CN112906476A - Airborne radar training sample selection method based on signal-to-noise-ratio loss - Google Patents

Airborne radar training sample selection method based on signal-to-noise-ratio loss Download PDF

Info

Publication number
CN112906476A
CN112906476A CN202110073198.6A CN202110073198A CN112906476A CN 112906476 A CN112906476 A CN 112906476A CN 202110073198 A CN202110073198 A CN 202110073198A CN 112906476 A CN112906476 A CN 112906476A
Authority
CN
China
Prior art keywords
aperture
sub
clutter
training sample
airborne radar
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.)
Granted
Application number
CN202110073198.6A
Other languages
Chinese (zh)
Other versions
CN112906476B (en
Inventor
李家烜
李明
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.)
University of Electronic Science and Technology of China
University of Electronic Science and Technology of China Zhongshan Institute
Original Assignee
University of Electronic Science and Technology of China
University of Electronic Science and Technology of China Zhongshan Institute
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 University of Electronic Science and Technology of China, University of Electronic Science and Technology of China Zhongshan Institute filed Critical University of Electronic Science and Technology of China
Priority to CN202110073198.6A priority Critical patent/CN112906476B/en
Publication of CN112906476A publication Critical patent/CN112906476A/en
Application granted granted Critical
Publication of CN112906476B publication Critical patent/CN112906476B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • 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
    • G01S7/36—Means for anti-jamming, e.g. ECCM, i.e. electronic counter-counter measures
    • 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
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
    • G06F17/10—Complex mathematical operations
    • G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F18/00—Pattern recognition
    • G06F18/20—Analysing
    • G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
    • G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
    • G—PHYSICS
    • G06—COMPUTING; CALCULATING OR COUNTING
    • G06F—ELECTRIC DIGITAL DATA PROCESSING
    • G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
    • G06F2218/02—Preprocessing
    • G06F2218/04—Denoising

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Data Mining & Analysis (AREA)
  • Mathematical Physics (AREA)
  • General Engineering & Computer Science (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Remote Sensing (AREA)
  • Pure & Applied Mathematics (AREA)
  • Mathematical Optimization (AREA)
  • Mathematical Analysis (AREA)
  • Artificial Intelligence (AREA)
  • Computational Mathematics (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Evolutionary Biology (AREA)
  • Evolutionary Computation (AREA)
  • Computing Systems (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Signal Processing (AREA)
  • Algebra (AREA)
  • Databases & Information Systems (AREA)
  • Software Systems (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention belongs to the field of airborne radar signal processing, and particularly provides a method for selecting an airborne radar training sample based on signal-to-noise ratio loss, which is used for eliminating the training sample inconsistent with the target distance ring clutter characteristic and ensuring the uniformity of the training sample used for clutter covariance matrix estimation in airborne radar space-time adaptive processing. Compared with the traditional method, the method has the advantages that the clutter characteristics of the target distance ring and the training sample are directly represented by using the sub-aperture covariance matrix, and the representation of the clutter characteristics is not influenced by other samples. In order to eliminate the influence of a target signal, the target distance ring possibly containing target signal components is eliminated in a Capon spectrum integral reconstruction-based mode, and compared with an orthogonal projection mode, clutter characteristics can be better reserved. The method adopts the output signal-to-noise ratio loss as the test statistic, and directly represents the approximation degree of the current sample and the CUT clutter characteristic.

Description

Airborne radar training sample selection method based on signal-to-noise-ratio loss
Technical Field
The invention belongs to the field of airborne radar signal processing, and particularly relates to an airborne radar training sample selection method which is used for eliminating training samples which are inconsistent with the characteristics of target distance ring clutter and ensuring the uniformity of training samples used for clutter covariance matrix estimation in airborne radar space-time adaptive processing.
Background
The space-time adaptive processing technology is an effective means for detecting a slow and small ground target under a strong clutter background by an airborne radar, and the key of the space-time adaptive processing lies in the estimation of a clutter covariance matrix of a target distance ring. However, in an actual application scenario, due to the influence of various non-ideal factors, such as a discrete strong scattering point, an interference target signal, intra-clutter motion, terrain variation, weather influence, and the like, the clutter characteristic of the training sample may be inconsistent with the target distance ring, and the estimation accuracy of the clutter covariance matrix of the target distance ring by using the training sample is reduced, so that the space-time adaptive processing performance is reduced, and the detection performance of the airborne radar on the ground target is deteriorated. Therefore, the original training samples are screened, the training samples which are inconsistent with the target distance ring clutter characteristic are removed, and the method has important significance for improving the detection performance of the airborne radar to the ground moving target.
At present, the existing airborne radar training sample selection method is mainly performed based on a generalized inner product, waveform similarity, terrain data and a sub-aperture covariance matrix. For example, in the document "Tang B, Tang J, Peng Y.detection of coherent samples based on loaded generated inner product method [ J ]. Digital Signal Processing,2012,22(4): 605-; the technique uses a sample covariance matrix calculated from the original training samples, but does not take into account clutter characteristics of the target range ring, when the clutter characteristics of most samples are not consistent with the target range ring, the selected samples will not have uniformity with the target range ring.
For example, in the document "Yifeng W, Tong W, Jianxin W, et al, robust tracking sampling selection al, terrestrial based on spectral analysis for space-time adaptive processing in noise interference environment [ J ]. radio resource & Navigation Iet,2015,9(7): 778-; the fundamental purpose of the Training sample Selection is to select the Training sample with the same covariance matrix as the target range ring, and the theoretical analysis in the document "Li H, Bao W, Hu J, et al.A Training sampling Selection Method Based on the System Identification for STAP [ J ]. Signal Processing,2017,142(jan.): 119-.
For example, in the document "Capraro C, Capraro G, Bradic I, et al, imaging digital residual data I n knowledge-aided space-time adaptive processing [ J ]. IEEE Transactions on Aerospace & electronic Systems,2006,42(3): 1080-; the method requires accurate matching of terrain data and radar echo signals, which is difficult to meet in practical application scenarios, and radar echo signals are affected by the irradiation angle under the same terrain, which is not considered at all in a method based on terrain data, so that the accuracy of screened samples can be seriously reduced.
For example, in the document "Wu Y, Wang T, Wu J, et al.conveying Sample Selection for Space-Time A significant Processing in Heterogeneous environment [ J ]. IEEE Geoscience and remove sensing letters,2014,12(4): 691) 695", a method for selecting a training Sample of an airborne radar based on a sub-aperture covariance matrix is disclosed, which eliminates a non-uniform training Sample by comparing the difference between a target distance ring and the sub-aperture covariance matrix of the training Sample, and the specific method is realized by subtracting two matrices to calculate a Frobenius norm; according to the method, in order to remove target components possibly existing in a target distance ring, orthogonal projection is carried out on target distance ring signals in an orthogonal subspace of a target guide vector, so that the target components are removed, the target guide vector cannot be orthogonal to a clutter subspace, clutter signal characteristics in the target distance ring are seriously damaged by adopting the orthogonal projection mode, so that the clutter characteristics of a screened training sample and the actual clutter characteristics of the target distance ring have larger deviation, and when the difference between the target distance ring and a training sample sub-aperture covariance matrix is measured by adopting a Frobenius norm, the Frobenius norm cannot accurately represent the approximation degree of the two covariance matrices, so that the screening efficiency is low.
Disclosure of Invention
The invention aims to provide a method for selecting training samples of an airborne radar based on signal-to-noise-ratio loss, aiming at the problems of the existing methods for selecting the training samples of the airborne radar; compared with the traditional method, the clutter suppression method has the advantages that the reconstructed sub-aperture covariance matrix is used for directly representing the clutter characteristics of the target distance ring and is more accurate, then the space-time adaptive filter designed by the sub-aperture covariance matrix of the sample distance ring is used for processing the target distance ring signal, the loss of the output signal-to-noise ratio is used as the test statistic, the clutter suppression capability of the selected sample on the target distance ring during space-time adaptive processing is directly measured, and the corresponding training sample is rejected when the value of the test statistic is lower than the screening threshold.
In order to achieve the purpose, the technical scheme adopted by the invention is as follows:
a method for selecting training samples of an airborne radar based on signal-to-noise-ratio loss is characterized by comprising the following steps:
step 1, estimating a target distance ring sub-aperture covariance matrix;
setting the number of array elements of the airborne radar as N and the number of transmitted pulses in the coherent processing interval as M, and obtaining a received signal X of a target distance ringCUT：
Figure BDA0002906665710000031
Mixing XCUTThe number of array elements divided into a series is N1The number of the transmitted pulses in the coherent processing interval is M1Sub-aperture signal of
Figure BDA0002906665710000032
Figure BDA0002906665710000033
Wherein the content of the first and second substances,
Figure BDA0002906665710000034
represents XCUTThe nth row and the mth column of elements;
estimating a target range ring sub-aperture covariance matrix using a sub-aperture smoothing technique
Figure BDA0002906665710000035
Figure BDA0002906665710000036
Step 2, estimating a target distance ring sub-aperture clutter covariance matrix;
defining a Capon integration region Ω according to the distribution of the clutter ridges in the space-time plane:
Ω＝{f:||f-f′||2≤ε,f′∈Π}，
wherein Π represents the set of points on the ridge of the clutter, and f' ∈ Π represents the clutterAny point on the ridge; epsilon is a preset constant:
Figure BDA0002906665710000037
then, the target range ring sub-aperture covariance matrix
Figure BDA0002906665710000038
The Capon spectrum of (a) is expressed as:
Figure BDA0002906665710000041
Figure BDA0002906665710000042
Figure BDA0002906665710000043
Figure BDA0002906665710000044
wherein f issRepresenting normalized spatial frequency, fdRepresents a normalized doppler frequency;
calculating a target distance ring sub-aperture clutter covariance matrix R0：
Figure BDA0002906665710000045
Step 3, estimating a sub-aperture covariance matrix of the training sample;
dividing data X (l) of the ith training sample by adopting a sub-aperture division mode which is the same as that of the target distance ring to obtain:
Figure BDA0002906665710000046
wherein the content of the first and second substances,
Figure BDA0002906665710000047
the nth row and mth column elements of X (l);
estimating a sub-aperture covariance matrix R (l) of the l training sample by using a sub-aperture smoothing technology;
step 4, calculating a weight vector of the space-time adaptive filter;
selecting a sub-aperture space-time guide vector outside any integral area as a target guide vector a, and calculating a weight vector w (l) of a space-time adaptive filter by using a sub-aperture covariance matrix of the ith training sample according to a minimum variance distortionless response criterion:
Figure BDA0002906665710000048
and 5, calculating the test statistic eta (l) of the ith training sample:
Figure BDA0002906665710000049
step 6, screening samples according to the test statistic eta (l);
and setting a screening threshold value mu, and rejecting the first training sample when the test statistic eta (l) of the first training sample is smaller than the screening threshold value mu.
Further, in step 2, the distribution of the clutter ridges in the space-time plane is:
according to the flight parameters of the aircraft: obtaining the normalized Doppler frequency f by the flight speed v, the yaw angle psi of the airborne radar and the flight direction of the airborne radar, the pitch angle theta of a specific counter point relative to the airborne platform and the observation direction beta of the airborne radardExpressed as:
Figure BDA0002906665710000051
obtaining a normalized spatial frequency fsExpressed as:
Figure BDA0002906665710000052
wherein, λ is the wavelength of the airborne radar, d is the array element spacing of the airborne radar, frThe pulse repetition frequency of the airborne radar;
by normalizing the spatial frequency fsAnd normalized Doppler frequency fdThe distribution of the clutter ridges in the spatio-temporal plane is determined.
Further, in step 6, the screening threshold μ is set to:
Figure BDA0002906665710000053
wherein k is a constant: k belongs to [0.1,0.01], and L is the total number of training samples.
The invention has the beneficial effects that:
the invention provides a method for selecting an airborne radar training sample based on signal-to-noise-ratio loss, which has the following advantages:
1) the clutter characteristics of the target distance ring are represented through the sub-aperture covariance matrix of the target distance ring, and compared with a mode that the clutter characteristics of the target distance ring are represented through a sample covariance matrix adopted based on a generalized inner product method, the method can avoid the influence of the heterogeneity of a training sample on the representation of the clutter characteristics of the target distance ring, and the representation result is only related to the target distance ring;
2) in order to eliminate target components possibly existing in the target distance ring sub-aperture covariance matrix, the invention reconstructs the target distance ring sub-aperture covariance matrix along a clutter ridge region in a space-time plane by using a Capon spectrum, and compared with an orthogonal projection mode adopted based on the sub-aperture covariance matrix, the Capon spectrum integral reconstruction mode adopted by the invention can better keep the clutter characteristic; according to the method, the distribution condition of the clutter ridges in the space-time plane is determined through the flight state parameters of the aircraft platform, and compared with a method based on terrain data, the method provided by the invention does not need to consider the problem of data matching, and is more easily suitable for actual working scenes;
3) the invention adopts signal-to-noise-ratio loss as test statistic to represent the strength of the space-time filter designed based on training sample data on the clutter suppression capability of a target distance ring, and the clutter suppression performance is better when the target distance ring is closer to the clutter characteristics of a training sample, namely the test statistic can represent the approximation degree of a target distance ring sub-aperture clutter covariance matrix and a training sample sub-aperture covariance matrix.
Drawings
Fig. 1 is a schematic flow chart of an airborne radar training sample selection method based on signal-to-noise-ratio loss in an embodiment of the present invention.
Fig. 2 is a schematic diagram of sub-aperture division of target distance ring data in the embodiment of the present invention.
Fig. 3 is a schematic diagram of a geometric model of an airborne platform of the airborne radar system in the embodiment of the invention.
FIG. 4 is a schematic diagram of the targets and clutter in the space-time plane in an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings and examples.
The invention provides a method for selecting an airborne radar training sample, which adopts output signal-to-noise-ratio loss as test statistic; in order to represent the clutter characteristic of the target distance ring, the method comprises the steps of firstly calculating a sub-aperture covariance matrix of the target distance ring, and performing Capon spectrum reconstruction in a space-time plane according to clutter ridge slope information (which can be estimated through flight state information) because target signal components may exist in the target distance ring to obtain the sub-aperture clutter covariance matrix of the target distance ring; then calculating a sub-aperture covariance matrix of a training sample, selecting a space-time guide vector outside an integral area in a space-time plane as a target guide vector, and designing a space-time adaptive filter according to a minimum variance distortionless response criterion; the filter designed by the training sample is used for processing the sub-aperture signal of the target distance ring, and the loss of the output signal-to-noise ratio is estimated to be used as the test statistic, so that the strength of the suppression capability of the selected sample on the target distance ring clutter when the selected sample is used for designing the space-time filter is directly represented, the output signal-to-noise ratio of the finally designed space-time filter is directly influenced, and the method is more direct and accurate.
In this embodiment, the following technical scheme is specifically adopted:
a method for selecting airborne radar training samples based on signal-to-noise-ratio loss is disclosed, the flow of which is shown in FIG. 1, and the method specifically comprises the following steps:
step 1, estimating a target distance ring sub-aperture covariance matrix;
assuming that the number of array elements of the airborne radar is N, the wavelength is lambda, the array element spacing is d, the number of transmitted pulses in the coherent processing interval is M, and the pulse repetition frequency is frThe number of training samples to be screened is L;
then, the received signal of the target range ring is
Figure BDA0002906665710000061
Figure BDA0002906665710000071
Dividing the array into a series of array elements with the number of N1The number of the transmitted pulses in the coherent processing interval is M1Sub-aperture signal of
Figure BDA0002906665710000072
Figure BDA0002906665710000073
Wherein the content of the first and second substances,
Figure BDA0002906665710000074
represents XCUTThe nth row and the mth column of elements; as shown in fig. 2;
estimating a target range ring sub-aperture covariance matrix using a sub-aperture smoothing technique
Figure BDA0002906665710000075
Comprises the following steps:
Figure BDA0002906665710000076
step 2, estimating a clutter covariance matrix of the target distance ring sub-aperture;
since the target range ring may contain the target signal, if directly utilized
Figure BDA0002906665710000077
The sub-aperture clutter covariance matrix representing the target distance ring is likely to have deviation, so the sub-aperture clutter covariance matrix of the target distance ring is estimated by using a Capon spectrum reconstruction mode
Figure BDA0002906665710000078
In a Capon spectrum in a space-time plane, determining the distribution of clutter ridges according to flight state parameters of the airplane, and finally performing integral reconstruction in a region near the clutter ridges to obtain a sub-aperture clutter covariance matrix of a target distance ring; the specific process is as follows:
the schematic diagram of the geometric model of the airborne platform of the airborne radar system is shown in FIG. 3, wherein the airplane flies along the direction of an x axis, the flying speed is v, the height of the airborne platform is H, and the array is parallel to an XOY plane; in the figure, P represents an inverse point on the ground, theta and
Figure BDA0002906665710000079
respectively representing the pitch angle and the azimuth angle of the array relative to the plane of the aircraft, and psi representing the yaw angle of the array relative to the flight direction of the aircraft; defining beta as the radar observation direction, can obtain：
Figure BDA00029066657100000710
Then, the normalized doppler frequency at point P can be expressed as:
Figure BDA00029066657100000711
the normalized spatial frequency at point P may be represented as:
Figure BDA0002906665710000081
for any point on the ground, it can be represented by β as a variable, according to fsAnd fdThe distribution condition of the clutter ridges in the space-time plane can be determined by the expression of (2);
when psi is 0°In the time, the airborne radar adopts a front side view working mode, so that the distribution of clutter ridges in the space-time plane is a straight line, the target and clutter distribution is shown in fig. 4, and the slope of the clutter ridges can be expressed as:
Figure BDA0002906665710000082
and the set of points on the ridge of the clutter is represented as pi, then any point on the ridge of the clutter is represented as f' ∈ pi, and the area of Capon integral reconstruction is defined as Ω:
Ω＝{f:||f-f′||2≤ε,f′∈Π}，
wherein epsilon is a predetermined constant, and is used for determining the range of integration, and is taken
Figure BDA0002906665710000083
f represents any point within the integration region Ω;
then the process of the first step is carried out,
Figure BDA0002906665710000084
the Capon spectrum of (a) can be expressed as:
Figure BDA0002906665710000085
wherein the content of the first and second substances,
Figure BDA0002906665710000086
representing space-time steering vectors, fsRepresenting normalized spatial frequency, fdRepresents a normalized doppler frequency; a-1Represents a matrix inversion operation, ·HWhich represents the transpose of the conjugate,
Figure BDA0002906665710000087
represents the Kronecker product;
Figure BDA0002906665710000088
Figure BDA0002906665710000089
the target range ring sub-aperture clutter covariance matrix based on Capon spectral reconstruction can be expressed as:
Figure BDA00029066657100000810
for calculation, the integration area can be divided into grid points, and then the integration area is uniformly divided into Q points by using summation instead of integration, and Q > N1M1Generally, Q is 100N1M1Then the reconstructed target range ring sub-aperture clutter covariance matrix can be expressed as:
Figure BDA00029066657100000811
wherein, s (f)si,fdi) E Ω, i is 1,2, …, and Q represents a space-time steering vector corresponding to the discretization grid point selected in the integration area;
step 3, estimating a sub-aperture covariance matrix of the training sample;
l training sample data are 1,2
Figure BDA0002906665710000091
Figure BDA0002906665710000092
Dividing training sample data X (l) by adopting a sub-aperture division mode which is the same as that of the target distance ring to obtain:
Figure BDA0002906665710000093
wherein the content of the first and second substances,
Figure BDA0002906665710000094
the nth row and mth column elements of X (l);
then the sub-aperture covariance matrix for the ith training sample is estimated using a sub-aperture smoothing technique as:
Figure BDA0002906665710000095
step 4, calculating a weight vector of the space-time adaptive filter;
selecting a sub-aperture space-time guide vector outside an integral area as a target guide vector a, and calculating a weight vector w (l) of a space-time adaptive filter by using a sub-aperture covariance matrix of the ith training sample according to a minimum variance distortionless response criterion;
then, according to the minimum variance undistorted response criterion, the solution expression of the weight vector w (l) of the space-time adaptive filter calculated by using the subaperture covariance matrix of the ith training sample is:
Figure BDA0002906665710000096
the following can be obtained:
Figure BDA0002906665710000097
step 5, calculating the test statistic eta (l) of the ith training sample;
processing the target range ring subaperture signal by using a filter with weight vector w (l), estimating the loss of the output signal-to-noise ratio, and taking the loss as a test statistic eta (l) corresponding to the ith training sample, wherein the covariance matrix of the target range ring clutter signal is represented by the covariance matrix of the target range ring subaperture clutter estimated in the step 2, namely
Figure BDA0002906665710000101
Step 6, screening samples according to the test statistic eta (l);
because the test statistic of the invention is the signal-to-noise-ratio loss of the filter with weight vector w (l) to the target distance ring data processing output, the strength of the first training sample used for the space-time adaptive filter design to the target distance ring clutter suppression capability is directly represented, and the greater the eta (l) value is, the stronger the target distance ring clutter suppression capability is; therefore, when the training samples are screened, after a threshold value is set, the training samples lower than the threshold value are removed; one method that can be used for setting the threshold value in this embodiment is:
Figure BDA0002906665710000102
wherein k is [0.1,0.01 ].
While the invention has been described with reference to specific embodiments, any feature disclosed in this specification may be replaced by alternative features serving the same, equivalent or similar purpose, unless expressly stated otherwise; all of the disclosed features, or all of the method or process steps, may be combined in any combination, except mutually exclusive features and/or steps.

Claims (3)

1. A method for selecting training samples of an airborne radar based on signal-to-noise-ratio loss is characterized by comprising the following steps:
step 1, estimating a target distance ring sub-aperture covariance matrix;
setting the number of array elements of the airborne radar as N and the number of transmitted pulses in the coherent processing interval as M, and obtaining a received signal X of a target distance ringCUT：
Figure FDA0002906665700000011
Mixing XCUTThe number of array elements divided into a series is N1The number of the transmitted pulses in the coherent processing interval is M1Sub-aperture signal of
Figure FDA0002906665700000012
Figure FDA0002906665700000013
Wherein the content of the first and second substances,
Figure FDA0002906665700000014
represents XCUTThe nth row and the mth column of elements;
estimating a target range ring sub-aperture covariance matrix using a sub-aperture smoothing technique
Figure FDA0002906665700000015
Figure FDA0002906665700000016
Step 2, estimating a target distance ring sub-aperture clutter covariance matrix;
defining a Capon integration region Ω according to the distribution of the clutter ridges in the space-time plane:
Ω＝{f:||f-f′||2≤ε,f′∈Π}，
wherein pi represents a set of points on the heterowave ridge, and f' epsilon represents any point on the heterowave ridge; epsilon is a preset constant:
Figure FDA0002906665700000017
then, the target range ring sub-aperture covariance matrix
Figure FDA0002906665700000018
The Capon spectrum of (a) is expressed as:
Figure FDA0002906665700000019
Figure FDA00029066657000000110
Figure FDA00029066657000000111
Figure FDA00029066657000000112
wherein f issRepresenting normalized spatial frequency, fdRepresents a normalized doppler frequency;
calculating a target distance ring sub-aperture clutter covariance matrix R0：
Figure FDA0002906665700000021
Step 3, estimating a sub-aperture covariance matrix of the training sample;
dividing data X (l) of the ith training sample by adopting a sub-aperture division mode which is the same as that of the target distance ring to obtain:
Figure FDA0002906665700000022
wherein the content of the first and second substances,
Figure FDA0002906665700000023
the nth row and mth column elements of X (l);
estimating a sub-aperture covariance matrix R (l) of the l training sample by using a sub-aperture smoothing technology;
step 4, calculating a weight vector of the space-time adaptive filter;
selecting a sub-aperture space-time guide vector outside any integral area as a target guide vector a, and calculating a weight vector w (l) of a space-time adaptive filter by using a sub-aperture covariance matrix of the ith training sample according to a minimum variance distortionless response criterion:
Figure FDA0002906665700000024
and 5, calculating the test statistic eta (l) of the ith training sample:
Figure FDA0002906665700000025
step 6, screening samples according to the test statistic eta (l);
and setting a screening threshold value mu, and rejecting the first training sample when the test statistic eta (l) of the first training sample is smaller than the screening threshold value mu.
2. The method for selecting training samples of airborne radar based on signal-to-noise ratio loss according to claim 1, wherein in the step 2, the distribution of the clutter ridges in the space-time plane is as follows:
according to the flight parameters of the aircraft: obtaining the normalized Doppler frequency f by the flight speed v, the yaw angle psi of the airborne radar and the flight direction of the airborne radar, the pitch angle theta of a specific counter point relative to the airborne platform and the observation direction beta of the airborne radardExpressed as:
Figure FDA0002906665700000026
obtaining a normalized spatial frequency fsExpressed as:
Figure FDA0002906665700000031
wherein, λ is the wavelength of the airborne radar, d is the array element spacing of the airborne radar, frThe pulse repetition frequency of the airborne radar;
by normalizing the spatial frequency fsAnd normalized Doppler frequency fdThe distribution of the clutter ridges in the spatio-temporal plane is determined.
3. The method for selecting training samples of airborne radar based on signal-to-noise-ratio loss according to claim 1, wherein in the step 6, the screening threshold μ is set as:
Figure FDA0002906665700000032
wherein k is a constant: k belongs to [0.1,0.01], and L is the total number of training samples.
CN202110073198.6A 2021-01-20 2021-01-20 Airborne radar training sample selection method based on signal-to-noise-ratio loss Active CN112906476B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110073198.6A CN112906476B (en) 2021-01-20 2021-01-20 Airborne radar training sample selection method based on signal-to-noise-ratio loss

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110073198.6A CN112906476B (en) 2021-01-20 2021-01-20 Airborne radar training sample selection method based on signal-to-noise-ratio loss

Publications (2)

Publication Number Publication Date
CN112906476A true CN112906476A (en) 2021-06-04
CN112906476B CN112906476B (en) 2022-11-04

Family

ID=76116369

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110073198.6A Active CN112906476B (en) 2021-01-20 2021-01-20 Airborne radar training sample selection method based on signal-to-noise-ratio loss

Country Status (1)

Country Link
CN (1) CN112906476B (en)

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114895262A (en) * 2022-04-08 2022-08-12 西安电子科技大学 Clutter suppression method based on space-time spectrum characteristics and orthogonal projection of wind turbine
CN116047462A (en) * 2023-03-31 2023-05-02 中国人民解放军空军预警学院 Method and device for selecting optimal array element number and array element spacing of end-shooting array airborne radar
CN114895262B (en) * 2022-04-08 2024-07-09 西安电子科技大学 Clutter suppression method based on space-time spectrum characteristics and orthogonal projection of wind motor

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101819269A (en) * 2010-03-19 2010-09-01 清华大学 Space-time adaptive processing method under non-homogeneous clutter environment
US20120249361A1 (en) * 2011-04-04 2012-10-04 Zafer Sahinoglu Method for Detecting Targets Using Space-Time Adaptive Processing
CN103018727A (en) * 2011-09-27 2013-04-03 中国科学院电子学研究所 Sample-training-based non-stationary clutter suppression method of vehicle-mounted radar
CN105785339A (en) * 2016-03-21 2016-07-20 西安电子科技大学 Airborne radar clutter covariance matrix estimation method in inhomogeneous clutter environment
CN107180259A (en) * 2017-05-15 2017-09-19 电子科技大学 A kind of STAP training sample selection methods based on System Discrimination
CN107664759A (en) * 2017-06-12 2018-02-06 南京邮电大学 It is a kind of that excellent airborne radar moving target detection method is selected based on sparse nonuniform sample
CN110618402A (en) * 2019-08-22 2019-12-27 西安空间无线电技术研究所 Auxiliary clutter sample screening method based on sub-aperture smoothing
CN111695087A (en) * 2020-05-27 2020-09-22 上海微波技术研究所（中国电子科技集团公司第五十研究所） Method and system for detecting composite mine detection target

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN101819269A (en) * 2010-03-19 2010-09-01 清华大学 Space-time adaptive processing method under non-homogeneous clutter environment
US20120249361A1 (en) * 2011-04-04 2012-10-04 Zafer Sahinoglu Method for Detecting Targets Using Space-Time Adaptive Processing
CN103018727A (en) * 2011-09-27 2013-04-03 中国科学院电子学研究所 Sample-training-based non-stationary clutter suppression method of vehicle-mounted radar
CN105785339A (en) * 2016-03-21 2016-07-20 西安电子科技大学 Airborne radar clutter covariance matrix estimation method in inhomogeneous clutter environment
CN107180259A (en) * 2017-05-15 2017-09-19 电子科技大学 A kind of STAP training sample selection methods based on System Discrimination
CN107664759A (en) * 2017-06-12 2018-02-06 南京邮电大学 It is a kind of that excellent airborne radar moving target detection method is selected based on sparse nonuniform sample
CN110618402A (en) * 2019-08-22 2019-12-27 西安空间无线电技术研究所 Auxiliary clutter sample screening method based on sub-aperture smoothing
CN111695087A (en) * 2020-05-27 2020-09-22 上海微波技术研究所（中国电子科技集团公司第五十研究所） Method and system for detecting composite mine detection target

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
龚清勇，朱兆达: ""非均匀环境下雷达动目标检测的空时自适应处理"", 《应用科学学报》 *

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN114895262A (en) * 2022-04-08 2022-08-12 西安电子科技大学 Clutter suppression method based on space-time spectrum characteristics and orthogonal projection of wind turbine
CN114895262B (en) * 2022-04-08 2024-07-09 西安电子科技大学 Clutter suppression method based on space-time spectrum characteristics and orthogonal projection of wind motor
CN116047462A (en) * 2023-03-31 2023-05-02 中国人民解放军空军预警学院 Method and device for selecting optimal array element number and array element spacing of end-shooting array airborne radar

Also Published As

Publication number Publication date
CN112906476B (en) 2022-11-04

Similar Documents

Publication Publication Date Title
CN108051809B (en) Moving target imaging method and device based on Radon transformation and electronic equipment
US7064702B1 (en) System, method and computer program product for reducing quadratic phase errors in synthetic aperture radar signals
US7916068B2 (en) Generalized inner product method and apparatus for improved detection and discrimination
Capraro et al. Implementing digital terrain data in knowledge-aided space-time adaptive processing
US7652617B2 (en) Radar microsensor for detection, tracking, and classification
CN108761419B (en) Low-altitude wind shear wind speed estimation method based on self-adaptive processing of combined space-time main channel
CN111352102A (en) Multi-target number detection method and device based on frequency modulation continuous wave radar
US6226321B1 (en) Multichannel parametric adaptive matched filter receiver
EP1485729B1 (en) System and method for target signature calculation and recognition
WO2014066360A1 (en) Sensor system and method for determining target location using sparsity-based processing
CN109901162B (en) Long-baseline imaging STAP method suitable for distributed geosynchronous orbit SAR
CN111401168B (en) Multilayer radar feature extraction and selection method for unmanned aerial vehicle
CN110554391A (en) low-altitude wind shear wind speed estimation method based on DDD-GMB
Sakar et al. Analysis of geometrical approximations in signal reconstruction methods for multistatic SAR constellations with large along-track baseline
Kim et al. Extrapolation-RELAX estimator based on spectrum partitioning for DOA estimation of FMCW radar
CN106970358B (en) Optimization method for angular Doppler registration of clutter spectrum of non-normal side-looking array radar
CN112906476B (en) Airborne radar training sample selection method based on signal-to-noise-ratio loss
CN103760540B (en) Based on moving target detect and the method for parameter estimation of reconstruction signal and 1-norm
CN113466797B (en) Double-base SAR space-time clutter suppression method based on clutter ridge matching sparse recovery
US7046192B2 (en) Radar process for classifying or identifying helicopters
CN113093187A (en) CSAR ground moving target tracking method without road information assistance
CN109581366B (en) Discrete sidelobe clutter identification method based on target steering vector mismatch
Zhang et al. A novel clutter covariance matrix estimation method based on feature subspace for space-based early warning radar
CN114966674A (en) Staring radar low-altitude target height measurement method based on transmitted waveform information
CN112835034B (en) Dual-channel radar ground height measurement system and method

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
GR01 Patent grant
GR01 Patent grant