CN109932716A - A kind of low target micro-Doppler feature extracting method - Google Patents
A kind of low target micro-Doppler feature extracting method Download PDFInfo
- Publication number
- CN109932716A CN109932716A CN201910173828.XA CN201910173828A CN109932716A CN 109932716 A CN109932716 A CN 109932716A CN 201910173828 A CN201910173828 A CN 201910173828A CN 109932716 A CN109932716 A CN 109932716A
- Authority
- CN
- China
- Prior art keywords
- target
- matrix
- signal
- sparse
- micro
- 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
Links
- 238000000034 method Methods 0.000 title claims abstract description 33
- 239000013598 vector Substances 0.000 claims abstract description 69
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 7
- 239000011159 matrix material Substances 0.000 claims description 53
- 238000000605 extraction Methods 0.000 claims description 34
- 230000005855 radiation Effects 0.000 claims description 29
- 238000004891 communication Methods 0.000 claims description 12
- 238000005070 sampling Methods 0.000 claims description 7
- 238000010276 construction Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 6
- 230000008569 process Effects 0.000 claims description 5
- 230000008859 change Effects 0.000 claims description 3
- 238000006243 chemical reaction Methods 0.000 claims description 3
- 238000005286 illumination Methods 0.000 claims description 3
- 238000007781 pre-processing Methods 0.000 claims description 3
- 230000001629 suppression Effects 0.000 claims description 3
- 230000009466 transformation Effects 0.000 claims description 3
- 230000008447 perception Effects 0.000 claims description 2
- 239000000284 extract Substances 0.000 abstract description 4
- 238000012360 testing method Methods 0.000 description 11
- 238000010586 diagram Methods 0.000 description 4
- 238000004458 analytical method Methods 0.000 description 3
- 230000008901 benefit Effects 0.000 description 3
- 238000000354 decomposition reaction Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000007123 defense Effects 0.000 description 2
- 238000002592 echocardiography Methods 0.000 description 2
- 230000035515 penetration Effects 0.000 description 2
- 238000000342 Monte Carlo simulation Methods 0.000 description 1
- 230000003471 anti-radiation Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000005094 computer simulation Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 230000007774 longterm Effects 0.000 description 1
- 238000010295 mobile communication Methods 0.000 description 1
- 238000012544 monitoring process Methods 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012549 training Methods 0.000 description 1
- 238000012795 verification Methods 0.000 description 1
Landscapes
- Radar Systems Or Details Thereof (AREA)
Abstract
A kind of low target micro-Doppler feature extracting method is disclosed, including the following steps: the reception and pretreatment of the target echo signal through external sort algorithm irradiation;Target fine motion frequency is extracted and constructs redundant observation vector model;Construct target micro-Doppler feature Parameter Dictionary;The micro-Doppler feature of low target extracts.Under the premise of land clutter and other clutters can be filtered out well, the present invention can realize that good micro-Doppler feature extracts to the low target under the conditions of Gaussian white noise channel.
Description
Technical Field
The invention relates to an external radiation source radar signal processing technology, in particular to a low-altitude target micro Doppler feature extraction method.
Background
At present, low-altitude areas are gradually opened, a large number of general aircrafts, especially unmanned aerial vehicles, enter the low-altitude field, the war electromagnetic environment is increasingly complex, and great pressure is caused to air defense early warning and air training in important areas. The external radiation source radar has good four-resistance (anti-radiation missile resistance, interference resistance, stealth resistance and low altitude penetration resistance). Especially, the civil communication signal-based external radiation source radar can make full use of the existing mobile communication facilities to construct a large-range passive detection network and has strong survivability in war environments. In order to maintain the national ownership and the state-of-the-earth defense space, the micro Doppler feature of the low-altitude target is effectively extracted, so that the low-altitude and ultra-low-altitude penetration target can be effectively tracked and imaged, the attribute, intention and threat degree of the low-altitude target are judged, and the method is one of the key factors for realizing more effective low-altitude monitoring and obtaining the war initiative.
The currently common micro-doppler feature extraction methods include: the time-frequency analysis is combined with a Hough Transform (HT) method, and the method is a common micro Doppler feature extraction method due to the characteristics of simplicity, effectiveness and the like. However, the extraction precision is influenced by the size of the radar echo data matrix, the window length of time-frequency analysis and the resolution of time-frequency analysis images. An orthogonal matching pursuit method is an efficient feature extraction technology proposed by y.c. Pati in 1993. The method is based on a matching tracking algorithm, atoms in a dictionary are subjected to orthogonalization processing according to a Schmidt orthogonalization method, and signals are decomposed on an overcomplete orthogonal basis to realize iterative decomposition of the signals. The method has high decomposition efficiency, high convergence rate, simplicity and feasibility, and is still applicable under the undersampling condition. However, the method has the greatest defects that when the extracted characteristic parameters are increased, the number of atoms in the dictionary is greatly increased, and the calculation amount of algorithm decomposition is increased in a geometric series manner.
Disclosure of Invention
Aiming at the problems in the prior art, the invention provides a low-altitude target micro Doppler feature extraction method, which comprises the following steps:
the first step is as follows: reception and preprocessing of target echo signals illuminated by external radiation sources
The external radiation source transmits LTE communication signals, after the LTE communication signals irradiate a target, the signals reflected by the target are called target echo signals, and after the receiver receives the signals, the receiver samples and receives the signals and outputs received signals;
the communication signal output by the external radiation source is
Wherein N is the subcarrier frequency number, K is the code length, N is the index of the subcarrier frequency, K is the index of the code length, anIs the weighting factor of the nth sub-carrier frequency, AreIs a real part modulation coefficient, AimFor imaginary modulation factor, TpFor Orthogonal Frequency Division Multiplexing (OFDM) symbol duration, f0Is the carrier frequency, and Δ f is the sub-carrier frequency interval; wherein Δ f satisfies the OFDM orthogonality condition
Wherein, i, j is the number of the sub-carrier frequency, which means that any two sub-carrier frequencies satisfy the orthogonality condition, that is
Δf=1/Tp(3)
The received target echo signal is
Wherein, tmIs a slow time, tkFor fast time, σ is the scattering coefficient of the target, r (t)m) The sum of the distances between the target and the external radiation source and the distance between the target and the receiver are respectively, and c is the speed of light in vacuum;
the reference signal is a direct wave signal from a receiver to an external radiation source and can be expressed as
Wherein D is the distance between the external radiation source and the receiver;
assuming that the multipath clutter in the received signal is better suppressed and the clean target echo and direct wave are obtained, the echo signal is conjugate multiplied by the reference signal to obtain the baseband echo which can be expressed as
Wherein,is s isr(tk,tm) The conjugate signal of (a);
the second step is that: extracting target micro-motion frequency and constructing multiple observation vector model
(a) Multiple observation vector model
Let there be a one-dimensional vector x ═ x (1), …, x (U)]TCoordinates α under the U-dimensional orthonormal basis psi only contain B nonzero elements, if B < U, then α is called B sparse, wherein x (1), …, x (U) are coordinates under a conventional coordinate system, psi is called sparse basis matrix of x, and the existence of observation matrix phi irrelevant to psi is setV×U(V < U) Linear Observation of vector xThe observation result can be recorded as y phi x phi psi α theta α, wherein theta phi psi is called a perception matrix, the original N-dimensional signal vector x is reduced to an M-dimensional vector y, and when the constraint isometry condition is met, the sparse representation of the signal vector x can be obtained by solving the minimum l0Norm resolution, i.e.
α=arg min||α||0(8)
Where α is a sparse representation of the vector x, α satisfies the condition y ═ Φ Ψ α;
consider that if there are L sources with the same sparse structure, let X ═ X1,…,xL]TX is a number of L X, i.e. X1,…,xLThe formed matrix can adopt the same observation matrix, and the observation result can be expressed as
Y=ΦX=Φ[x1,…,xL](9)
The model is called a multiple observation vector MMV model, and the multiplicity is L; solving is carried out by sparse matrix rank minimization by using the prior information that signals have the same sparse structure, namely the following equation sets are combined
Wherein A is a sparse representation of X, αlRepresents the coordinate of the l-th source under the U-dimensional orthonormal base and satisfies the condition of Y-phi psi [ α1,…,αL];
(b) Construction of multiple observation vector model
Aiming at the baseband echo signal output in the first step, if the time is tmAnd a fast time tkRespectively sampling M 'and N', the echo signal can be expressed in a matrix form Sr∈RN′×M′Wherein R represents a real number matrix; due to the change of the target attitude, the observed echo matrix can not be directly used as a multiple observation vector Sr(ii) a Due to the fact thatHere, the fast time t 'of the same target posture is extracted'sThe baseband echo of (1) constructs a multiple observation vector model with the same sparse structure, so that the sparse expression of the multiple observation vector model can be solved;
because the echo at a certain fast time needs to be extracted, the prior information of the target micro-motion frequency is needed; extracting a target micro-motion frequency f by utilizing sinusoidal frequency modulation Fourier Bessel transformation; constructing a multiple observation vector model of the baseband echo by using the extracted micro-motion frequency f ', wherein f' is the estimation result of the extracted target micro-motion frequency
S(f′)=[sb(t′s1,t′m1),sb(t′s2,t′m2),…,sb(t′sN′,t′mN′)](11)
Wherein T ═ 1/f 'is the target inching period, T'siIs a marked fast time instant, t'miIs a fast time t'siCorresponding slow time series, i.e. t'mi=t′si,t′si+T,...,t′si+LT,The number of occurrences of the same attitude at the radar illumination time t,denotes rounding down, sb(t′si,t′mi) Is represented by t'si,t′miA baseband echo signal that is a parameter, i 1, 2.., N';
the third step: constructing a target micro Doppler feature parameter dictionary
For a specific external radiation source signal, a received signal in practical application contains a target echo signal, a direct wave signal and other clutter, wherein the clutter signal is filtered by an existing clutter suppression method;
for the preprocessed target echo signal, a multiple observation vector model is constructed according to the second step, the orthogonality of OFDM signals is utilized to extract the echo of a certain sub-carrier frequency for processing, and then the carrier frequency of the target in the baseband echo is a known quantity; assuming that the number of signal amplitudes and phases in the sparse dictionary are P terms and Q terms respectively, the sparse dictionary can be expressed as { A }1,…,APAnd { theta }1,…,θQF 'is used as a parameter, and the sparse dictionary Ψ (f') can be expressed as a matrix
Wherein the sparse dictionary matrix Ψ (f ') has a size of M' × PQ,for extracted slow time instants, row vectors in the matrixA feature vector representing the target micro-doppler, where m is 1, …, L, and
wherein,
the fourth step: micro-Doppler feature extraction of low-altitude targets
After the communication radiation source signal feature data set dictionary construction is completed, for a target baseband echo signal, the micro Doppler feature extraction of a low-altitude target is realized by adopting an orthogonal matching tracking algorithm;
using the micro-doppler feature parameter dictionary constructed in the third step, x ═ Ψ α can be rewritten as
S(f′)=Ψ(f′)X (15)
The size of the source matrix S (f ') is M multiplied by L, the size of the sparse representation matrix X is PQ multiplied by L, and the size of the sparse base matrix psi (f ') is M ' multiplied by PQ;
defining a down-sampling rate of
η=L/M′ (16)
Where M' is the number of slow time sample points;
under the condition of noise, the sparse expression solution of the MMV model with multiple observation vectors can pass through l0Solving for norm minimization
X=arg min||X||0(17)
Wherein X satisfies the conditionε is the allowable error;
in the solving process, if and only if the formulas (13) and (14) satisfy Ap=AiAnd thetaq=θiX tintin matrix Ψ (f)', aq+(p-1)QIf the row is not zero, extracting information such as amplitude, phase and the like of the target echo signal by extracting a non-zero row vector in the matrix, and obtaining information such as the rotation radius, the initial phase and the like of the target by formula conversion, namely extracting the micro Doppler feature of the low altitude target.
The method has the advantages that the micro Doppler characteristic information of the target can be obtained by directly processing the echo signal without carrying out characteristic extraction from the image, and the problem that the resolution of the image influences the extraction precision of the characteristic parameters is solved; the multi-observation vector model is adopted to process the target baseband echo signal, the extraction precision of the characteristic parameter is improved along with the increase of the weight of the model, and meanwhile, a better low-altitude target micro Doppler characteristic extraction function can be realized under the condition that the weight reaches a certain value; compared with a method adopting a single observation vector model, the method has the advantage that the time consumption for extracting the target characteristic parameters is less. Therefore, the method has stronger robustness on the micro Doppler feature extraction of the low-altitude target under different signal-to-noise ratios.
Drawings
FIG. 1 is a diagram showing the positional relationship between an object and external radiation sources and receivers;
FIG. 2 shows a graph of the target baseband echo signal sparse solution results;
FIG. 3 shows a Doppler curve of scattering points of a target, wherein FIG. 3(a) shows a target micro-Doppler theoretical curve and FIG. 3(b) shows a target micro-Doppler reconstruction curve;
FIG. 4 shows normalized mean square error at different multiplicity, where FIG. 4(a) shows normalized mean square error for radius of rotation at different multiplicity and 4(b) shows normalized mean square error for initial phase at different multiplicity;
FIG. 5 shows the time consumption of the low-altitude target micro-Doppler feature extraction method of the invention under different multiples;
fig. 6 shows a schematic flow diagram of the method of the invention.
Detailed Description
The technical scheme and the implementation process of the invention are described in detail by combining specific examples.
The low-altitude target micro Doppler feature extraction method comprises the following steps:
the first step is as follows: receiving and preprocessing a target echo signal irradiated by an external radiation source;
a position relation diagram among the target, the external radiation source and the receiver is shown in fig. 1, the selected external radiation source transmits Long Term Evolution (LTE) communication signals, after the signals irradiate the target, the signals reflected by the target are called target echo signals, and after the signals are received by the receiver, the signals are sampled and received signals are output;
the communication signal output by the external radiation source is
Where N is the subcarrier number, K is the code length (related to the modulation scheme), N is the index of the subcarrier frequency, K is the index of the code length, anIs the weighting factor of the nth sub-carrier frequency, AreIs a real part modulation coefficient, AimFor imaginary modulation factor, TpFor Orthogonal Frequency Division Multiplexing (OFDM) symbol duration, f0Is the carrier frequency, and Δ f is the sub-carrier frequency interval;
wherein Δ f satisfies the OFDM orthogonality condition
Where i, j is the number of the sub-carrier frequency, which means that any two sub-carrier frequencies satisfy the orthogonality condition, i.e. it is
Δf=1/Tp(3)
The received target echo signal is
Wherein, tmIs a slow time, tkFor fast time, σ is the scattering coefficient of the target, r (t)m) The sum of the distances between the target and the external radiation source and the distance between the target and the receiver are respectively, and c is the speed of light in vacuum;
the reference signal is a direct wave signal from a receiver to an external radiation source and can be expressed as
Wherein D is the distance between the external radiation source and the receiver;
assuming that the multipath clutter in the received signal is better suppressed and the clean target echo and direct wave are obtained, the echo signal is conjugate multiplied by the reference signal to obtain the baseband echo which can be expressed as
Wherein,is s isr(tk,tm) The conjugate signal of (a);
the second step is that: extracting target micro-motion frequency and constructing a multiple observation vector model;
in a practical environment, strong scattering points on aerial targets and spatial targets are always limited, so that echoes of the aerial targets and the spatial targets have sparsity per se. The invention constructs a multiple observation vector model for the baseband echo, and reduces the data volume of the echo matrix, thereby reducing the operation amount of the feature extraction algorithm. The multiple observation vector model is a general popularization of a single observation vector model, and is used for observing signals with the same sparse structure for multiple times, namely, each column of the signals is sparse, the positions of non-zero elements are the same, and the multiple observation vector model has higher reconstruction precision and better anti-noise capability than the single observation vector model;
(a) multiple observation vector model
Let there be a one-dimensional vector x ═ x (1), …, x (U)]TCoordinates α under the U-dimensional orthonormal basis Ψ contain only B (B < U) non-zero elements, then α is said to be B sparse, where x (1), …,x (u) is the coordinates in a conventional coordinate system. Ψ is referred to as the sparse basis matrix for x. Let us assume that there is an observation matrix phi that is independent of ΨV×U(V < U) linear observation is performed on the vector x, and the observation result can be recorded as y Φ x Φ Ψ α Θ α, where Θ Φ Ψ is referred to as the sensing matrix, the original N-dimensional signal vector x is reduced to an M-dimensional vector y., and when a constraint isometry condition (which is well known to those skilled in the art) is satisfied, the sparse representation of the signal vector x can be obtained by solving the minimum l0Norm resolution, i.e.
α=arg min||α||0(8)
Where α is a sparse representation of the vector x, α satisfies the condition y ═ Φ Ψ α;
consider that if there are L sources with the same sparse structure, let X ═ X1,…,xL]T(X is a number of X's, i.e. X1,…,xLFormed matrix), the same observation matrix may be used and the observations may be represented as
Y=ΦX=Φ[x1,…,xL](9)
This model is called the multiple observation vector (MMV) model, with a multiplicity of L. By using the prior information that the signals have the same sparse structure, the solution can be carried out by the sparse matrix rank minimization, namely, the following equation sets are connected
Wherein A is a sparse representation of X, αlRepresents the coordinate of the l-th source under the U-dimensional orthonormal base and satisfies the condition of Y-phi psi [ α1,…,αL];
(b) Construction of multiple observation vector model
Aiming at the baseband echo signal output in the first step, if the time is tmAnd a fast time tkRespectively sampling M 'and N', then echo signalCan be expressed as a matrix form Sr∈RN′×M′Where R represents a real number matrix. Due to the change of the target attitude, the observed echo matrix can not be directly used as a multiple observation vector Sr. Therefore, a fast time t 'of the same target attitude is extracted'sThe baseband echo of (1) constructs a multiple observation vector model with the same sparse structure, so that the sparse expression of the multiple observation vector model can be solved;
since echoes at a certain fast time instant need to be extracted, a priori information on the target micro-motion frequency is needed. The method extracts the target micro-motion frequency f by using the sinusoidal frequency modulation Fourier Bessel transformation, can directly calculate the baseband echo, extracts the target micro-motion frequency, and has higher precision. Constructing a multiple observation vector model of the baseband echo by using the extracted micro-motion frequency f ', wherein f' is the estimation result of the extracted target micro-motion frequency
S(f′)=[sb(t′s1,t′m1),sb(t′s2,t′m2),…,sb(t′sN′,t′mN′)](11)
Wherein T ═ 1/f 'is the target inching period, T'stIs a marked fast time instant, t'miIs a fast time t'siCorresponding slow time series, i.e. t'mi=t′si,t′si+T,...,t′si+LT,The number of occurrences of the same attitude at the radar illumination time t,denotes rounding down, sb(t′si,t′mi) Is represented by t'si,t′miA baseband echo signal that is a parameter, i 1, 2.., N';
the third step: constructing a target micro Doppler characteristic parameter dictionary;
after a baseband echo multiple observation vector model is constructed, feature extraction needs to be carried out on the model, and a target micro Doppler feature parameter dictionary is constructed; the amplitude, the initial phase and the micro-motion frequency of the target baseband echo signal are used as micro-Doppler characteristic parameters of a target, and the micro-motion frequency of the target is obtained by other methods, so that the amplitude and the initial phase of the target baseband echo signal form a characteristic vector;
for a specific external radiation source signal, a received signal in practical application contains a target echo signal, a direct wave signal and other clutter, wherein the clutter signal is filtered by an existing clutter suppression method;
and for the preprocessed target echo signals, constructing a multiple observation vector model according to the second step, and extracting the echo of a certain sub-carrier frequency by utilizing the orthogonality of Orthogonal Frequency Division Multiplexing (OFDM) signals for processing, wherein the carrier frequency of the target in the baseband echo is a known quantity. Assuming that the number of signal amplitudes and phases in the sparse dictionary are P terms and Q terms respectively, the sparse dictionary can be expressed as { A }1,…,APAnd { theta }1,…,θQF ' is used as a parameter, and the sparse dictionary Ψ (f ') with f ' as a parameter can be expressed as a matrix
Wherein the sparse dictionary matrix Ψ (f ') has a size of M' × PQ,for extracted slow time instants, row vectors in the matrixA feature vector representing the target micro-doppler, where m is 1, …, L, and
wherein,
the fourth step: extracting micro Doppler features of the low-altitude target;
after the construction of a communication radiation source signal feature data set dictionary is completed, for a target baseband echo signal, the invention adopts an orthogonal matching tracking algorithm to realize the extraction of the micro Doppler feature of a low-altitude target;
using the micro-doppler feature parameter dictionary constructed in the third step, x ═ Ψ α can be rewritten as
S(f′)=Ψ(f′)X (15)
The size of the source matrix S (f ') is M multiplied by L, the size of the sparse representation matrix X is PQ multiplied by L, and the size of the sparse base matrix psi (f ') is M ' multiplied by PQ;
defining a down-sampling rate of
η=L/M′ (16)
Where M' is the number of slow time sample points;
under noisy conditions, the sparse representation of the multiple observation vector (MMV) model can be solved by l0Norm minimization for solving
X=arg min||X||0(17)
Wherein X satisfies the conditionε is the allowable error;
in the solving process, if and only if the formulas (13) and (14) satisfy Ap=AiAnd thetaq=θiX in the matrix Ψ (f'), of|q+(p-1)QIf the row is not zero, then passNon-zero row vectors in the matrix are extracted, information such as amplitude, phase and the like of a target echo signal can be extracted from the non-zero row vectors, and information such as a rotation radius, an initial phase and the like of a target is obtained through formula conversion, so that extraction of the low-altitude target micro Doppler feature is realized.
Example testing
In order to verify the feasibility and robustness of the method in target micro Doppler feature extraction, the experiment carries out computer simulation verification on the micro Doppler features of a plurality of scattering points on a rotating target, the rotating frequency of the micro Doppler features is 25Hz, and the simulation result is analyzed.
Testing one: feasibility test for low-altitude target micro Doppler theoretical feature extraction
The feasibility of the method for extracting the low-altitude target micro Doppler features is tested by 3 target rotation points for feature extraction, wherein the initial phases are pi/2, 5 pi/3 and 4 pi/3 respectively, the rotation radiuses are 0.68m, 3.12m and 2.94m respectively, and the test result is shown in figure 2. And reconstructing a micro Doppler curve of the scattering point of the rotating target by using the extraction result and comparing the micro Doppler curve with a micro Doppler theoretical curve of the scattering point of the rotating target, wherein the result is shown in fig. 3.
As can be seen from the test results in FIG. 2, the target has three rotational scattering points, and the rotational radius and the initial phase of the target are calculated to be {0.680m, 1.571rad }, {2.932m, 5.759rad }, {3.136m, 3.884rad }, respectively. From the result of fig. 3, it can be seen that the micro doppler reconstruction curve of the scattering point of the target is relatively close to the theoretical curve thereof, and although there is a certain deviation on the initial phase, the rotation period and the amplitude are relatively close, so that a better function of extracting the micro doppler feature of the low altitude target can be realized. The test results prove the feasibility of extracting the low-altitude target micro Doppler features.
And (2) testing: robustness test for low-altitude target micro-Doppler theoretical feature extraction
In order to verify the robustness of the method, the influence of Gaussian noise is considered, the SNR is changed from-10 dB to 20dB at equal intervals of 5dB, and 1000 Monte Carlo experiments are carried out. Fig. 4 shows the result of the feature extraction method for extracting the low-altitude target micro-doppler feature under different signal-to-noise ratios. From the test results of fig. 4, it can be known that the method can achieve a better function of extracting the low-altitude target micro doppler features when the feature parameter extraction accuracy is increased along with the increase of the model weight and the weight reaches a certain value. Fig. 5 shows the time-consuming result of the feature extraction method for extracting the low-altitude target micro-doppler features under different down-sampling rates. From the test results of fig. 5, compared with the method using a single observation vector model, the method has the advantages that the time consumption for extracting the target characteristic parameters is less, and the characteristic extraction can be performed more quickly and effectively. The test results also prove the robustness of the method for extracting the low-altitude target micro Doppler features under different signal to noise ratios.
For clarity of illustration, FIG. 6 shows a schematic flow diagram of the method of the present invention.
The invention provides a low-altitude target micro Doppler feature extraction method under the condition of a Gaussian white noise channel, which can be applied to automatic extraction of target micro Doppler features.
Claims (1)
1. A low-altitude target micro Doppler feature extraction method comprises the following steps:
the first step is as follows: reception and preprocessing of target echo signals illuminated by external radiation sources
The external radiation source transmits LTE communication signals, after the LTE communication signals irradiate a target, the signals reflected by the target are called target echo signals, and after the receiver receives the signals, the receiver samples and receives the signals and outputs received signals;
the communication signal output by the external radiation source is
Wherein N is the subcarrier frequency number, K is the code length, N is the index of the subcarrier frequency, K is the index of the code length, anIs the weighting factor of the nth sub-carrier frequency, AreIs a real part modulation coefficient, AimFor imaginary modulation factor, TpFor orthogonal frequency division multiplexing OFDM symbol duration, f0Is the carrier frequency, and Δ f is the sub-carrier frequency interval;
wherein Δ f satisfies the OFDM orthogonality condition
Wherein, i, j is the number of the sub-carrier frequency, which means that any two sub-carrier frequencies satisfy the orthogonality condition, that is
Δf=1/Tp(3)
The received target echo signal is
Wherein, tmIs a slow time, tkFor fast time, σ is the scattering coefficient of the target, r (t)m) The sum of the distances between the target and the external radiation source and the distance between the target and the receiver are respectively, and c is the speed of light in vacuum;
the reference signal is a direct wave signal from a receiver to an external radiation source and can be expressed as
Wherein D is the distance between the external radiation source and the receiver;
assuming that the multipath clutter in the received signal is better suppressed and the clean target echo and direct wave are obtained, the echo signal is conjugate multiplied by the reference signal to obtain the baseband echo which can be expressed as
Wherein,is s isr(tk,tm) The conjugate signal of (a);
the second step is that: extracting target micro-motion frequency and constructing multiple observation vector model
(a) Multiple observation vector model
Let there be a one-dimensional vector x ═ x (1), …, x (U)]TCoordinates α under the U-dimensional orthonormal basis psi only contain B nonzero elements, if B < U, then α is called B sparse, wherein x (1), …, x (U) are coordinates under a conventional coordinate system, psi is called sparse basis matrix of x, and the existence of observation matrix phi irrelevant to psi is setV×U(V < U) carrying out linear observation on the vector x, wherein the observation result can be recorded as y phi x phi psi α phi theta α, and in the formula, theta phi psi is called a perception matrix, the original N-dimensional signal vector x is reduced to an M-dimensional vector y, and when the constraint isometry condition is met, the sparse representation of the signal vector x can be realized by solving the minimum l0Norm resolution, i.e.
α=argmin||α||0(8)
Where α is a sparse representation of the vector x, α satisfies the condition y ═ Φ Ψ α;
consider that if there are L sources with the same sparse structure, let X ═ X1,…,xL]TX is a number of L X, i.e. X1,…,xLThe formed matrix can adopt the same observation matrix, and the observation result can be expressed as
Y=ΦX=Φ[x1,…,xL](9)
The model is called a multiple observation vector MMV model, and the multiplicity is L; solving is carried out by sparse matrix rank minimization by using the prior information that signals have the same sparse structure, namely the following equation sets are combined
Wherein A is a sparse representation of X, αlRepresents the coordinates of the l-th source under the U-dimensional orthonormal basis and satisfies the condition Y-phi psi [ α1,…,αL];
(b) Construction of multiple observation vector model
Aiming at the baseband echo signal output in the first step, if the time is tmAnd a fast time tkRespectively sampling M 'and N', the echo signal can be expressed in a matrix form Sr∈RN′×M′Wherein R represents a real number matrix; due to the change of the target attitude, the observed echo matrix can not be directly used as a multiple observation vector Sr(ii) a Therefore, a fast time t 'of the same target attitude is extracted'sThe baseband echo of (1) constructs a multiple observation vector model with the same sparse structure, so that the sparse expression of the multiple observation vector model can be solved;
because the echo at a certain fast time needs to be extracted, the prior information of the target micro-motion frequency is needed; extracting a target micro-motion frequency f by utilizing sinusoidal frequency modulation Fourier Bessel transformation; constructing a multiple observation vector model of the baseband echo by using the extracted micro-motion frequency f ', wherein f' is the estimation result of the extracted target micro-motion frequency
S(f′)=[sb(t′s1,t′m1),sb(t′s2,t′m2),…,sb(t′sN′,t′mN′)](11)
Wherein T ═ 1/f 'is the target inching period, T'siIs a marked fast time instant, t'miIs a fast time t'siCorresponding slow time series, i.e. t'mi=t′si,t′si+T,...,t′si+LT,The number of occurrences of the same attitude at the radar illumination time t,denotes rounding down, sb(t′si,t′mi) Is represented by t'si,t′miA baseband echo signal that is a parameter, i 1, 2.., N';
the third step: constructing a target micro Doppler feature parameter dictionary
For a specific external radiation source signal, a received signal in practical application contains a target echo signal, a direct wave signal and other clutter, wherein the clutter signal is filtered by an existing clutter suppression method;
for the preprocessed target echo signal, a multiple observation vector model is constructed according to the second step, the orthogonality of OFDM signals is utilized to extract the echo of a certain sub-carrier frequency for processing, and then the carrier frequency of the target in the baseband echo is a known quantity; assuming that the number of signal amplitudes and phases in the sparse dictionary are P terms and Q terms respectively, the sparse dictionary can be expressed as { A }1,…,APAnd { theta }1,…,θQF ' is used as a parameter, and the sparse dictionary Ψ (f ') with f ' as a parameter can be expressed as a matrix
Wherein the sparse dictionary matrix Ψ (f ') has a size of M' × PQ,for decimated slow time instants, row vectors in the matrixA feature vector representing the target micro-doppler, where m is 1, …, L, and
wherein,
the fourth step: micro-Doppler feature extraction of low-altitude targets
After the communication radiation source signal feature data set dictionary construction is completed, for a target baseband echo signal, the micro Doppler feature extraction of a low-altitude target is realized by adopting an orthogonal matching tracking algorithm;
using the micro-doppler feature parameter dictionary constructed in the third step, x ═ Ψ α can be rewritten as
S(f′)=Ψ(f′)X (15)
The size of the source matrix S (f ') is M multiplied by L, the size of the sparse representation matrix X is PQ multiplied by L, and the size of the sparse base matrix psi (f ') is M ' multiplied by PQ;
defining a down-sampling rate of
η=L/M′ (16)
Where M' is the number of slow time sample points;
under the condition of noise, the sparse expression solution of the MMV model with multiple observation vectors can pass through l0Norm minimization for solving
X=argmin||X||0(17)
Wherein X satisfies the conditionε is the allowable error;
in the solving process, if and only if the formulas (13) and (14) satisfy Ap=AiAnd thetaq=θiX tintin matrix Ψ (f)', aq+(p-1)QIf the row is not zero, extracting information such as amplitude, phase and the like of a target echo signal by extracting a non-zero row vector in the matrix, and obtaining information such as a rotation radius, an initial phase and the like of a target by formula conversion, namely extracting the micro Doppler feature of the low altitude target.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910173828.XA CN109932716B (en) | 2019-03-03 | 2019-03-03 | Low-altitude target micro Doppler feature extraction method |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910173828.XA CN109932716B (en) | 2019-03-03 | 2019-03-03 | Low-altitude target micro Doppler feature extraction method |
Publications (2)
Publication Number | Publication Date |
---|---|
CN109932716A true CN109932716A (en) | 2019-06-25 |
CN109932716B CN109932716B (en) | 2023-02-24 |
Family
ID=66986669
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910173828.XA Active CN109932716B (en) | 2019-03-03 | 2019-03-03 | Low-altitude target micro Doppler feature extraction method |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN109932716B (en) |
Cited By (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110716201A (en) * | 2019-09-10 | 2020-01-21 | 中国人民解放军空军工程大学 | Space rotation target video ISAR imaging method based on transmitted pulse period delay design |
CN111157987A (en) * | 2020-01-03 | 2020-05-15 | 中南大学 | Human body target micro Doppler frequency estimation method based on extended Bessel model |
CN111708011A (en) * | 2020-07-10 | 2020-09-25 | 南京天朗防务科技有限公司 | Micro Doppler velocity measurement method based on compressed sensing |
CN112616132A (en) * | 2020-12-16 | 2021-04-06 | 同济大学 | Low-altitude air-ground unmanned aerial vehicle channel multipath tracking method based on geometric prior model |
CN112924944A (en) * | 2021-02-02 | 2021-06-08 | 西安电子工程研究所 | Vehicle target micro-motion signal suppression method based on time-frequency spectrum entropy estimation |
CN113466824A (en) * | 2021-09-06 | 2021-10-01 | 成都锐芯盛通电子科技有限公司 | Unmanned aerial vehicle identification method based on two-dimensional phased array radar |
CN115388712A (en) * | 2022-10-31 | 2022-11-25 | 济钢防务技术有限公司 | Intelligent laser weapon system control method |
Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20160021390A1 (en) * | 2014-07-15 | 2016-01-21 | Alcatel-Lucent Usa, Inc. | Method and system for modifying compressive sensing block sizes for video monitoring using distance information |
US9520051B1 (en) * | 2015-06-29 | 2016-12-13 | Echocare Technologies Ltd. | System and method for implementing personal emergency response system based on UWB interferometer |
US20160379475A1 (en) * | 2015-06-29 | 2016-12-29 | Echocare Technologies Ltd. | Human motion feature extraction in personal emergency response systems and methods |
US20170213109A1 (en) * | 2014-03-31 | 2017-07-27 | Los Alamos National Security, Llc | Classification of multispectral or hyperspectral satellite imagery using clustering of sparse approximations on sparse representations in learned dictionaries obtained using efficient convolutional sparse coding |
CN107728124A (en) * | 2017-09-08 | 2018-02-23 | 中国电子科技集团公司信息科学研究院 | A kind of more radar dynamic regulating methods and device based on comentropy |
WO2018045566A1 (en) * | 2016-09-09 | 2018-03-15 | 深圳大学 | Random pulse doppler radar angle-doppler imaging method based on compressed sensing |
CN108693509A (en) * | 2018-04-08 | 2018-10-23 | 中国人民解放军海军航空大学 | Frequency control battle array radar Ullage frequency focuses moving-target integration detection method |
-
2019
- 2019-03-03 CN CN201910173828.XA patent/CN109932716B/en active Active
Patent Citations (7)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US20170213109A1 (en) * | 2014-03-31 | 2017-07-27 | Los Alamos National Security, Llc | Classification of multispectral or hyperspectral satellite imagery using clustering of sparse approximations on sparse representations in learned dictionaries obtained using efficient convolutional sparse coding |
US20160021390A1 (en) * | 2014-07-15 | 2016-01-21 | Alcatel-Lucent Usa, Inc. | Method and system for modifying compressive sensing block sizes for video monitoring using distance information |
US9520051B1 (en) * | 2015-06-29 | 2016-12-13 | Echocare Technologies Ltd. | System and method for implementing personal emergency response system based on UWB interferometer |
US20160379475A1 (en) * | 2015-06-29 | 2016-12-29 | Echocare Technologies Ltd. | Human motion feature extraction in personal emergency response systems and methods |
WO2018045566A1 (en) * | 2016-09-09 | 2018-03-15 | 深圳大学 | Random pulse doppler radar angle-doppler imaging method based on compressed sensing |
CN107728124A (en) * | 2017-09-08 | 2018-02-23 | 中国电子科技集团公司信息科学研究院 | A kind of more radar dynamic regulating methods and device based on comentropy |
CN108693509A (en) * | 2018-04-08 | 2018-10-23 | 中国人民解放军海军航空大学 | Frequency control battle array radar Ullage frequency focuses moving-target integration detection method |
Non-Patent Citations (2)
Title |
---|
QI-FANG HE ET.AL: "Micro-Doppler parameter estimation via multiple measurement vector model", 《2017 IEEE INTERNATIONAL CONFERENCE ON SIGNAL PROCESSING, COMMUNICATIONS AND COMPUTING (ICSPCC)》 * |
陈是扦 等: "基于参数化解调的旋转目标微多普勒频率提取方法", 《上海航天》 * |
Cited By (14)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN110716201B (en) * | 2019-09-10 | 2024-01-26 | 中国人民解放军空军工程大学 | Space rotation target video ISAR imaging method based on emission pulse period delay design |
CN110716201A (en) * | 2019-09-10 | 2020-01-21 | 中国人民解放军空军工程大学 | Space rotation target video ISAR imaging method based on transmitted pulse period delay design |
CN111157987B (en) * | 2020-01-03 | 2021-11-02 | 中南大学 | Human body target micro Doppler frequency estimation method based on extended Bessel model |
CN111157987A (en) * | 2020-01-03 | 2020-05-15 | 中南大学 | Human body target micro Doppler frequency estimation method based on extended Bessel model |
CN111708011B (en) * | 2020-07-10 | 2022-06-03 | 南京天朗防务科技有限公司 | Micro Doppler velocity measurement method based on compressed sensing |
CN111708011A (en) * | 2020-07-10 | 2020-09-25 | 南京天朗防务科技有限公司 | Micro Doppler velocity measurement method based on compressed sensing |
CN112616132B (en) * | 2020-12-16 | 2022-04-01 | 同济大学 | Low-altitude air-ground unmanned aerial vehicle channel multipath tracking method based on geometric prior model |
CN112616132A (en) * | 2020-12-16 | 2021-04-06 | 同济大学 | Low-altitude air-ground unmanned aerial vehicle channel multipath tracking method based on geometric prior model |
CN112924944A (en) * | 2021-02-02 | 2021-06-08 | 西安电子工程研究所 | Vehicle target micro-motion signal suppression method based on time-frequency spectrum entropy estimation |
CN112924944B (en) * | 2021-02-02 | 2023-06-13 | 西安电子工程研究所 | Vehicle target inching signal suppression method based on time-frequency spectrum entropy estimation |
CN113466824A (en) * | 2021-09-06 | 2021-10-01 | 成都锐芯盛通电子科技有限公司 | Unmanned aerial vehicle identification method based on two-dimensional phased array radar |
CN113466824B (en) * | 2021-09-06 | 2021-11-26 | 成都锐芯盛通电子科技有限公司 | Unmanned aerial vehicle identification method based on two-dimensional phased array radar |
CN115388712A (en) * | 2022-10-31 | 2022-11-25 | 济钢防务技术有限公司 | Intelligent laser weapon system control method |
CN115388712B (en) * | 2022-10-31 | 2023-01-31 | 济钢防务技术有限公司 | Intelligent laser weapon system control method |
Also Published As
Publication number | Publication date |
---|---|
CN109932716B (en) | 2023-02-24 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN109932716B (en) | Low-altitude target micro Doppler feature extraction method | |
CN110412559B (en) | Non-coherent fusion target detection method for MIMO radar of distributed unmanned aerial vehicle | |
CN108051809B (en) | Moving target imaging method and device based on Radon transformation and electronic equipment | |
CN104515971B (en) | Airborne single-station passive positioning method for multiple broadband targets | |
Chabriel et al. | Adaptive target detection techniques for OFDM-based passive radar exploiting spatial diversity | |
Liu et al. | LPI radar signal detection based on radial integration of Choi-Williams time-frequency image | |
Du et al. | Robust classification scheme for airplane targets with low resolution radar based on EMD-CLEAN feature extraction method | |
US7289060B1 (en) | Super high range resolution amplitude and range principal scatterer (SHARP) classifier | |
Del Arroyo et al. | WiMAX OFDM for passive SAR ground imaging | |
CN110174650B (en) | Meteorological radar wind power plant clutter suppression method based on two-dimensional joint interpolation | |
CN109856635A (en) | CSAR ground moving target refocusing imaging method | |
CN109669182B (en) | Passive bistatic SAR moving/static target joint sparse imaging method | |
Fang et al. | Improved passive SAR imaging with DVB-T transmissions | |
CN107942327A (en) | Single channel HRWS SAR imaging methods based on impulse phase coding | |
CN113253251B (en) | FDA-MIMO radar detection method and system when target speed is unknown | |
Li et al. | A fast cross-range scaling algorithm for ISAR images based on the 2-D discrete wavelet transform and pseudopolar Fourier transform | |
Wang et al. | MIMO SAR system using digital implemented OFDM waveforms | |
CN109143235B (en) | Ground moving target detection method for double-base forward-looking synthetic aperture radar | |
CN107748364A (en) | Low wind field speed estimation method based on contraction multistage wiener filter | |
Jang et al. | A single-snapshot localization for monostatic FDA-MIMO radar | |
CN112731304A (en) | Arc array radar clutter suppression method based on azimuth angle domain filtering | |
CN115575921B (en) | Pitching-direction-based multichannel multi-interference-base suppression interference suppression method | |
CN102222230A (en) | Multi-input multi-output imaging radar target identification method based on parameterization dictionary | |
Qiuchen et al. | ISAR cross-range scaling based on the MUSIC technique | |
Zhao et al. | Three‐dimensional high‐resolution MIMO radar imaging via OFDM modulation and unitary ESPRIT |
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 |