms "connection" and "fixation" should be understood in a broad sense. For example, "fixation" can refer to fixed connection, detachable connection, or integration; "connection" can refer to mechanical connection or electrical connection. Those skilled in the art will understand the specific meanings of the above terms in the present invention based on specific circumstances.

[0022] In addition, the technical solutions between the various embodiments of the present invention can be combined with each other, but it must be based on the fact that ordinary technicians in this field can implement it. When the combination of technical solutions is mutually contradictory or cannot be implemented, it should be deemed that such a combination of technical solutions does not exist and is not within the scope of protection required by the present invention.

[0023] The following will further introduce the specific implementation methods, as well as the technical difficulties and inventive points of this invention in combination with this design example.

[0024] Example 1: High-resolution forward imaging of point targets:

[0025] like Figure 1As shown in FIG, a forward-looking FMCW radar imaging method based on sparse sampling intervals is characterized by first using a uniform linear array to perform sampling at different times through platform motion, and obtaining preliminary imaging results of each channel through range-angle estimation; then, through adaptive beamforming processing, a null is formed at the false target; finally, the imaging results of each channel are coherently accumulated to obtain high-resolution range-angle imaging. The specific steps are as follows:

[0026] Step 1: Place a single-transmitter, multiple-receiver uniform linear array radar at the front of the platform, perpendicular to the platform's motion direction, and transmit an FMCW signal. Utilize the platform's motion to obtain sparsely sampled echo data, which is then mixed to obtain a sparsely sampled difference frequency signal. The details are as follows:

[0027] Figure 2 is the radar forward imaging geometric model, and Table 1 lists the radar parameters.

[0028] Table 1

[0029] Parameter Value Carrier frequency f0 / (GHz) 77 Bandwidth B / (GHz) 1 Modulation period T / (us) 100 Number of receiving elements M 8 Wavelength λ / (m) 0.004 Element spacing d / (m) 0.002 Velocity v / (m / s) 20

[0030] A single-transmitter, multi-receiver uniform linear array moves at a speed of v = 20 m / s along the normal direction of the array. The number of receiving elements M = 8 and the element spacing d = 0.002 m. k is the target, θ k is the angle between the kth target and the normal direction of the uniform linear array, that is, the target angle. R0 = 20m is the initial distance between the radar and the target, T s =400us is the sampling time interval, and the number of sampling times is 25 times.

[0031] R k (t,m) is the instantaneous slant range between the mth receiving element of the radar and the target, R k (t, s) is the instantaneous slant range between the transmitting array element and the target, which can be expressed as:

[0032]

[0033] Among them, R0(k,m) is the initial distance from the mth receiving element of the radar to the target, and R0(k,s) is the initial distance from the radar transmitting element to the target.

[0034] Since the array is arranged compactly, the difference between R0(k,m) and R0(k,s) is much smaller than the range resolution of the system and can be considered to be approximately equal.

[0035] Assuming that R0(k,m) is approximately equal to R0(k,s), we can get:

[0036] R k (t,m)=R k (t,s)+dmsinθk

[0037] If the transmitted signal is FMCW, it can be expressed as:

[0038]

[0039] in, Represents a single cycle transmission signal, f0 is the carrier frequency, T m is the modulation period, modulation slope β=B / T m , B represents bandwidth, is the fast time, that is, the time within the modulation cycle, the cycle number n = 0, 1, ... N-1, and N is the total number of modulation cycles, that is, the number of sampling times, and j represents the imaginary part.

[0040] After the transmitted signal is reflected by the target, the echo signal s received by the mth receiving element is r (t,m) is:

[0041]

[0042] Among them, the receiving array element number m=0,1,…M-1, K is the total number of targets, is the round-trip time of the signal between the mth receiving element and the kth target, which is expressed as:

[0043]

[0044] Where t is the total time, t n is the slow time, and t n =inT m , the period number n=0,1,…N-1, i is the sampling interval coefficient, when i>1, it means sparse sampling, c represents the speed of light;

[0045] The echo signal of the mth receiving element is mixed with the reference signal to obtain the difference frequency signal, which is the sparsely sampled difference frequency signal s beat (t,m):

[0046]

[0047] Step 2: Perform distance offset compensation on the sparsely sampled difference frequency signal, filter out the remaining video items, and perform distance movement correction to construct a sparsely sampled distance-angle difference frequency signal model, as follows:

[0048] Ignore the third-order and higher-order terms, and k (t,s) in fast time Performing Taylor expansion at , we get:

[0049]

[0050] Among them, R n is the instantaneous slope distance under the “stop-go-stop” model, k n is the error caused by the continuous motion of the platform within the modulation period;

[0051]

[0052] Then s beat (t,m) can be re-expressed as:

[0053]

[0054] Among them, the third exponential term corresponds to the distance offset error, and the fourth exponential term has It can be seen that its impact on the result is far less than that of the third exponential term, so the fourth exponential term can be ignored; the fifth exponential term is the residual video phase generated by the difference frequency processing, which is eliminated by distance domain matched filtering.

[0055] The frequency error Δf caused by the third exponential term is:

[0056]

[0057] The corresponding distance error ΔR in the distance dimension is:

[0058]

[0059] Among them, f d is the Doppler frequency, and λ is the wavelength.

[0060] The ratio of range offset error to range resolution is defined as Δp:

[0061]

[0062] Among them, f m is the signal modulation frequency, the distance resolution ρ r =c / 2B.

[0063] When Δp is greater than 0.5, the range dimension will experience focus blur. For high-frequency, high-speed FMCW radar platforms, the Doppler frequency is often very large, which may lead to a non-negligible impact of the range offset error, which needs to be compensated during the imaging process.

[0064] For the first exponential term R n Do Taylor expansion and ignore higher-order terms, the expression is:

[0065]

[0066] In the above formula, the second term represents the range walk and the third term represents the range curvature. In the forward viewing area, compared with the oblique viewing area, the angle θ between the target and the array is k The value of is very small, and the influence of distance movement is much greater than that of distance bending. When the accuracy requirement is not particularly high, distance bending can be ignored.

[0067] After the above compensation and neglect processing, the expression of the front view difference frequency signal s is obtained beat (t,n,m) is as follows:

[0068] In the above formula, the first exponential term corresponds to distance information, the second exponential term corresponds to velocity and angle information, and indicates that distance and velocity are coupled, and the third exponential term corresponds to angle information, and indicates that distance and array element position are coupled. In most scenarios, the distance between the radar and the target is much greater than the array length, so the effect of array element position on distance can be ignored.

[0069] Fast Time Do time-frequency replacement, that is Then the Keystone transform is used to eliminate the distance-velocity coupling and the slow time t n Perform scale transformation, so Then the front video frequency signal S′ after Keystone transformation beat (f r ,n,m) is expressed as:

[0070]

[0071] Among them, f r represents the instantaneous frequency, t′ n Represents the slow time after scale transformation.

[0072] The above formula is the difference frequency signal after distance offset compensation, residual video item filtering and distance movement correction, based on which the sparse sampling distance-angle difference frequency signal model X is constructed. m (θ,R), expressed as follows:

[0073] X m (θ, R)=A(θ)S(k)B(R)+ε

[0074] Where R represents distance, θ represents angle, A(θ) is the slow-time domain angle flow matrix, S(k) is the signal matrix, B(R) is the distance flow matrix, and ε is Gaussian white noise.

[0075] A(θ)=[a(θ1)a(θ2)…a(θ K )] N×K

[0076] S(k)=diag[s1 s2…s K ] K×K

[0077]

[0078] in,

[0079]

[0080] s k =exp(-j2πdmsinθ k / λ)exp(j4πR0(k,s) / λ)

[0081] b(R k )=exp(j4πf r R0(k,s) / c)

[0082] Where T is the matrix transpose symbol, diag(·) represents the diagonal matrix, L is the number of fast time domain sampling points, a(θ k ) represents the steering vector, s k Represents the sampled signal.

[0083] Step 3: Based on the sparsely sampled range-angle difference frequency signal model, perform range-dimensional FFT and angle-dimensional CBF processing on the data of each channel to obtain a preliminary range-angle image of the target containing grating lobes and blur, as follows:

[0084] Assume that there is a target P on one side of the radar platform trajectory k1 (R k sinθ k ,R k cosθ k ), after receiving the target echo, FFT is performed in the fast time domain to achieve range focusing. The range focusing result of the mth receiving array element is f(f r ,m) is:

[0085]

[0086] For target P k1 The slow time domain data corresponding to the distance unit is used to perform CBF angle estimation to obtain the distance-angle estimation f(m,R k ,θ k ),like Figure 3 As shown in (a), the angle estimation results are as follows Figure 3 (b) shown.

[0087] According to the previous analysis, we know that in the obtained distance-angle image, in addition to the real target P k1 In addition, there will be a fuzzy target P k2The false targets caused by grating lobes, thus the range-angle estimation results for:

[0088]

[0089] Where Q is the total amount of grating lobes on one side of the motion trajectory, q represents the grating lobe number on one side of the motion trajectory, and the target complex scattering coefficient set is

[0090] The range-angle images of each channel are coherently accumulated, and the imaging result of the M receiving array element channels, that is, the preliminary range-angle image F(R,θ) of the target including grating lobes and blur, is:

[0091]

[0092] in, is an array manifold containing blurred targets and grating lobes, and a(θ k )=[1exp(-j2πdsinθ k / λ)…exp(-j2π(M-1)dsinθ k / λ)] T .

[0093] Step 4: Based on the preliminary range-angle image of the target containing grating lobes and blur, a real array is used to roughly estimate the angle of the target in different range units, and the angle between the grating lobes and the blurred target is calculated. Through adaptive beamforming, a null is generated at the position of the false target. The range-angle imaging results of each channel are weighted and coherently accumulated to obtain a range-angle image without blur and grating lobes, as shown below:

[0094] By utilizing the degrees of freedom of the real array, adaptive beamforming is realized in the spatial domain according to the LCMV criterion, and a null is formed at the position of the false target in the angle dimension. The cost function of LCMV is:

[0095] and

[0096] Among them, R XX represents the autocorrelation matrix of the signal, f represents the constrained response vector, w represents the weight vector, and H represents the conjugate transpose.

[0097] The Lagrange multiplier algorithm is used to solve the above constrained optimization problem, and the weight vector w of LCMV is obtained as:

[0098]

[0099] The imaging results of each channel are weighted to obtain the imaging results after deblurring and suppressing grating lobes, that is, the range-angle image without blur and grating lobes. Expressed as:

[0100]

[0101] Imaging results such as Figure 4 As shown. Figure 4 (a) is the distance-angle imaging result after algorithm processing, Figure 4 (b) is the imaging angle-dimensional slice after algorithm processing, and the false targets are suppressed by more than 30dB.

[0102] Example 2: Comparison with other traditional algorithms:

[0103] In this embodiment, the proposed algorithm is compared with the traditional CBF and DBF algorithms. The CBF algorithm uses an 8-element real array for imaging, and the number of continuous sampling of the DBS algorithm is the same as that of the proposed algorithm, which is 30 times.

[0104] like Figure 5 As shown in (a), the two targets are at angles of 45° and 47°, respectively. The proposed algorithm effectively suppresses false targets, and the target angles are estimated to be 44.9° and 46.9°, respectively, indicating that the algorithm can effectively distinguish the two targets. In contrast, the CBF algorithm has poor resolution and is insufficient to distinguish the two targets. The DBS algorithm has higher resolution than the CBF algorithm, but still cannot effectively distinguish the two targets and suffers from Doppler ambiguity.

[0105] like Figure 5 As shown in (b), the angles of the two point targets are 45° and -47°, respectively. The proposed algorithm estimates the target angles to be 44.9° and -46.9°, respectively, verifying the effectiveness of the algorithm. In comparison, the CBF algorithm estimates poorly. The DBS algorithm has a higher resolution than the CBF algorithm, but still lower than the proposed algorithm.

Claims

1. A FMCW radar forward imaging method based on sparse sampling intervals, characterized in that: First, a uniform linear array is used to perform sampling at different times through platform movement, and preliminary imaging results of each channel are obtained through range-angle estimation. Then, adaptive beamforming is used to form a null at the false target. Finally, the imaging results of each channel are coherently accumulated to obtain high-resolution range-angle imaging.

2. The FMCW radar forward imaging method based on sparse sampling interval according to claim 1, characterized in that: Here are the steps: Step 1: A single-transmitter, multi-receiver uniform linear array radar is deployed at the front end of the platform, perpendicular to the platform's motion direction. It transmits an FMCW signal and uses the platform's motion to obtain sparsely sampled echo data. After mixing, a sparsely sampled difference frequency signal is obtained, and the process proceeds to step 2. Step 2: Perform distance offset compensation, filter out the remaining video items, and perform distance movement correction. Build a sparsely sampled distance-angle difference frequency signal model, and then proceed to step 3. Step 3: Perform range-dimensional FFT and angle-dimensional CBF processing on the data of each channel to obtain a preliminary range-angle image of the target. The preliminary range-angle image of the target includes grating lobes and blur, and then proceed to step 4. Step 4: For targets in different range units, use the real array to roughly estimate their angles and calculate the angles of the grating lobes and blurred targets. Through adaptive beamforming, a null is generated at the position of the false target. The range and angle imaging results of each channel are weighted and coherently accumulated to obtain a range-angle image without ambiguity and grating lobes.

3. The FMCW radar forward imaging method based on sparse sampling interval according to claim 2, characterized in that: In step 1, a single-transmitter, multi-receiver uniform linear array radar is placed at the front end of the platform, perpendicular to the platform's motion direction, to transmit an FMCW signal. Utilizing the platform's motion, sparsely sampled echo data is obtained, and after mixing, a sparsely sampled difference frequency signal is obtained, as follows: A single-transmitter, multi-receiver uniform linear array radar moves at a uniform speed v along the array normal direction. The total number of receiving array elements is M, the distance between adjacent array elements is d, and s represents the transmitting array element. Let P k is the target, θ k is the angle between the kth target and the normal direction of the uniform linear array, that is, the target angle; R0 is the initial distance between the radar and the target, T s is the sampling time interval, t represents the full time; R k (t,m) is the instantaneous slant range between the mth receiving element of the radar and the target, R k (t, s) is the instantaneous slant range between the transmitting array element and the target, which can be expressed as: Where R0(k,m) is the initial distance from the mth receiving element of the radar to the target, and R0(k,s) is the initial distance from the radar transmitting element to the target; Assuming that R0(k,m) is approximately equal to R0(k,s), we can get: R k (t,m)=R k (t,s)+dmsinθ k If the transmitted signal is FMCW, it can be expressed as: in, Represents a single cycle transmission signal, f0 is the carrier frequency, T m is the modulation period, modulation slope β=B / T m , B represents bandwidth, is the fast time, i.e. the time within the modulation cycle, the cycle number n = 0, 1, ... N-1, and N is the total number of modulation cycles, i.e. the number of sampling times, and j represents the imaginary part; After the transmitted signal is reflected by the target, the echo signal s received by the mth receiving element is r (t,m) is: Among them, the receiving array element number m=0,1,…M-1, K is the total number of targets, is the round-trip time of the signal between the mth receiving element and the kth target, which is expressed as: Where t is the total time, t n is the slow time, and t n =inT m , the period number n=0,1,…N-1, i is the sampling interval coefficient, when i>1, it means sparse sampling, c represents the speed of light; The echo signal of the mth receiving element is mixed with the reference signal to obtain the difference frequency signal, which is the sparsely sampled difference frequency signal s beat (t,m):

4. The FMCW radar forward imaging method based on sparse sampling interval according to claim 3, characterized in that: In step 2, the sparsely sampled difference frequency signal is compensated for distance offset, the remaining video items are filtered out, and the distance movement correction is performed to construct a sparsely sampled distance-angle difference frequency signal model, as follows: Ignore the third-order and higher-order terms, and k (t,s) in fast time Performing Taylor expansion at , we get: Among them, R n is the instantaneous slope distance under the "stop-go-stop" model, k n is the error caused by the continuous motion of the platform within the modulation period; Then s beat (t,m) can be re-expressed as: Among them, the third exponential term corresponds to the distance offset error, and the fourth exponential term has It can be seen that its impact on the result is far less than the third exponential term, so the fourth exponential term can be ignored; the fifth exponential term is the residual video phase generated by the difference frequency processing, which is eliminated by distance domain matched filtering: The frequency error Δf caused by the third exponential term is: The corresponding distance error ΔR in the distance dimension is: Among them, f d is the Doppler frequency, λ is the wavelength; The ratio of range offset error to range resolution is defined as Δp: Among them, f m is the signal modulation frequency, the distance resolution ρ r =c / 2B; When Δp is greater than 0.5, the range dimension will experience focus blur. For high-frequency, high-speed FMCW radar platforms, the Doppler frequency is often very large, resulting in a non-negligible impact of the range offset error, which needs to be compensated during the imaging process. For the first exponential term R n Do Taylor expansion and ignore higher-order terms, the expression is: In the above formula, the second term represents the range walk and the third term represents the range curvature. In the forward viewing area, compared with the oblique viewing area, the angle θ between the target and the array is k The value of is very small, and the influence of distance movement is much greater than that of distance bending. When the accuracy requirement is not particularly high, distance bending can be ignored. After the above compensation and neglect processing, the expression of the front view difference frequency signal s is obtained beat (t,n,m) is as follows: In the above formula, the first exponential term corresponds to distance information, the second exponential term corresponds to velocity and angle information, and indicates that there is a coupling relationship between distance and velocity, and the third exponential term corresponds to angle information, and indicates that there is a coupling relationship between distance and array element position; Fast Time Do time-frequency replacement, that is Then the Keystone transform is used to eliminate the distance-velocity coupling and the slow time t n Perform scale transformation, so Then the front video frequency signal S′ after Keystone transformation beat (f r ,n,m) is expressed as: Among them, f r represents the instantaneous frequency, t n ′ represents the slow time after scale transformation; The above formula is the difference frequency signal after distance offset compensation, residual video item filtering and distance movement correction, based on which the sparse sampling distance-angle difference frequency signal model X is constructed. m (θ,R), expressed as follows: X m (θ, R)=A(θ)S(k)B(R)+ε Where R represents distance, θ represents angle, A(θ) is the slow time domain angle flow matrix, S(k) is the signal matrix, B(R) is the distance flow matrix, and ε is Gaussian white noise. A(θ)=[a(θ1)a(θ2)…a(θ K )] N×K S(k)=diag[s1 s2…s K ] K×K in, s k =exp(-j2πdmsinθ k / λ)exp(j4πR0(k,s) / λ) b(R k )=exp(j4πf r R0(k,s) / c) Where T is the matrix transpose symbol, diag(·) represents the diagonal matrix, L is the number of fast time domain sampling points, a(θ k ) represents the steering vector, s k Represents the sampled signal.

5. The FMCW radar forward imaging method based on sparse sampling interval according to claim 4, characterized in that: In step 3, based on the sparsely sampled range-angle difference frequency signal model, the data of each channel is processed by performing range-dimensional FFT and angle-dimensional CBF to obtain a preliminary range-angle image of the target containing grating lobes and blur, as follows: Assume that there is a target P on one side of the radar platform trajectory k1 (R k sinθ k ,R k cosθ k ), after receiving the target echo, FFT is performed in the fast time domain to achieve range focusing. The range focusing result of the mth receiving array element is f(f r ,m) is: For target P k1 The slow time domain data corresponding to the distance unit is used to perform CBF angle estimation to obtain the distance-angle estimation f(m,R k ,θ k ); Then we can get the distance-angle image, in addition to the real target P k1 In addition, there will be a fuzzy target P k2 The false targets caused by grating lobes, thus the range-angle estimation results for: Where IQ is the total amount of grating lobes on one side of the motion trajectory, q represents the grating lobe number on one side of the motion trajectory, and the target complex scattering coefficient set is The range-angle images of each channel are coherently accumulated, and the imaging result of the M receiving array element channels, that is, the preliminary range-angle image F(R,θ) of the target including grating lobes and blur, is: in, is an array manifold containing blurred targets and grating lobes, and a(θ k )=[1exp(-j2πdsinθ k / λ)…exp(-j2π(M-1)dsinθ k / l)] T 。 6. The FMCW radar forward imaging method based on sparse sampling interval according to claim 5, characterized in that: In step 4, based on the preliminary range-angle image of the target containing grating lobes and blur, a real array is used to roughly estimate the angle of the target in different range units, and the angle between the grating lobes and the blurred target is calculated. Through adaptive beamforming, a null is generated at the position of the false target. The range-angle imaging results of each channel are weighted and coherently accumulated to obtain an unambiguous and grating-lobe-free range-angle image, as follows: By utilizing the degrees of freedom of the real array, adaptive beamforming is realized in the spatial domain according to the LCMV criterion, and a null is formed at the position of the false target in the angle dimension. The cost function of LCMV is: and Among them, R XX represents the autocorrelation matrix of the signal, f represents the constraint response vector, w represents the weight vector, and H represents the conjugate transpose; The Lagrange multiplier algorithm is used to solve the above constrained optimization problem, and the weight vector w of LCMV is obtained as: The imaging results of each channel are weighted to obtain the imaging results after deblurring and suppressing grating lobes, that is, the range-angle image without blur and grating lobes. Expressed as:

Images