CN112835005B - Micro Doppler feature extraction method based on super-resolution target tracking - Google Patents

Micro Doppler feature extraction method based on super-resolution target tracking Download PDF

Info

Publication number
CN112835005B
CN112835005B CN202011644573.XA CN202011644573A CN112835005B CN 112835005 B CN112835005 B CN 112835005B CN 202011644573 A CN202011644573 A CN 202011644573A CN 112835005 B CN112835005 B CN 112835005B
Authority
CN
China
Prior art keywords
micro
target
motion
frequency
doppler
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.)
Active
Application number
CN202011644573.XA
Other languages
Chinese (zh)
Other versions
CN112835005A (en
Inventor
胡文
赵月
熊清
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 Yunhefeng Intelligent Technology Co ltd
Original Assignee
Jiangsu Yunhefeng Intelligent Technology Co ltd
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 Jiangsu Yunhefeng Intelligent Technology Co ltd filed Critical Jiangsu Yunhefeng Intelligent Technology Co ltd
Priority to CN202011644573.XA priority Critical patent/CN112835005B/en
Publication of CN112835005A publication Critical patent/CN112835005A/en
Application granted granted Critical
Publication of CN112835005B publication Critical patent/CN112835005B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

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
    • 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
    • G01S7/418—Theoretical aspects

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Physics & Mathematics (AREA)
  • General Physics & Mathematics (AREA)
  • Radar, Positioning & Navigation (AREA)
  • Remote Sensing (AREA)
  • Radar Systems Or Details Thereof (AREA)

Abstract

The invention discloses a micro Doppler feature extraction method based on super-resolution target tracking, which comprises the following steps of S1: obtaining a broadband radar echo, ranging in a time domain by adopting a super-resolution algorithm, and performing super-resolution processing on a target one-dimensional range profile; s2: estimating the instantaneous Doppler frequency of the target by a time-frequency analysis method so as to obtain the instantaneous speed, compensating the translational speed of the target, extracting the micro Doppler frequency shift component, and calculating the micro speed; s3: introducing the micro-motion speed extracted in the S2 into a target tracking flow, constructing pseudo measurement, and performing sequential filtering to obtain a more accurate micro-motion track; s4: and acquiring a target micro-motion track, and updating the micro-motion parameters to be estimated. The method has the advantages that the target tracking algorithm is combined with the micro Doppler feature extraction process, the target tracking precision is improved by utilizing the Doppler measurement information, and the high-quality micro motion curve is obtained, so that more accurate micro motion parameter estimation is obtained.

Description

Micro Doppler feature extraction method based on super-resolution target tracking
Technical Field
The invention relates to the field of radar signal processing, in particular to a micro Doppler feature extraction method based on super-resolution target tracking.
Background
The radar transmits electromagnetic signals to a moving target, the target moves radially relative to the radar, and the radar echo generates frequency shift for a long time; in addition to body movement, if there is a micro-motion of the target or any structural component of the target in the direction of the radar line of sight, such micro-motion will cause additional frequency modulation on the echo signal and produce a side frequency near the doppler shift frequency produced by body movement, which is defined by v.c.chen as the micro-doppler effect.
The problem of extraction of micro Doppler features has been a lot of research results in recent years, however, the existing micro Doppler feature extraction method only analyzes the micro motion features in the frequency domain, extracts the micro Doppler curve based on the time-frequency diagram, and ignores the time domain features of the target micro motion. And the micro-motion amplitude of the target is difficult to accurately estimate in a time domain because the micro-motion of the target is weak compared with the translation of the target main body, so that the micro-motion amplitude becomes a difficult point in research.
Disclosure of Invention
In order to solve the problems, the technical scheme adopted by the invention is as follows:
a micro Doppler feature extraction method based on super-resolution target tracking comprises the following steps:
the method comprises the steps of firstly, obtaining broadband radar target echo, sampling pulses in the target echo, obtaining one-dimensional distance direction data among a plurality of pulses, and constructing a distance dimensional covariance matrix R according to the one-dimensional distance direction data among the pulses r ；
Step two, the distance dimension covariance matrix R r Decomposing the characteristic value, selecting the selected characteristic value according to the strategy of selecting the characteristic value, and forming the noise subspace G by the characteristic vector corresponding to the selected characteristic value noise Setting a predetermined signal vector r (tau) i ) Will be associated with said noise subspace G according to a spectral peak search strategy noise Defining the signal vector with the minimum matching degree of the inner characteristic vector as a jogging signal vector;
thirdly, obtaining the echo delay corresponding to the micro-motion signal vector, wherein the relationship between the echo delay and the radial distance of the target scattering point is as follows:
Figure GDA0003848596250000021
calculating distance information of a scattering point of a target according to the formula I and the echo delay of the micro-motion signal vector, and generating a first target micro-motion track according to the distance information;
fourthly, performing time-frequency transformation processing on the micro-motion signal vector obtained in the second step to calculate micro-motion speed information of the target scattering point;
the fourth step comprises the generation of translation compensation frequency f according to a translation frequency estimation strategy u Calculating the corresponding instantaneous frequency f according to the micro-motion signal vector m And superposing the instantaneous frequency and the translational compensation frequency to generate a jogging frequency f d I.e. f d ＝f m +f u According to a formula, two pairs of the inching frequencies f d Performing time-frequency transformation processing to calculate the micro-motion speed information of the scattering point of the target,
Figure GDA0003848596250000022
and calculating micro-motion speed information of the scattering point of the target through time-frequency transformation, introducing the micro-Doppler speed obtained by extraction into a first target micro-motion track, further accurately estimating the micro-motion track of the target, introducing the micro-Doppler speed into Kalman filtering, constructing pseudo measurement, performing sequential filtering, further accurately estimating the position of the target, introducing a sequential filtering algorithm into a micro-Doppler feature extraction process, and obtaining a high-quality micro-motion curve while improving the tracking accuracy of the target by using Doppler measurement information so as to obtain more accurate micro-motion parameter estimation.
And fifthly, performing parameter optimization on the first target micro-motion track according to the micro-motion speed information of the target scattering point and generating a second micro-motion track.
Preferably, the distance dimensional covariance matrix R r ＝E{X r (X r ) H In which E { X } r (X r ) H Is a pair of X r (X r ) H The statistical average is calculated and the average is calculated,
the eigenvalue selection strategy comprises maintaining the distance dimension covariance matrix R r When the first eigenvalue is more than 10 times of the l-th eigenvalue, namely the third equation is satisfied, the l-th and the following eigenvalues are taken as the selected eigenvalues,
Figure GDA0003848596250000031
preferably, the proposed signal vector
Figure GDA0003848596250000032
The spectral peak search strategy comprises constructing a pseudo spectrum P r (τ i )，
Figure GDA0003848596250000033
Calculating the pseudo-spectrum P r (τ i ) And the characteristic vector corresponding to the peak value is the inching signal vector.
Preferably, the fourth step further includes superimposing distance units in which the plurality of inching signal vectors are located to obtain a one-dimensional time sequence z (u), and performing time-frequency transformation on the one-dimensional time sequence z (u):
Figure GDA0003848596250000034
where g (t) is a window function, STFT z The frequency corresponding to the maximum value point in the (t, f) is the instantaneous frequency f of the target scattering point m 。
Preferably, the translational frequency estimation strategy comprises calculating the translational compensation frequency f according to a formula u ，
Figure GDA0003848596250000041
III, wherein f d max Is STFT z Frequency, f, corresponding to the maximum point in (t, f) d min Is STFT z And (t, f) the frequency corresponding to the maximum and minimum point.
Preferably, the distance information of the scattering point of the target calculated in the third step includes a distance measurement value
Figure GDA0003848596250000042
And angle measurement
Figure GDA0003848596250000043
The third step comprises:
(a) Measuring the distance
Figure GDA0003848596250000044
And angle measurement
Figure GDA0003848596250000045
Conversion to fluteUnder the karr coordinate system:
Figure GDA0003848596250000046
Figure GDA0003848596250000047
(b) Structure false measurement
Figure GDA0003848596250000048
Figure GDA0003848596250000049
Defining measurement vectors
Figure GDA00038485962500000410
Figure GDA00038485962500000411
Wherein the content of the first and second substances,
Figure GDA00038485962500000412
Figure GDA00038485962500000413
velocity components in the x and y directions respectively,
Figure GDA00038485962500000414
is eta k The switching error of (a) is determined,
Figure GDA00038485962500000415
to measure noise, h k Is a measurement function;
(c) Calculating a measurement vector
Figure GDA00038485962500000416
Measuring a mean value and a covariance matrix of conversion errors;
(d) Measuring the vector
Figure GDA0003848596250000051
And said pseudo-metric
Figure GDA0003848596250000052
Decorrelation;
(e) For the measurement vector
Figure GDA0003848596250000053
And (3) performing tracking filtering:
(i) Calculating a target state predicted value and calculating a predicted covariance;
(ii) Calculating a Kalman gain;
(iii) Updating the target state;
(f) Measuring the false
Figure GDA0003848596250000054
Performing extended Kalman filtering estimation;
(g) Acquiring a target micro-motion track, and updating the micro-motion amplitude estimation:
Figure GDA0003848596250000055
the method has the advantages that the distance dimension covariance matrix R is constructed for the corresponding one-dimensional distance direction data between the pulses by constructing the distance-slow time matrix r And decomposing the characteristic values, taking the characteristic vectors corresponding to smaller characteristic values to form a noise subspace, constructing a pseudo spectrum, matching the signal vectors with the noise characteristic vectors, wherein the smaller the matching degree is, the greater the separation degree is, and the smaller the correlation is, reconstructing the micro-motion signal vectors by the method, finding the echo delay corresponding to the micro-motion signal vectors, calculating the distance information of the target scattering point, generating a first target micro-motion track according to the distance information, and performing preliminary micro-motion track tracking on the target scattering point.
The invention calculates the micro-motion speed information of the scattering point of the target through time-frequency transformation, and introduces the micro-Doppler speed obtained by extraction into the first target micro-motion track to further refine the micro-motion track of the target.
According to the method, the sequential filtering algorithm is introduced into the micro Doppler feature extraction process, and the high-quality micro motion curve is obtained while the tracking precision of the target is improved by using the Doppler measurement information, so that more accurate micro motion parameter estimation is obtained.
Drawings
Fig. 1 is a schematic flow diagram of a micro doppler feature extraction method based on super-resolution ranging target tracking according to the present invention;
fig. 2 is a schematic diagram of a pulse echo model in a micro doppler feature extraction method based on super-resolution ranging target tracking provided by the invention.
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
As shown in fig. 1, a micro doppler feature extraction method based on super resolution ranging target tracking provided by the embodiment of the present invention includes the following steps:
a micro Doppler feature extraction method based on super-resolution target tracking comprises the following steps:
the method comprises the steps of firstly, obtaining broadband radar target echo, sampling pulses in the target echo, obtaining one-dimensional distance direction data among a plurality of pulses, and constructing a distance dimensional covariance matrix R according to the one-dimensional distance direction data among the pulses r ；
Step two, for the distance dimension covariance matrix R r Decomposing the characteristic value, selecting the selected characteristic value according to the strategy selected by the characteristic value, and forming a noise element by the characteristic vector corresponding to the selected characteristic valueSpace G noise Setting a predetermined signal vector r (tau) i ) Will be associated with said noise subspace G according to a spectral peak search strategy noise Defining the signal vector with the minimum matching degree of the inner characteristic vector as a jogging signal vector;
thirdly, obtaining the echo delay corresponding to the micro-motion signal vector, wherein the relationship between the echo delay and the radial distance of the target scattering point is as follows:
Figure GDA0003848596250000061
and calculating distance information of a scattering point of the target according to the formula I and the echo delay of the micro-motion signal vector, and generating a first target micro-motion track according to the distance information. Constructing a distance dimension covariance matrix R for the corresponding one-dimensional distance direction data between pulses by constructing a distance-slow time matrix r And decomposing the characteristic values, taking the characteristic vectors corresponding to smaller characteristic values to form a noise subspace, constructing a pseudo spectrum, matching the signal vectors with the noise characteristic vectors, wherein the smaller the matching degree is, the greater the separation degree is, and the smaller the correlation is, reconstructing the micro-motion signal vectors by the method, finding the echo delay corresponding to the micro-motion signal vectors, calculating the distance information of the target scattering point, generating a first target micro-motion track according to the distance information, and performing preliminary micro-motion track tracking on the target scattering point.
Preferably, the method further comprises the following steps:
fourthly, performing time-frequency transformation processing on the micro-motion signal vector obtained in the second step to calculate micro-motion speed information of the target scattering point;
and fifthly, performing parameter optimization on the first target micro-motion track according to the micro-motion speed information of the target scattering point and generating a second micro-motion track. And calculating the micro-motion speed information of the scattering point of the target through time-frequency transformation, and introducing the extracted micro-Doppler speed into the first target micro-motion track to further refine the micro-motion track of the target.
Preferably, said step four comprisesGenerating translation compensation frequency f according to translation frequency estimation strategy u Calculating the corresponding instantaneous frequency f according to the inching signal vector m And superposing the instantaneous frequency and the translational compensation frequency to generate a jogging frequency f d I.e. f d ＝f m +f u According to a formula, two pairs of the inching frequencies f d Performing time-frequency transformation processing to calculate the micro-motion speed information of the scattering point of the target,
Figure GDA0003848596250000071
preferably, the distance dimensional covariance matrix R r ＝E{X r (X r ) H In which E { X } r (X r ) H Is a pair of X r (X r ) H The statistical average is calculated and the average is calculated,
the eigenvalue selection strategy comprises maintaining the distance dimension covariance matrix R r When the first eigenvalue is more than 10 times of the ith eigenvalue, namely the formula III is satisfied, the first and the following eigenvalues are taken as selected eigenvalues,
Figure GDA0003848596250000081
preferably, the proposed signal vector
Figure GDA0003848596250000082
The spectral peak search strategy comprises constructing a pseudo spectrum P r (τ i )，
Figure GDA0003848596250000083
Calculating the pseudo-spectrum P r (τ i ) The feature vector corresponding to the peak value is the inching signal vector.
Preferably, the fourth step further includes superimposing distance units in which the plurality of inching signal vectors are located to obtain a one-dimensional time sequence z (u), and performing time-frequency transformation on the one-dimensional time sequence z (u):
Figure GDA0003848596250000084
where g (t) is a window function, STFT z The frequency corresponding to the maximum value point in (t, f) is the instantaneous frequency f of the target scattering point m 。
Preferably, the translational frequency estimation strategy comprises calculating the translational compensation frequency f according to a formula u ，
Figure GDA0003848596250000091
III, wherein f d max Is STFT z Frequency, f, corresponding to the maximum point in (t, f) d min Is STFT z And (t, f) the frequency corresponding to the maximum and minimum value points.
Preferably, the distance information of the scattering point of the target calculated in the third step includes a distance measurement value
Figure GDA0003848596250000092
And angle measurement
Figure GDA0003848596250000093
The third step comprises:
(a) Measuring the distance
Figure GDA0003848596250000094
And angle measurement
Figure GDA0003848596250000095
Conversion to cartesian coordinates:
Figure GDA0003848596250000096
Figure GDA0003848596250000097
(b) Structure false measurement
Figure GDA0003848596250000098
Figure GDA0003848596250000099
Defining measurement vectors
Figure GDA00038485962500000910
Figure GDA00038485962500000911
Wherein, the first and the second end of the pipe are connected with each other,
Figure GDA00038485962500000912
Figure GDA00038485962500000913
velocity components in the x and y directions respectively,
Figure GDA00038485962500000914
is eta of k The switching error of (a) is determined,
Figure GDA00038485962500000915
to measure noise, h k Is a measurement function;
(c) Calculating a measurement vector
Figure GDA00038485962500000916
Measuring a mean value and a covariance matrix of conversion errors;
(d) Measuring the vector
Figure GDA00038485962500000917
And said pseudo-metric
Figure GDA00038485962500000918
Decorrelation;
(e) For the measurement vector
Figure GDA00038485962500000919
And (3) performing tracking filtering:
(i) Calculating a target state predicted value and calculating a predicted covariance;
(ii) Calculating a Kalman gain;
(iii) Updating the target state;
(f) Measuring the false
Figure GDA0003848596250000101
Performing extended Kalman filtering estimation;
(g) Acquiring a target micro-motion track, and updating the micro-motion amplitude estimation:
Figure GDA0003848596250000102
as shown in fig. 2, the second embodiment of the present invention is different from the first embodiment in that a detailed procedure for implementing the contents of the first embodiment is provided.
Acquiring a broadband radar echo, measuring distance in a time domain by adopting an MUSIC algorithm, and performing super-resolution processing on a target one-dimensional range profile;
(a) Broadband radar LFM monopulse emission waveform:
Figure GDA0003848596250000103
wherein T is the pulse repetition period, f 0 K is the chirp rate for the transmit frequency.
And (3) target echo model:
Figure GDA0003848596250000104
where tau is the echo time delay and,
Figure GDA0003848596250000105
r0 is the target relative radar radial distance,
and (3) dechirping frequency modulation:
Figure GDA0003848596250000106
each pulse is sampled, the intra-pulse sampling is fast time sampling, and the inter-pulse sampling is slow time sampling.
(b) Constructing a distance dimension covariance matrix R for one-dimensional distance direction data corresponding to each slow time r ，
R r ＝E{X r (X r ) H }
Wherein, E { X r (X r ) H Is a pair X r (X r ) H Statistical averaging, where time averaging may be substituted.
(c) To R is r Decomposing the characteristic value, and taking the characteristic vector corresponding to the smaller characteristic value to form a noise subspace G noise Here, the following method is adopted to determine the size of the noise subspace, the eigenvalues are arranged from large to small, when the first eigenvalue is more than 10 times of the l-th eigenvalue, the signal subspace is considered to be the eigenvector corresponding to the first l-1 eigenvalues, and the noise subspace is considered to be the eigenvector corresponding to the remaining eigenvalues:
Figure GDA0003848596250000111
and (3) searching a pseudo spectrum:
Figure GDA0003848596250000112
where N is the number of time sampling points
Wherein
Figure GDA0003848596250000113
As a signal vector, with tau i As independent variable, pseudo-spectrum P r (τ i ) For dependent variables, search for P r (τ i ) Peak value, τ corresponding to peak value i For echo time delay of scattering points of the object, r (tau) correspondingly i ) And then, obtaining the precise distance of the scattering point of the target by the reconstructed signal vector, wherein the relation between the echo delay and the radial distance of the scattering point of the target is as follows:
Figure GDA0003848596250000121
estimating instantaneous Doppler frequency through time-frequency transformation to obtain instantaneous speed, compensating the translational speed of the target, extracting a micromotion Doppler frequency shift component, and calculating micromotion speed;
(a) Superposing the distance units where the targets are located to obtain a one-dimensional time sequence z (u), and performing time-frequency transformation on the one-dimensional time sequence z (u):
Figure GDA0003848596250000122
where g (t) is a window function, STFT z Each element in (t, f) represents the component amplitude of the signal at time t, fourier transformed at frequency f;
STFT z and f corresponding to the maximum value point in (t, f) is the instantaneous frequency of the target scattering point.
For a discrete signal applied in practice:
Figure GDA0003848596250000123
(b) Extracting a Doppler curve from the time-frequency distribution diagram by adopting a W-V peak value detection method, and searching a frequency point with the maximum amplitude at each moment in a time-frequency plane as an instantaneous Doppler frequency shift corresponding to each moment:
Figure GDA0003848596250000124
(c) Calculating the instantaneous velocity by the instantaneous doppler shift:
Figure GDA0003848596250000125
λ is the wavelength of the emitted waveform
Wherein:
f d ＝f m +f u
f m is the target inching frequency, f u The target micro-motion component is in translation frequency, and the micro-motion amplitude is symmetrical relative to the target translation main body, so that the following conditions exist:
Figure GDA0003848596250000131
step three, introducing the micro Doppler velocity extracted in the step two into Kalman filtering, constructing pseudo measurement, carrying out sequential filtering, and further accurately positioning the target:
(a) Because the radar measurement and the target motion state are completely nonlinear, and the tracking estimation effect of the target by using nonlinear filtering is poor, the invention converts the position (the slant distance and the azimuth angle) measurement under a polar coordinate system into a Cartesian coordinate system:
Figure GDA0003848596250000132
Figure GDA0003848596250000133
wherein
Figure GDA0003848596250000134
The measured value of the distance is the measured value,
Figure GDA0003848596250000135
the angle measurement is taken.
(b) Introducing the Doppler measurements into a measurement equation, constructing pseudo-measurements to reduce the strong non-linearity between the Doppler measurements and the motion state of the target
Figure GDA0003848596250000136
Figure GDA0003848596250000137
Figure GDA0003848596250000138
Wherein the content of the first and second substances,
Figure GDA0003848596250000139
the velocity components in the x and y directions respectively,
Figure GDA00038485962500001310
is eta of k The switching error of (a) is determined,
Figure GDA00038485962500001311
in order to measure the vector, the measurement vector,
Figure GDA00038485962500001312
to measure noise, h k Is a measurement function and
Figure GDA00038485962500001313
(c) Calculating a mean and covariance matrix of the measurement transformation errors:
in an actual scenario, the true doppler velocity information and position cannot be known, and generally, in the case of measurement value suppression, the mathematical expectation is obtained for the true mean and covariance:
Figure GDA0003848596250000141
Figure GDA0003848596250000142
wherein the values of the elements are as follows:
Figure GDA0003848596250000143
Figure GDA0003848596250000144
Figure GDA0003848596250000145
Figure GDA0003848596250000146
Figure GDA0003848596250000147
Figure GDA0003848596250000148
Figure GDA0003848596250000149
Figure GDA00038485962500001410
Figure GDA00038485962500001411
wherein r is k As true value of distance, θ k For the real value of the azimuth angle,
Figure GDA00038485962500001412
is a measured value of the distance, and is,
Figure GDA00038485962500001413
the measured value of the speed is the measured value of the speed,
Figure GDA00038485962500001414
is a measure of the azimuth angle for which,
Figure GDA00038485962500001415
the difference is distance, azimuth angle, speed measurement error variance, rho is the correlation coefficient of the distance measurement error and the speed measurement error.
(d) The conversion errors of the pseudo-measurements and the position information are statistically correlated, so for sequential filtering, the correlation between them must be removed first, and the position and pseudo-measurements are de-correlated.
The covariance R k，α By position and pseudo-measurement
Figure GDA0003848596250000151
Two parts are written in blocks:
Figure GDA0003848596250000152
wherein, the first and the second end of the pipe are connected with each other,
Figure GDA0003848596250000153
Figure GDA0003848596250000154
order to
Figure GDA0003848596250000155
To obtain
Figure GDA0003848596250000156
And
Figure GDA0003848596250000157
and ordering:
Figure GDA0003848596250000158
in the equation
Figure GDA0003848596250000159
The left and right sides are respectively multiplied by B k From the Cholesky decomposition of the matrix, we can derive:
Figure GDA00038485962500001510
Figure GDA00038485962500001511
wherein the content of the first and second substances,
Figure GDA00038485962500001512
has a mean value of
Figure GDA00038485962500001513
Variance of
Figure GDA00038485962500001514
Figure GDA00038485962500001515
And
Figure GDA00038485962500001516
converting the measurement, and obtaining:
Figure GDA0003848596250000161
Figure GDA0003848596250000162
Figure GDA0003848596250000163
Figure GDA0003848596250000164
and are and
Figure GDA0003848596250000165
not correlated, the superscript epsilon indicates the quantity correlated with the pseudo-metrology.
(e) Position measurement tracking filtering:
calculating a target state predicted value, and simultaneously calculating a predicted covariance:
Figure GDA0003848596250000166
P k|k-1 ＝F k-1 ·P k-1|k-1 ·F′ k-1 +Q k-1
calculating a Kalman gain:
Figure GDA0003848596250000167
and (3) updating the target state:
Figure GDA0003848596250000168
and (3) covariance updating:
Figure GDA0003848596250000169
wherein, F k-1 For motion state transition matrix, G k-1 To control the matrix, u k-1 To control the vector, V k-1 Is white gaussian noise, and is a noise,
Figure GDA00038485962500001610
for the measurement matrix, P k|k-1 In order to predict the error correlation matrix, the error correlation matrix is,
Figure GDA00038485962500001611
in order to estimate the error correlation matrix, the error correlation matrix is estimated,
Figure GDA00038485962500001612
is the kalman gain.
(f) Filtering estimation of pseudo measurement:
Figure GDA00038485962500001613
Figure GDA0003848596250000171
Figure GDA0003848596250000172
wherein, the first and the second end of the pipe are connected with each other,
Figure GDA0003848596250000173
in that
Figure GDA0003848596250000174
The resulting jacobian matrix, namely:
Figure GDA0003848596250000175
Figure GDA0003848596250000176
is composed of
Figure GDA0003848596250000177
The second derivative of (a) constitutes:
Figure GDA0003848596250000178
by using
Figure GDA0003848596250000179
(1,2) to represent a position estimation error covariance matrix
Figure GDA00038485962500001710
The elements of the first row and the second column, and the like. A. The k Comprises the following steps:
Figure GDA00038485962500001711
step four, acquiring a target micro-motion track, and updating the micro-motion parameter estimation:
Figure GDA00038485962500001712
the working principle is as follows: the invention constructs a distance dimension covariance matrix R for corresponding one-dimensional distance direction data between pulses by constructing a distance-slow time matrix r Decomposing the characteristic values, taking the characteristic vectors corresponding to smaller characteristic values to form a noise subspace, constructing a pseudo spectrum, matching the signal vectors with the noise characteristic vectors, wherein the smaller the matching degree is, the greater the separation degree is, and the smaller the correlation is, reconstructing the micro-motion signal vectors by the method, finding the echo delay corresponding to the micro-motion signal vectors, calculating the distance information of the target scattering point, thereby realizing super-resolution ranging, generating a first target micro-motion track according to the distance information, and performing preliminary micro-motion track following on the target scattering pointTracing. And calculating the micro-motion speed information of the scattering point of the target through time-frequency transformation, and introducing the extracted micro-Doppler speed into the first target micro-motion track to further refine the micro-motion track of the target. And the micro Doppler velocity is introduced into Kalman filtering, pseudo measurement is constructed, sequential filtering is carried out, and the position of the target is further accurate.
The above description is only a preferred embodiment of the present invention, and it should be noted that, for those skilled in the art, several modifications and variations can be made without departing from the technical principle of the present invention, and these modifications and variations should also be regarded as the protection scope of the present invention.

Claims (7)

1. A micro Doppler feature extraction method based on super-resolution target tracking is characterized by comprising the following steps:
the method comprises the steps of firstly, obtaining broadband radar target echo, sampling pulses in the target echo, obtaining one-dimensional distance direction data among a plurality of pulses, and constructing a distance dimensional covariance matrix R according to the one-dimensional distance direction data among the pulses r ；
Step two, the distance dimension covariance matrix R r Decomposing the characteristic value, selecting the selected characteristic value according to the strategy of selecting the characteristic value, and forming the noise subspace G by the characteristic vector corresponding to the selected characteristic value noise Setting a predetermined signal vector r (tau) i ) Will be associated with said noise subspace G according to a spectral peak search strategy noise Defining the signal vector with the minimum matching degree of the inner characteristic vector as a jogging signal vector;
thirdly, obtaining the echo delay corresponding to the micro-motion signal vector, wherein the relationship between the echo delay and the radial distance of the target scattering point is as follows:
Figure FDA0003848596240000011
wherein: tau is echo time delay;
calculating distance information of a target scattering point according to the formula I and the echo delay of the micro-motion signal vector, and generating a first target micro-motion track according to the distance information;
fourthly, performing time-frequency transformation processing on the micro-motion signal vector obtained in the second step, and calculating micro-motion speed information of the target scattering point;
the fourth step comprises the generation of translation compensation frequency f according to a translation frequency estimation strategy u Calculating the corresponding instantaneous frequency f according to the micro-motion signal vector m And superposing the instantaneous frequency and the translational compensation frequency to generate a jogging frequency f d I.e. f d ＝f m +f u According to a formula, two pairs of the inching frequencies f d Performing time-frequency transformation processing to calculate the micro-motion speed information of the scattering point of the target,
Figure FDA0003848596240000012
wherein: λ is the emission waveform wavelength;
and calculating the micro-motion speed information of the scattering point of the target through time-frequency transformation, introducing the extracted micro-Doppler speed into a first target micro-motion track, further accurately measuring the micro-motion track of the target, introducing the micro-Doppler speed into Kalman filtering, constructing pseudo measurement, performing sequential filtering, further accurately measuring the position of the target, introducing a sequential filtering algorithm into the micro-Doppler feature extraction process, and obtaining a high-quality micro-motion curve while improving the tracking precision of the target by using the Doppler measurement information so as to obtain more accurate micro-motion parameter estimation.
2. The micro-doppler feature extraction method for super-resolution ranging target tracking according to claim 1, further comprising:
and fifthly, performing parameter optimization on the first target micro-motion track according to the micro-motion speed information of the target scattering point and generating a second micro-motion track.
3. The method of micro-Doppler feature extraction for super-resolution ranging target tracking according to claim 1, wherein the distance dimensional covariance matrix
R r ＝E{X r (X r ) H In which E { X } r (X r ) H Is a pair of X r (X r ) H The statistical average is calculated and the average is calculated,
the eigenvalue selection strategy comprises the distance dimension covariance matrix R r When the first eigenvalue is more than 10 times of the ith eigenvalue, namely the formula III is satisfied, the first and the following eigenvalues are taken as selected eigenvalues,
Figure FDA0003848596240000021
4. the micro-Doppler feature extraction method for super-resolution ranging target tracking according to claim 1, wherein the proposed signal vector
Figure FDA0003848596240000022
The spectral peak search strategy includes constructing a pseudo-spectrum P r (τ i )，
Figure FDA0003848596240000023
Calculating the pseudo-spectrum P r (τ i ) The eigenvector corresponding to the peak value is the inching signal vector;
wherein: k is the frequency modulation slope; t is the pulse repetition period.
5. The method for extracting micro-doppler features for tracking over-resolution ranging targets according to claim 2, wherein the fourth step further comprises stacking distance units where a plurality of micro-motion signal vectors are located to obtain a one-dimensional time sequence z (u), and performing time-frequency transformation on the one-dimensional time sequence z (u):
Figure FDA0003848596240000031
where g (t) is a window function, STFT z The frequency corresponding to the maximum value point in (t, f) is the instantaneous frequency f of the target scattering point m 。
6. The micro-Doppler feature extraction method for super-resolution ranging target tracking according to claim 5, wherein the translational frequency estimation strategy comprises calculating a translational compensation frequency f according to a formula u ，
Figure FDA0003848596240000032
III, wherein f dmax Is STFT z Frequency, f, corresponding to the maximum point in (t, f) dmin Is STFT z And (t, f) the frequency corresponding to the maximum and minimum value points.
7. The method for extracting micro-Doppler features for super-resolution ranging target tracking according to claim 1, wherein the distance information of the scattering points of the target calculated in the third step comprises distance measurement values
Figure FDA0003848596240000033
And angle measurement
Figure FDA0003848596240000034
The third step comprises:
(a) Measuring the distance
Figure FDA0003848596240000035
And angle measurement
Figure FDA0003848596240000036
Conversion to cartesian coordinates:
Figure FDA0003848596240000037
Figure FDA0003848596240000038
(b) Structure false measurement
Figure FDA0003848596240000039
Figure FDA00038485962400000310
Defining a measurement vector
Figure FDA0003848596240000041
Figure FDA0003848596240000042
Wherein the content of the first and second substances,
Figure FDA0003848596240000043
Figure FDA0003848596240000044
Figure FDA0003848596240000045
respectively velocity components in the x and y directions,
Figure FDA0003848596240000046
is eta k The error of the conversion of (2) is,
Figure FDA0003848596240000047
to measure noise, h k For the purpose of the measurement function,
Figure FDA0003848596240000048
is a velocity measurement;
(c) Calculating a measurement vector
Figure FDA0003848596240000049
Measuring a mean value and a covariance matrix of conversion errors;
(d) Measuring the vector
Figure FDA00038485962400000410
And said pseudo-metric
Figure FDA00038485962400000411
Decorrelation;
(e) For the measurement vector
Figure FDA00038485962400000412
And (3) performing tracking filtering:
(i) Calculating a target state predicted value and calculating a predicted covariance;
(ii) Calculating a Kalman gain;
(iii) Updating the target state;
(f) Measure the said fake
Figure FDA00038485962400000413
Performing extended Kalman filtering estimation;
(g) Acquiring a target micro-motion track, and updating the micro-motion amplitude estimation:
Figure FDA00038485962400000414
CN202011644573.XA 2020-12-31 2020-12-31 Micro Doppler feature extraction method based on super-resolution target tracking Active CN112835005B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202011644573.XA CN112835005B (en) 2020-12-31 2020-12-31 Micro Doppler feature extraction method based on super-resolution target tracking

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202011644573.XA CN112835005B (en) 2020-12-31 2020-12-31 Micro Doppler feature extraction method based on super-resolution target tracking

Publications (2)

Publication Number Publication Date
CN112835005A CN112835005A (en) 2021-05-25
CN112835005B true CN112835005B (en) 2022-11-11

Family

ID=75927167

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202011644573.XA Active CN112835005B (en) 2020-12-31 2020-12-31 Micro Doppler feature extraction method based on super-resolution target tracking

Country Status (1)

Country Link
CN (1) CN112835005B (en)

Families Citing this family (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113985384B (en) * 2021-12-28 2022-05-24 中南大学 Spatial target translation compensation method and device based on template matching
CN115219990A (en) * 2022-07-13 2022-10-21 陕西黄河集团有限公司 Feature extraction method of one-dimensional range profile
CN116679301B (en) * 2023-07-28 2023-10-20 西安电子科技大学 Method for rapidly reconstructing target range profile of broadband radar in super-resolution mode

Family Cites Families (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
GB869808A (en) * 1958-12-01 1961-06-07 Marconi Wireless Telegraph Co Improvements in or relating to radar systems
CN105652272B (en) * 2015-12-31 2018-03-30 西北工业大学 High-resolution imaging method is tieed up using the distance of signal discrete frequency component
CN110501696B (en) * 2019-06-28 2022-05-31 电子科技大学 Radar target tracking method based on Doppler measurement adaptive processing
CN111443335B (en) * 2020-03-17 2022-03-25 中山大学 Method, system and device for estimating target micro-motion parameters of broadband radar and storage medium

Also Published As

Publication number Publication date
CN112835005A (en) 2021-05-25

Similar Documents

Publication Publication Date Title
CN112835005B (en) Micro Doppler feature extraction method based on super-resolution target tracking
CN107132534B (en) Optimization method for high-speed radar target frequency domain detection
CN106646447B (en) Radar target long time integration detection method based on linear frequency modulation continuous wave
CN101408616B (en) Inverse synthetic aperture radar imaging distance aligning method applicable to low signal-noise ratio data
CN109188385B (en) Method for detecting high-speed weak target under clutter background
KR101628154B1 (en) Multiple target tracking method using received signal strengths
CN111551922B (en) Three-dimensional space double/multi-base radar high-speed target detection method
Huang et al. Radar high speed small target detection based on keystone transform and linear canonical transform
CN101738606A (en) Method for detecting coherent integration of radar target based on generalized Doppler filter bank
CN110133654B (en) High-orbit satellite SAR moving target detection method
CN109324322A (en) A kind of direction finding and target identification method based on passive phased array antenna
CN111007503B (en) Moving target focusing and positioning method and system based on frequency spectrum accurate positioning
JP5044358B2 (en) Radar equipment
CN109901162B (en) Long-baseline imaging STAP method suitable for distributed geosynchronous orbit SAR
CN106597437A (en) High-frequency airborne FM continuous wave SAR motion compensation and imaging processing method
CN109613507B (en) Detection method for high-order maneuvering target radar echo
CN111856466A (en) Efficient ISAR (inverse synthetic aperture radar) translation compensation method for complex moving target
CN115034993A (en) Method and device for defocusing removal processing of moving target image and storage medium
CN109782249B (en) Two-target correlation time delay estimation algorithm
Page et al. Detection and tracking of moving vehicles with Gotcha radar systems
CN109507654B (en) LS-based phase information calculation method in complex environment
CN112415512B (en) SAR moving target focusing method based on advance and retreat method and golden section method
Zhu et al. Robust moving targets detection and velocity estimation using multi-channel and multi-look SAR images
CN106772373A (en) For the SAR imaging methods of any ground moving object
Ding et al. A novel iterative inner-pulse integration target detection method for bistatic radar

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