CN111044996A - LFMCW radar target detection method based on dimension reduction approximate message transfer - Google Patents

LFMCW radar target detection method based on dimension reduction approximate message transfer Download PDF

Info

Publication number
CN111044996A
CN111044996A CN201911333074.6A CN201911333074A CN111044996A CN 111044996 A CN111044996 A CN 111044996A CN 201911333074 A CN201911333074 A CN 201911333074A CN 111044996 A CN111044996 A CN 111044996A
Authority
CN
China
Prior art keywords
target
dimension reduction
data cube
detection
time domain
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201911333074.6A
Other languages
Chinese (zh)
Inventor
晋本周
阙中元
张小飞
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
JIANGSU HUIWEIXUN INFORMATION TECHNOLOGY Co.,Ltd.
Original Assignee
Nanjing University of Aeronautics and Astronautics
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Nanjing University of Aeronautics and Astronautics filed Critical Nanjing University of Aeronautics and Astronautics
Priority to CN201911333074.6A priority Critical patent/CN111044996A/en
Publication of CN111044996A publication Critical patent/CN111044996A/en
Pending legal-status Critical Current

Links

Images

Classifications

    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S13/00—Systems using the reflection or reradiation of radio waves, e.g. radar systems; Analogous systems using reflection or reradiation of waves whose nature or wavelength is irrelevant or unspecified
    • G01S13/02—Systems using reflection of radio waves, e.g. primary radar systems; Analogous systems
    • G01S13/04—Systems determining presence of a target
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
    • G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
    • G—PHYSICS
    • G01—MEASURING; TESTING
    • G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
    • G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
    • G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
    • G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
    • G01S7/414—Discriminating targets with respect to background clutter

Abstract

A LFMCW radar target detection method based on dimension reduction approximate message transfer comprises the following steps: the receiving end carries out conventional coherent accumulation on the single-frame baseband data cube to obtain a three-dimensional frequency domain data cube; performing target pre-detection on the data cube after the coherent accumulation to obtain pre-detection target point trace information; calculating a dimension reduction observation matrix based on the trace point information obtained by target pre-detection, arranging a baseband data cube into a vector, and establishing a radar dimension reduction observation model; reconstructing a target signal by using a radar dimension reduction observation model through a generalized approximate message transfer algorithm; and carrying out constant false alarm detection by using the reconstructed signal. Compared with the traditional target detection method based on fast Fourier transform, the method provided by the invention has better target detection performance.

Description

LFMCW radar target detection method based on dimension reduction approximate message transfer
Technical Field
The invention belongs to the technical field of radar target detection, and particularly relates to an LFMCW radar target detection method based on dimension reduction approximate message passing (RD-GAMP).
Background
Conventional linear signal processing methods, such as Fast Fourier Transform (FFT), have been widely used for radar signal processing. Taking the vehicle-mounted LFMCW radar as an example, the baseband signal obtained by subjecting the received signal to deskew processing can be modeled as the sum of a series of complex point frequency signals in the space domain, the fast time domain and the slow time domain, respectively. To improve the signal-to-noise ratio (SNR) of the target, a fast fourier transform is employed to achieve efficient signal processing. However, to reduce clutter and sidelobes of strong targets, the sampled data is typically weighted before performing a fast fourier transform, resulting in a loss of signal-to-noise ratio.
The method based on compressed sensing provides a feasible scheme for solving the problems, and the target can be directly reconstructed by utilizing the sparsity of the signal. In many application scenarios, the signal of interest is sparse, since the number of targets is usually much smaller than the number of resolution cells. Compared with the traditional signal processing method, the method based on compressed sensing directly reconstructs based on original data, and the signal processing loss can be reduced.
In recent years, a method called Approximate Message Passing (AMP) in the field of compressed sensing has received attention from many scholars. The approximate message transmission method is obtained by utilizing Gaussian and quadratic approximation on the basis of a circulation belief propagation method in a graph model. Based on the idea of message passing, the method can be used to solve the problem of base tracking or base tracking noise reduction. At large system limits, the phase change curve of the approximate message passing method approaches l1The phase change curve of the norm minimization method has better signal recovery performance. Research by some scholars shows that the approximate message transmission method can be further popularized and applied to the situation with arbitrary signal prior distribution and noise distribution. This generalized approach is referred to as Generalized Approximate Messaging (GAMP). The generalized approximate message transmission method provides a way to efficiently solve the problem of sparse signal reconstruction.
However, there are also many factors to consider when applying the generalized approximate messaging approach. In practice, the goal is not on grid points, so the granularity of grid points needs to be a compromise between accuracy and algorithm stability. Although the finer grid points can reduce the reconstruction error theoretically, the problem of unstable values may occur. In addition, although the generalized approximate message passing method is low in computational complexity, it is difficult to directly apply the generalized approximate message passing method to radar signal processing. The reason is that the observation matrix dimension in the radar observation model is very high, and the direct application of the generalized approximate message transfer algorithm still needs a large amount of calculation and storage, which is difficult to satisfy in practical application.
Disclosure of Invention
The invention provides a target detection method of an LFMCW radar based on dimension reduction approximate message transfer, aiming at the problem of signal processing loss of a traditional target detection method of fast Fourier transform, which reduces the dimension of a signal observation model through target pre-detection, utilizes the dimension reduction observation model and a generalized approximate message transfer algorithm to reconstruct signals and detect targets, reduces the signal processing loss of a conventional method and improves the target detection performance.
In order to achieve the purpose, the invention adopts the following technical scheme:
a LFMCW radar target detection method based on dimension reduction approximate message transfer is characterized by comprising the following steps:
1) the receiving end obtains a single-frame baseband data cube according to the received signal, and windowing is respectively carried out on the single-frame baseband data cube in a fast time domain, a slow time domain and a space domain;
2) for the windowed baseband data cube, realizing coherent accumulation of a target signal based on three-dimensional fast Fourier transform (3D-FFT) to obtain a three-dimensional frequency domain data cube;
3) performing target pre-detection on the three-dimensional frequency domain data cube to obtain pre-detected target trace information;
4) calculating a dimension reduction observation matrix based on the pre-detected target trace information, arranging the baseband data cubes according to vectors, and establishing a radar dimension reduction observation model;
5) reconstructing a target signal based on a radar dimension reduction observation model;
6) and carrying out constant false alarm detection by using the reconstructed target signal.
In order to optimize the technical scheme, the specific measures adopted further comprise:
further, in step 1), before the baseband data cube is obtained, deskewing is performed on the received signal, so as to obtain a difference frequency signal.
Further, in step 1), the window functions include a chebyshev window and a hamming window, and different window functions are respectively selected for a fast time domain, a slow time domain and a space domain.
Further, in the step 3), for a three-dimensional frequency domain data cube obtained after fast Fourier transform, performing three-dimensional target pre-detection on each unit; and determining the target pre-detection threshold according to the requirement of dimension reduction, and recording the corresponding fast time domain unit number, slow time domain unit number and airspace unit number of each trace point passing through the detection threshold.
Further, in step 4), based on the pre-detected target trace information, a dimension reduction observation matrix is calculated, specifically as follows:
401) the baseband data cube is represented in vector form: assuming that the distance unit, the pulse number and the antenna unit number are N, k and L respectively, sequentially arranging the baseband data cubes according to the distance unit, the airspace unit and the pulse index to obtain an observation vector
Figure BDA0002328013160000021
402) Constructing a dimension reduction observation matrix: assume that the number of target traces to be pre-detected is IpdIthpdThe Doppler, space domain frequency and fast time domain difference frequency corresponding to the point trace are respectively
Figure BDA0002328013160000022
And
Figure BDA0002328013160000023
then, the dimension reduction observation matrix is:
Aκ＝[aκ(0)，...，aκ(ipd)，...，aκ(Ipd-1)]，
wherein the content of the first and second substances,
Figure BDA0002328013160000031
symbol
Figure BDA0002328013160000032
Which represents the product of the Kronecker reaction,
Figure BDA0002328013160000033
and
Figure BDA0002328013160000034
respectively as follows:
Figure BDA0002328013160000035
Figure BDA0002328013160000036
Figure BDA0002328013160000037
403) establishing a radar dimension reduction observation model:
r＝Akxk+w
wherein the content of the first and second substances,
Figure BDA0002328013160000038
xκ(ipd) The complex amplitude of the target is represented by corresponding Doppler frequency, space domain frequency and fast time domain difference frequency
Figure BDA0002328013160000039
And
Figure BDA00023280131600000310
w denotes a noise vector.
Further, in step 5), a generalized approximate message passing algorithm is used for reconstructing the target signal.
The invention has the beneficial effects that: the invention provides an LFMCW radar target detection method based on dimension reduction approximate message transfer by performing coherent accumulation and target pre-detection on a radar baseband data cube; selecting a corresponding pre-detection threshold according to the actual requirement on the dimensionality reduction degree, and further adjusting the number of pre-detection target point traces obtained by a target pre-detector; constructing a dimension reduction observation model according to the trace result of the pre-detection target point, greatly reducing the dimension of an observation matrix, being beneficial to reducing the calculated amount of a subsequent algorithm, and further solving the problem that the reconstruction algorithm cannot be directly applied due to overhigh dimension of the observation matrix; and reconstructing signals and detecting the target by using a dimension reduction observation model and a generalized approximate message transfer algorithm. The method reduces the loss of signal-to-noise ratio of signal processing because a window function is not used for weighting the sampling data, thereby having better target detection performance.
Drawings
Fig. 1 is a schematic flow chart of an LFMCW radar target detection method based on dimension reduction approximate message transfer.
Fig. 2 is a schematic diagram of a receive baseband data cube.
FIG. 3 is a simulation result of target pre-detection before constructing a dimension-reduced observation model.
Fig. 4 shows the simulation results of the reconstruction of the target signal by the GAMP algorithm and the RD-GAMP algorithm under the condition of the coarse grid.
Fig. 5 shows the simulation results of reconstruction of the target signal by the GAMP algorithm and the RD-GAMP algorithm under the fine grid condition.
FIG. 6 is a ROC curve of the object detection method of the present invention.
FIG. 7 is a Doppler distance spectrum of measured data used to verify the performance of the proposed method of the present invention.
Fig. 8 is a graph of false alarm rate for target detection with the present invention and conventional methods.
FIG. 9 is a comparison of target detection performance curves for the present invention and conventional methods.
Detailed Description
The present invention will now be described in further detail with reference to the accompanying drawings.
The LFMCW radar target detection method based on dimension reduction approximate message passing as shown in fig. 1 includes the following steps:
1) the receiving end respectively windows the single-frame baseband data cube in a fast time domain, a slow time domain and a space domain;
2) for the windowed baseband data cube, realizing coherent accumulation of a target signal based on three-dimensional fast Fourier transform (3D-FFT) to obtain a three-dimensional frequency domain data cube;
3) performing target pre-detection on the three-dimensional frequency domain data cube to obtain pre-detected target trace information;
4) calculating a dimension reduction observation matrix based on the trace point information of the target pre-detection, arranging the baseband data cubes according to vectors, and establishing a radar dimension reduction observation model;
5) reconstructing a target signal by using a baseband data cube and a dimension reduction observation matrix through a generalized approximate message transfer algorithm;
6) and carrying out constant false alarm detection by using the reconstructed signal.
The invention is suitable for linear frequency modulation pulse continuous wave system (LFMCW) radars.
Fig. 2 is a schematic diagram of a baseband data cube after deskew of received signals.
The transmitted signal is a chirp, assuming the number of pulses is K. The system antenna unit is L. At the receiver, the signal received by the ith antenna element may be represented as:
Figure BDA0002328013160000041
wherein L is more than or equal to 0 and less than or equal to L-1, K is more than or equal to 0 and less than or equal to K-1, fcIs the clutter frequency, P is the number of targets,
Figure BDA0002328013160000042
is the p-th target complex amplitude. Tau isp(T) is the target delay, TIFor pulse intervals, μ is the chirp rate, fsp，pIs the spatial domain frequency, and ω (t) is additive white gaussian noise. rect is a rectangular window function. After deskew, the received signal is:
Figure BDA0002328013160000043
let τ bemaxDenotes the maximum target delay, 0 < τp(t)≤TmaxThe k-th pulse effective observation time interval is (kT)I+Tmax，(k+1)TI]. Considering the valid time interval, ignoring range-doppler coupling and range walk, the deskewed signal can be further approximated as:
Figure BDA0002328013160000051
wherein f isd，p，fr，pAnd fsp，pRespectively expressed as the doppler, difference and spatial frequencies, σ, of the pth targetpTo target complex amplitude, ω (t) is additive white Gaussian noise.
At the Nyquist rate of rl(t) after sampling and discretization, the single frame of data can be represented as a three-dimensional data cube, as shown in FIG. 2. Arranging the data cube into a vector r, specifically, a space-fast-time domain data matrix corresponding to the k-th pulse
Figure BDA0002328013160000052
Arranged according to a line to obtain a vector r corresponding to the kth pulsekThen, arranging all pulse data according to the pulse sequence number to obtain the current frame observation vector r. Then, the single frame echo data can be represented in the form of a matrix as follows:
r＝Ax+w (4)
wherein, A represents an observation matrix,
Figure BDA0002328013160000053
x is the target vector to be reconstructed and w is the noise vector. A. thed、AspAnd ArRespectively representing slow time, spatial and fast time domain observation matrices. Md、MspAnd MrRespectively representing the number of grid points divided by the three domains. In general, Md≥K、MspNot less than L and Mr≥N。AdM ofdThe columns are as follows:
Figure BDA0002328013160000054
Aspm ofspThe columns are as follows:
Figure BDA0002328013160000055
Arm ofrThe columns are as follows:
Figure BDA0002328013160000056
suppose that for a specific target, the Doppler frequency, the spatial domain frequency domain and the fast time difference frequency are respectively
Figure BDA0002328013160000057
Crown
Figure BDA0002328013160000058
Complex amplitude of sigmamThen x (m) ═ σmWherein m is mdmspmr. Here, assume that the target is located On a grid point (On-grid). When the target is away from the grid (Off-grid), as long as the grid point division is fine enough, the method can still realize the successful reconstruction of the target.
The dimensionality of the matrix A and the vector x in the formula (4) is very high, and the calculation amount for directly applying the GAMP algorithm to detect the target is very large and is difficult to bear. Meanwhile, in the case of an off-grid, the reconstruction accuracy is poor. Based on the reasons, the invention provides the LFMCW radar target detection by adopting a dimension reduction approximate message transmission method.
The following will explain the practical implementation process and advantages of the method by combining the experimental results.
Step 1: the receiving end respectively windows the single-frame baseband data cube in a fast time domain, a slow time domain and a space domain; the window function can be selected as common window functions such as a Chebyshev window, a Hamming window and the like according to actual conditions, and different window functions can be selected for a fast time domain, a slow time domain and a space domain;
step 2: and for the windowed data cube, realizing coherent accumulation of a target signal based on three-dimensional fast Fourier transform (3D-FFT) to obtain a three-dimensional frequency domain data cube.
And step 3: performing target pre-detection on the three-dimensional frequency domain data cube to obtain pre-detected target trace information; and the target pre-detection threshold is determined according to the requirement of dimension reduction. And recording the number of the corresponding fast time domain unit, slow time domain unit and airspace unit for each trace point passing through the detection threshold.
And 4, step 4: calculating a dimension reduction observation matrix based on the trace point information of the target pre-detection, arranging the baseband data cubes according to vectors, and establishing a radar dimension reduction observation model; the method comprises the following steps:
firstly, expressing a baseband three-dimensional data cube into a vector form; assuming that the distance unit, the pulse number and the antenna unit number are N, k and L respectively, arranging the three-dimensional data cubes of the baseband according to the distance unit, the airspace unit and the pulse index in sequence to obtain an observation vector
Figure BDA0002328013160000061
Next, a dimension reduction observation matrix is constructed. Suppose the number of pre-detected traces is IpdIthpdThe Doppler, space domain frequency and fast time domain difference frequency corresponding to the point trace are respectively
Figure BDA0002328013160000062
And
Figure BDA0002328013160000063
from this, the dimension reduction observation matrix can be obtained as:
Aκ＝[aκ(0)，...，aκ(ipd)，...，aκ(Ipd-1)]，(8)
wherein the content of the first and second substances,
Figure BDA0002328013160000064
symbol
Figure BDA0002328013160000065
Which represents the product of the Kronecker reaction,
Figure BDA0002328013160000066
and
Figure BDA0002328013160000067
respectively as follows:
Figure BDA0002328013160000068
Figure BDA0002328013160000069
Figure BDA00023280131600000610
and finally, establishing a dimension reduction observation model:
r＝Aκxκ+w，(13)
wherein the content of the first and second substances,
Figure BDA00023280131600000611
xκ(ipd) The complex amplitude of the target is represented by corresponding Doppler frequency, space domain frequency and fast time domain difference frequency
Figure BDA0002328013160000071
And
Figure BDA0002328013160000072
and 5: and (5) reconstructing the target. Based on known observation vector r and dimension reduction observation matrix AкReconstructing the target signal by adopting a generalized approximate message transfer algorithm, wherein the obtained signal reconstruction result is
Figure BDA0002328013160000073
Step 6: based on
Figure BDA0002328013160000074
The constant false alarm rate detection can be carried out by adopting a common constant false alarm rate detection methodThe method is carried out.
In the simulation experiments of fig. 3, 4 and 5, the effect of dimension reduction was verified. In order to compare the dimensionality Reduction (RD) -GAMP with the GAMP, the simulation is performed in a fast time domain as an example. Suppose the echo signal contains 5 targets, the signal-to-noise ratio is-10 dB, and the sampling frequency fsSame as the echo signal bandwidth B, fsAnd B is 1000Hz, and the number of sampling points N in the fast time domain is 1000.
FIG. 3 is a simulation result of target pre-detection before constructing a dimension reduction observation model. As can be seen from FIG. 3, in the simulation, the number of target pre-detected post-detection threshold traces is 21, which is much smaller than the number of original sample points after FFT. And constructing a dimension reduction observation model based on the target pre-detection result, so that the calculation amount of a subsequent reconstruction algorithm can be greatly reduced.
Fig. 4 and 5 illustrate the reconstruction effect of the GAMP and RD-GAMP algorithms in different grid point granularities. Fig. 4 shows the reconstruction results of the two algorithms in the case of the coarse grid, and fig. 5 shows the reconstruction results of the two algorithms in the case of the fine grid. From the reconstruction result, in the case of the coarse mesh, the reconstruction results of both algorithms have a large amplitude reconstruction error. In the case of fine meshes, the amplitude reconstruction error of the RD-GAMP algorithm is reduced, and the reconstruction effect of the GAMP algorithm is deteriorated. From the comparative description of the two figures, the RD-GAMP algorithm has better and more stable target reconstruction performance compared with the GAMP algorithm.
Fig. 6 is a Receiver Operating Characteristic (ROC) curve for a target detection method based on dimension reduction approximation messaging. In the simulation, the number L of antenna elements is 10, the number K of pulses is 20, and the number N of fast time sampling points is 500. For comparison, an ROC curve of a target detection method based on the traditional FFT method is given at the same time. The fast time domain and the slow time domain both adopt Chebyshev windows, the weighting depth is minus 40dB, and the space domain adopts Taylor windows. In each simulation, 20 targets are added randomly, and target angles, distances and Doppler are generated randomly for 200 times of simulation. As can be seen from the figure, the RD-AMP-based object detection method is significantly superior to the conventional FFT-based object detection method.
Fig. 7, 8, and 9 verify that the method of the present invention can be performed based on measured data. The range-doppler spectrum of a certain ground-based radar echo is shown in fig. 7, and known data contains an unmanned aerial vehicle target, the SNR is 18.1dB, and due to the high SNR of the cooperative target, in order to compare the performance of the conventional algorithm with the RD-GAMP algorithm, the following analysis method is adopted: 1) adding noise into the data, and evaluating the performance of a conventional method and an RD-GAMP method under different noise increment conditions based on a Monte Carlo method; 2) in addition to the known targets, other targets are used as false alarms, and the detection probability and the false alarm rate of the noise area are counted.
Performance of conventional method and RD-AMP method for example, as shown in fig. 8 and 9, the two methods compare detection performance at the same false alarm rate. In order to inhibit strong clutter side lobes, a Chebyshev window is adopted in a slow time domain in a conventional algorithm, and the weighting depth is-70 dB. As can be seen from FIG. 8, the false alarm rates of both algorithms are 6 × 10-5Left and right. Note that the false alarm rate is not constant when the noise rise is less than 10dB because there are other targets in the data, and when the noise rise is greater, the other targets are no longer detected and the false alarm rate tends to be constant. Under different noise increment conditions, the detection probability curves of the two are shown in fig. 9, and it can be seen that the RD-GAMP has better anti-noise capability, i.e. better performance, and when the detection probability is 0.5, the RD-GAMP has about 2dB performance gain compared with the conventional method. The reason is that the RD-GAMP is directly reconstructed based on the original data, and there is no loss of signal processing caused by conventional processing windowing.
According to the LFMCW radar target detection method based on the dimension reduction approximate message transfer, the dimension of a signal observation model is effectively reduced through a target pre-detection method, the calculated amount of an algorithm is reduced, and the reconstruction robustness is improved. Meanwhile, the signal is reconstructed by using a dimension reduction observation model and a generalized approximate message transfer algorithm, and the target detection is carried out, so that the detection performance better than that of the traditional target detection method based on fast Fourier transform can be realized.
It should be noted that the terms "upper", "lower", "left", "right", "front", "back", etc. used in the present invention are for clarity of description only, and are not intended to limit the scope of the present invention, and the relative relationship between the terms and the terms is not limited by the technical contents of the essential changes.
The above is only a preferred embodiment of the present invention, and the protection scope of the present invention is not limited to the above-mentioned embodiments, and all technical solutions belonging to the idea of the present invention belong to the protection scope of the present invention. It should be noted that modifications and embellishments within the scope of the invention may be made by those skilled in the art without departing from the principle of the invention.

Claims (6)

1. A LFMCW radar target detection method based on dimension reduction approximate message transfer is characterized by comprising the following steps:
1) the receiving end obtains a single-frame baseband data cube according to the received signal, and windowing is respectively carried out on the single-frame baseband data cube in a fast time domain, a slow time domain and a space domain;
2) for the windowed baseband data cube, realizing coherent accumulation of a target signal based on three-dimensional fast Fourier transform (3D-FFT) to obtain a three-dimensional frequency domain data cube;
3) performing target pre-detection on the three-dimensional frequency domain data cube to obtain pre-detected target trace information;
4) calculating a dimension reduction observation matrix based on the pre-detected target trace information, arranging the baseband data cubes according to vectors, and establishing a radar dimension reduction observation model;
5) reconstructing a target signal based on a radar dimension reduction observation model;
6) and carrying out constant false alarm detection by using the reconstructed target signal.
2. The LFMCW radar target detection method based on dimension reduction approximate message passing as claimed in claim 1, wherein: in the step 1), before the baseband data cube is obtained, the received signal is subjected to deskew processing, so that a difference frequency signal is obtained.
3. The LFMCW radar target detection method based on dimension reduction approximate message passing as claimed in claim 1, wherein: in the step 1), the window functions comprise a Chebyshev window and a Hamming window, and different window functions are respectively selected for a fast time domain, a slow time domain and a space domain.
4. The LFMCW radar target detection method based on dimension reduction approximate message passing as claimed in claim 1, wherein: in the step 3), performing three-dimensional target pre-detection on each unit for a three-dimensional frequency domain data cube obtained after fast Fourier transform; and determining the target pre-detection threshold according to the requirement of dimension reduction, and recording the corresponding fast time domain unit number, slow time domain unit number and airspace unit number of each trace point passing through the detection threshold.
5. The LFMCW radar target detection method based on dimension reduction approximate message passing as claimed in claim 1, wherein: in step 4), based on the pre-detected target trace information, calculating a dimensionality reduction observation matrix, specifically as follows:
401) the baseband data cube is represented in vector form: assuming that the distance unit, the pulse number and the antenna unit number are N, k and L respectively, sequentially arranging the baseband data cubes according to the distance unit, the airspace unit and the pulse index to obtain an observation vector
Figure FDA0002328013150000011
402) Constructing a dimension reduction observation matrix: assume that the number of target traces to be pre-detected is IpdIthpdThe Doppler, space domain frequency and fast time domain difference frequency corresponding to the point trace are respectively
Figure FDA0002328013150000012
And
Figure FDA0002328013150000013
then, the dimension reduction observation matrix is:
Aκ＝[aκ(0)，...，aκ(ipd)，...，aκ(Ipd-1)]，
wherein the content of the first and second substances,
Figure FDA0002328013150000021
symbol
Figure FDA0002328013150000022
Which represents the product of the Kronecker reaction,
Figure FDA0002328013150000023
and
Figure FDA0002328013150000024
respectively as follows:
Figure FDA0002328013150000025
Figure FDA0002328013150000026
Figure FDA0002328013150000027
403) establishing a radar dimension reduction observation model:
r＝Aκxκ+w
wherein the content of the first and second substances,
Figure FDA0002328013150000028
xκ(ipd) The complex amplitude of the target is represented by corresponding Doppler frequency, space domain frequency and fast time domain difference frequency
Figure FDA0002328013150000029
And
Figure FDA00023280131500000210
w tableA noise vector is shown.
6. The LFMCW radar target detection method based on dimension reduction approximate message passing as claimed in claim 1, wherein: in the step 5), a generalized approximate message transfer algorithm is used for reconstructing the target signal.
CN201911333074.6A 2019-12-20 2019-12-20 LFMCW radar target detection method based on dimension reduction approximate message transfer Pending CN111044996A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201911333074.6A CN111044996A (en) 2019-12-20 2019-12-20 LFMCW radar target detection method based on dimension reduction approximate message transfer

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201911333074.6A CN111044996A (en) 2019-12-20 2019-12-20 LFMCW radar target detection method based on dimension reduction approximate message transfer

Publications (1)

Publication Number Publication Date
CN111044996A true CN111044996A (en) 2020-04-21

Family

ID=70238430

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201911333074.6A Pending CN111044996A (en) 2019-12-20 2019-12-20 LFMCW radar target detection method based on dimension reduction approximate message transfer

Country Status (1)

Country Link
CN (1) CN111044996A (en)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111965615A (en) * 2020-07-28 2020-11-20 南京航空航天大学 Radar target detection method based on estimation before detection
CN114296039A (en) * 2021-12-01 2022-04-08 南京航空航天大学 LFMCW radar target constant false alarm detection method and device based on sparse reconstruction

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5440311A (en) * 1993-08-06 1995-08-08 Martin Marietta Corporation Complementary-sequence pulse radar with matched filtering and Doppler tolerant sidelobe suppression preceding Doppler filtering
FR2933497A1 (en) * 2008-07-03 2010-01-08 Claude Goutelard METHODS AND SYSTEMS FOR CODED EMISSION AND ANTENNA RECEPTION, IN PARTICULAR FOR RADAR
CN102707266A (en) * 2012-05-24 2012-10-03 北京理工大学 Radar with anti-interference and multi-target identification functions and detection method thereof
CN107229041A (en) * 2016-03-23 2017-10-03 西安中兴新软件有限责任公司 A kind of MIMO radar Wave arrival direction estimating method and device
US9791563B1 (en) * 2014-01-08 2017-10-17 National Technology & Engineering Solutions Of Sandia, Llc Joint synthetic aperture radar plus ground moving target indicator from single-channel radar using compressive sensing
CN109975807A (en) * 2019-03-27 2019-07-05 东南大学 A kind of reduced order subspace angle-measuring method suitable for millimeter wave trailer-mounted radar
CN110007282A (en) * 2019-03-15 2019-07-12 南京航空航天大学 Continuous wave system 1-bit radar target reconstruction problem dimension reduction method
CN110363151A (en) * 2019-07-16 2019-10-22 中国人民解放军海军航空大学 Based on the controllable radar target detection method of binary channels convolutional neural networks false-alarm

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US5440311A (en) * 1993-08-06 1995-08-08 Martin Marietta Corporation Complementary-sequence pulse radar with matched filtering and Doppler tolerant sidelobe suppression preceding Doppler filtering
FR2933497A1 (en) * 2008-07-03 2010-01-08 Claude Goutelard METHODS AND SYSTEMS FOR CODED EMISSION AND ANTENNA RECEPTION, IN PARTICULAR FOR RADAR
CN102707266A (en) * 2012-05-24 2012-10-03 北京理工大学 Radar with anti-interference and multi-target identification functions and detection method thereof
US9791563B1 (en) * 2014-01-08 2017-10-17 National Technology & Engineering Solutions Of Sandia, Llc Joint synthetic aperture radar plus ground moving target indicator from single-channel radar using compressive sensing
CN107229041A (en) * 2016-03-23 2017-10-03 西安中兴新软件有限责任公司 A kind of MIMO radar Wave arrival direction estimating method and device
CN110007282A (en) * 2019-03-15 2019-07-12 南京航空航天大学 Continuous wave system 1-bit radar target reconstruction problem dimension reduction method
CN109975807A (en) * 2019-03-27 2019-07-05 东南大学 A kind of reduced order subspace angle-measuring method suitable for millimeter wave trailer-mounted radar
CN110363151A (en) * 2019-07-16 2019-10-22 中国人民解放军海军航空大学 Based on the controllable radar target detection method of binary channels convolutional neural networks false-alarm

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
BENZHOU JIN; XIAOFEI ZHANG; BO ZHAO; GANG WU: "One-Bit LFM Pulse Radar: Harmonic Analysis and Target Reconstruction", 《IEEE ACCESS》 *

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111965615A (en) * 2020-07-28 2020-11-20 南京航空航天大学 Radar target detection method based on estimation before detection
CN114296039A (en) * 2021-12-01 2022-04-08 南京航空航天大学 LFMCW radar target constant false alarm detection method and device based on sparse reconstruction

Similar Documents

Publication Publication Date Title
CN110275158B (en) Broadband radar echo signal parameter estimation method based on Bayesian compressed sensing
CN106772253B (en) Radar clutter suppression method under non-uniform clutter environment
CN107450055B (en) High-speed maneuvering target detection method based on discrete linear frequency modulation Fourier transform
CN105807267A (en) MIMO radar extended target detection method
US8760340B2 (en) Processing radar return signals to detect targets
CN110007282B (en) Continuous wave system 1-bit radar target reconstruction problem dimension reduction method
CN112346030B (en) Super-resolution direction-of-arrival estimation method for unmanned aerial vehicle group
CN113238211B (en) Parameterized adaptive array signal detection method and system under interference condition
CN113050059A (en) Group target focusing super-resolution direction of arrival estimation method by using co-prime array radar
CN114609623B (en) Target detection method and device of monopulse radar and computer equipment
CN111045002A (en) Maneuvering target coherent accumulation method based on TRT and SNuFFT
CN111044996A (en) LFMCW radar target detection method based on dimension reduction approximate message transfer
CN111007487B (en) Multi-base radar target detection method based on time reversal
Xue et al. Knowledge-based target detection in compound Gaussian clutter with inverse Gaussian texture
CN107229040B (en) high-frequency radar target detection method based on sparse recovery space-time spectrum estimation
CN108845318B (en) Satellite-borne high-resolution wide-range imaging method based on Relax algorithm
CN112904280B (en) Transmitting and receiving combined polarization optimization method for time-sharing full-polarization radar system
US11269052B2 (en) Signal processing method
CN112014807A (en) Self-adaptive clutter suppression method for frequency agile radar
CN112255608A (en) Radar clutter self-adaptive suppression method based on orthogonal projection
CN113030928B (en) Polarization radar extended target self-adaptive detection method and system in non-uniform environment
CN114152918A (en) Anti-intermittent main lobe interference method based on compressed sensing
CN112835006A (en) Method and system for tracking radar small-target detection on sea based on interframe accumulation
CN113109776A (en) Angular flicker suppression method based on rearrangement time-frequency analysis
CN110907930A (en) Vehicle-mounted radar target detection and estimation method and device based on angle estimation

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
TA01 Transfer of patent application right

Effective date of registration: 20220111

Address after: Room 131, zone a, incubation building, innovation park, Baixia high tech park, No. 1, Guanghua Road, Qinhuai District, Nanjing, Jiangsu 210000

Applicant after: JIANGSU HUIWEIXUN INFORMATION TECHNOLOGY Co.,Ltd.

Address before: No.29, Jiangjun Avenue, Nanjing, Jiangsu Province, 210000

Applicant before: Nanjing University of Aeronautics and Astronautics

TA01 Transfer of patent application right
RJ01 Rejection of invention patent application after publication

Application publication date: 20200421

RJ01 Rejection of invention patent application after publication