CN104777467B - Object detection method based on frequency scan antenna - Google Patents

Abstract

The invention discloses a kind of object detection method based on frequency scan antenna, target quick detection and angle super-resolution are realized by handling raw radar data, is specifically included：Set up echo signal model；Subspace average treatment is carried out to original echoed signals, signal matrix is obtained；Signal matrix to acquisition asks for autocorrelation matrix；Eigenvalues Decomposition is carried out to the signal autocorrelation matrix of acquisition, characteristic value and characteristic vector space is obtained, and the characteristic vector space is divided into signal subspace and noise subspace；The signal subspace and noise subspace of acquisition is orthogonal, and construct FS MUSIC puppet spectrums；Spectrum peak search is carried out to gained puppet spectrum, it is exactly target echo direction to find out the corresponding angle of multiple maximum points.The method of the present invention need not be by array received, it is only necessary to which frequency scan antenna transmission signal, loudspeaker or waveguide port receive signal, handles simple and quick, is suitable for the quick detection of multiple targets.

Description

Object detection method based on frequency scan antenna

Technical field

The present invention relates to Radar Signal Detection technical field, relate more specifically to a kind of target based on frequency scan antenna Detection method.

Background technology

Target detection is always the Important Problems of the association area research such as radar, communication, locally fixed in such as cellphone subscriber Many application scenarios such as location Based service, radar target positioning, there is widespread demand in position, WiFi network.Its medium wave reaches Direction estimation (DOA) receives much concern always in recent years as an importance of target detection.Most of Mutual coupling System is all by the use of array antenna as receiver, and the array signal of reception is handled with MUSIC algorithms, but this germline System, which wants raising resolution ratio, to be needed to increase element number of array, so as to cause cost increase, power consumption increase, circuit structure complicated.

Frequency scan antenna is the frequency of the class antenna that beam position changes with frequency, wherein millimeter wave and Terahertz section The main application scenarios of scanning antenna have：For the Area Surveillance RADAR on battlefield, airport etc., to 2 dimensional region scanning imagery, its In dimension utilize the beam scanning of FREQUENCY CONTROL, another dimension utilizes antenna rotation；For spacecraft safe landing Terrain detection radar, the detection to obstacles such as touchdown area rock, slopes is realized using frequency scan antenna；For Terahertz two Dimension imaging, wherein orientation utilize frequency scanning detection target.But these radars have a defect, in frequency scanning direction Resolution ratio be limited by beam angle.The method for improving frequency scanning directional resolution is typically increase antenna aperture to reduce day Line beam angle, but this mode can bring the lifting of difficulty of processing, and design complexities increase, and many times Due to the limitation that radar system is overall, it is impossible to allow the size of antenna very big.Therefore, frequency scanning direction is improved from algorithm Target resolution capability, as urgent problem to be solved.At present, MUSIC algorithms are applied to based on frequency scanning there is not yet having The report of antenna system.

The content of the invention

In view of this, it is a primary object of the present invention to propose a kind of object detection method based on frequency scan antenna, To realize target quick detection and angle super-resolution by handling raw radar data.The method of the present invention need not be by array Receive, it is only necessary to frequency scan antenna transmission signal, loudspeaker or waveguide port receive signal, are suitable for the quick of multiple targets Detection.The inventive method may apply to terahertz imaging, Area Surveillance RADAR, Non-Destructive Testing, target and quickly recognize and track Deng field.

To achieve these goals, as one aspect of the present invention, the invention provides one kind based on frequency scanning day The object detection method of line, comprises the following steps：

Step S1：Set up echo signal model；

Step S2：Subspace average treatment is carried out to original echoed signals, signal matrix is obtained；

Step S3：Autocorrelation matrix is asked for the signal matrix obtained in step S2；

Step S4：Eigenvalues Decomposition is carried out to the signal autocorrelation matrix obtained in step S3, characteristic value and feature is obtained Vector space, and the characteristic vector space is divided into signal subspace and noise subspace；

Step S5：The signal subspace and noise subspace that are obtained in step S4 is orthogonal, and construct FS-MUSIC Puppet spectrum；

Step S6：Spectrum peak search is carried out to pseudo- spectrum obtained by step S5, it is exactly mesh to find out the corresponding angle of multiple maximum points Mark echo direction.

Wherein, step S1 includes：

Echo-signal is expressed as X=AS+U, wherein X=[x (f₁), x (f₂), x (f₃) ..., x (f_N)]^T, it is that N*1 is tieed up back Each element in ripple signal vector, the vector is the signal of the different frequency received, x (f_n) it is abbreviated as x_n, n=1, 2 ..., N；

A=[α (f, β₁) α (f, β₂) ... α (f, β_M)], it is that N*M dimension arrays are oriented to matrix, each of wherein A is classified as one Array steering vector；

α (f, β_m)=[a (f₁, β_m)a(f₂, β_m)...a(f_N, β_m)]^T, m=1,2..., M, wherein a (f_n, β_m) it is frequency f_n Lower β_mThe dual station directivity factor in direction；

S=[s (β₁)s(β₂)...s(β_M)]^T, it is β_mThe backscattering coefficient of direction target, s (β_m) it is abbreviated as s_m, m=1, 2 ..., M；

U=[u (f₁)u(f₂)...u(f_N)]^TFor the noise of different frequency, x (f_n) it is abbreviated as x_n, n=1,2 ..., N.

Wherein, the dual station directivity factor a (f in step S1_n, β_m) include amplitude or phase information.

Wherein, step S2 includes：

Subspace average treatment is carried out to original echoed signals X：X=[x^{1}x^{2}...x^{P}],

Wherein：

x^{1}=[x₁, x₂..., x_Q]^T

x^{2}=[x₂, x₃..., x_Q+1]^T,

x^{P}=[x_p, x_P+1..., x_N]^T, P is subspace number, and P ＞ D (set D as target number), and Q is per sub-spaces Element number, and P+Q+1=M.

Wherein, step S3 includes：

Autocorrelation matrix is asked for according to following formula to the signal matrix that is obtained in step S2：

Wherein, P is subspace number, and H represents conjugate transposition.

Wherein, step S4 includes：

Eigenvalues Decomposition is carried out to the signal autocorrelation matrix obtained in step S3, characteristic value is obtainedWith feature to Quantity spaceWhereinAnd λ₁≥λ₂≥...≥λ_Q, the corresponding characteristic vector space of characteristic valueWhereinIt is characterized value λ₁(i=1,2 ..., Q) corresponding characteristic vector；

Known target number is D, so only D maximum characteristic value is relevant with signal, is denoted as signal characteristic valueCharacteristic value equally related to noise is λ_N={ λ_D+1, λ_D+2..., λ_Q}；

Characteristic vector space is divided into signal subspace and noise subspace, wherein signal subspaceDimension is Q*D；Noise subspaceDimension is Q* (Q-D).

Wherein, step S5 includes：

Array steering vector in step S1 is denoted as more common form：α (f, β)；

The signal subspace and noise subspace that are obtained in step S4 is orthogonal, it can thus be appreciated that array steering vector with Noise subspace is also orthogonal, i.e.,Wherein ε is a very small number；And

Construct FS-MUSIC puppet spectrums：

Wherein, β={ β₁, β₂..., β_M}。

Wherein, it is the pseudo- spectrum formula with weighted type that the formula of FS-MUSIC puppet spectrums is constructed described in step S5.

Wherein, also include after step S6：

Step S7：By the tranmitting frequency and scanning angle relation, the multiple target that step S6 is obtained of frequency scanning antenna Echo direction is converted into multiple frequencies；And

Step S8：The signal extraction of the obtained the multiple frequencies of step S7 and neighbouring frequency band is come out, enters horizontal pulse pressure Contracting, obtains the two dimensional image of target.

Wherein, to be applied to terahertz imaging, Area Surveillance RADAR, Non-Destructive Testing, target quick for the object detection method Identification and tracking field.

Understand that method of the invention has the advantages that based on above-mentioned technical proposal：(1) this method is swept using frequency The beam scanning characteristic of antenna is retouched, realizes that target is differentiated in distance to the two dimension with orientation；Relative to other target acquisitions Mode, with simple system, the advantage that cost is low, detection speed is fast；(2) FS-MUSIC angles oversubscription proposed by the present invention is utilized Distinguish algorithm, it is possible to achieve angle-resolved less than 3dB beam angles；(3) using the object detection method of the present invention, by Fig. 3 The corresponding frequency-region signal of two crest angles is extracted, and carries out the two-dimensional position information that process of pulse-compression obtains target, such as Shown in Fig. 4, two targets are in distance to distinguishable with orientation.

Brief description of the drawings

Fig. 1 is the principle schematic of the object detection method based on frequency scan antenna of the present invention；

Fig. 2 is the flow chart of the FS-MUSIC angle super-resolution algorithms of the present invention；

Fig. 3 is the result of the FS-MUSIC algorithms of the present invention, beam angle-scanning angle figure；

Fig. 4 is the result of the object detection method of the present invention, the two-dimensional position information schematic diagram of target.

Embodiment

For the object, technical solutions and advantages of the present invention are more clearly understood, below in conjunction with specific embodiment, and reference Accompanying drawing, the present invention is described in further detail.

The invention discloses a kind of FS-MUSIC angle super-resolution algorithms based on frequency scan antenna, achievable wave beam is swept Retouch the angle super-resolution that direction is less than 3dB beam angles；Another further aspect, the invention discloses a kind of based on frequency scan antenna Object detection method, as shown in figure 1, wave beam is in a scanning direction, can obtain target range to the information with orientation, its Middle orientation is differentiated to be realized using the beam scanning function and FS-MUSIC algorithms of frequency scan antenna, and distance uses arteries and veins to resolution Rush compress technique realization.

The FS-MUSIC algorithm steps of the present invention are as follows：

Step S1：Set up echo signal model.Echo-signal can be expressed as X=AS+U；

Wherein X=[x (f₁), x (f₂), x (f₃) ..., x (f_N)]^T, it is that N*1 dimension echo-signals are vectorial, it is each in the vector Element is the signal of the different frequency received, x (f_n) it is abbreviated as x_n, n=1,2 ..., N；

A=[α (f, β₁) α (f, β₂) ... α (f, β_M)] (1)

A is that N*M dimension arrays are oriented to matrix, and each of wherein A is classified as an array steering vector；

α (f, β_m)=[a (f₁, β_m)a(f₂, β_m)...a(f_N, β_m)]^T, m=1,2..., M, wherein a (f_n, β_m) it is frequency f_n Lower β_mThe dual station directivity factor in direction；

S=[s (β₁)s(β₂)...s(β_M)]^T, it is β_mThe backscattering coefficient of direction target, s (β_m) it is abbreviated as s_m, m=1, 2 ..., M；

U=[u (f₁)u(f₂)...u(f_N)]^TFor the noise of different frequency, x (f_n) it is abbreviated as x_n, n=1,2 ..., N.

Step S2：Subspace average treatment is carried out to original echoed signals X：X=[x^{1}x^{2}...x^{p}],

Wherein：

x^{1}=[x₁, x₂..., x_Q]^T,

x^{2}=[x₂, x₃..., x_Q+1]^T,

x^{P}=[x_p, x_P+1..., x_N]^T, it is subspace number to make P.

Step S3：Autocorrelation matrix is asked for the signal matrix obtained in step S2：

Wherein H represents conjugate transposition.

Step S4：Eigenvalues Decomposition is carried out to the signal autocorrelation matrix obtained in step S3, characteristic value is obtainedAnd spy Levy vector spaceWhereinAnd λ₁≥λ₂≥...≥λ_Q, the descending corresponding feature of characteristic value to Measure and beAssuming that known target number is D, so only D maximum characteristic value is relevant with signal, it is denoted as Signal characteristic valueCharacteristic value equally related to noise isBy feature Vector space is divided into signal subspace and noise subspace, and wherein signal subspace isDimension is Q*D；Noise subspace isDimension is Q* (Q-D).

Step S5：Array steering vector in step S1 is denoted as more common form：α (f, β).In step S4 The signal subspace and noise subspace of acquisition are orthogonal, it can thus be appreciated that array steering vector is also orthogonal with noise subspace, i.e.,Wherein ε is a very small number.

Construct FS-MUSIC puppet spectrums：

Wherein β={ β₁, β₂..., β_M}。

Step S6：To pseudo- spectrum obtained by step S5, spectrum peak search is carried out, it is exactly mesh to find out the corresponding angle of D maximum point Mark echo direction.

In above-mentioned algorithm, the dual station directivity factor a (f in step S1_n, β_m) can be amplitude form, it can also include Phase information.

Pseudo- spectrum form is not limited to formula provided by the present invention in step S5, the pseudo- spectrum formula that can also have weighted type Etc. various forms of pseudo- spectrum formula.

The invention also discloses a kind of object detection method based on frequency scan antenna, step is as follows：

Step G1：By the tranmitting frequency and scanning angle relation, the D target echo side that step S6 is obtained of frequency scanning antenna To being converted into D frequency.

Step G2：The signal extraction of D obtained frequency of step G1 and neighbouring frequency band is come out, pulse compression is carried out, obtains To the two dimensional image of target.

By actually detected test, method of the invention has following advantage：

(1) this method utilizes the beam scanning characteristic of frequency scan antenna, realizes target in distance to two with orientation Dimension is differentiated；Relative to other target acquisition modes, with simple system, the advantage that cost is low, detection speed is fast.

(2) FS-MUSIC angle super-resolutions algorithm proposed by the present invention is utilized, it is possible to achieve less than the angle of 3dB beam angles Degree is differentiated.The algorithm is verified below by experiment.Terahertz frequency scanning target acquisition experiment parameter sets as follows Shown in table 1：

Table 1

Parameter type	Numerical value	Unit
			Bandwidth range	230~330	GHz
Beam-scanning angles	- 36~-10	Degree
			3dB beam angles	2.2~2.5	Degree
Target range	1.57	m
			Target interval	1.6	Degree

Two angle on target intervals are less than 3dB beam angles, and directly receiving data can not differentiate.Fig. 3 is to be calculated with FS-MUSIC The result that method is handled, it can be found that two targets be apparent from be resolved out, 1.78 ° of angle interval, error is allowing to connect By within the scope of.

(3) using the object detection method of the present invention, the corresponding frequency-region signal of two crest angles in Fig. 3 is extracted Come, carry out process of pulse-compression and obtain the two-dimensional position information of target, as shown in figure 4, two targets in distance to equal with orientation It is distinguishable.

Particular embodiments described above, has been carried out further in detail to the purpose of the present invention, technical scheme and beneficial effect Describe in detail bright, it should be understood that the foregoing is only the present invention specific embodiment, be not intended to limit the invention, it is all Within the spirit and principles in the present invention, any modification, equivalent substitution and improvements done etc. should be included in the protection of the present invention Within the scope of.

Claims (9)

1. a kind of object detection method based on frequency scan antenna, comprises the following steps：

Step S1：Set up echo signal model：Echo-signal is expressed as X=AS+U,

Wherein X=[x (f₁), x (f₂), x (f₃) ..., x (f_N)]^T, it is each element in N*1 dimension echo-signal vectors, the vector For the signal of the different frequency received, x (f_n) it is abbreviated as x_n, n=1,2 ..., N；

A=[α (f, β₁) α (f, β₂) ... α (f, β_M)], it is that N*M dimension arrays are oriented to matrix, each of wherein A is classified as a battle array Row steering vector；

α (f, β_m)=[a (f₁, β_m) a(f₂, β_m) ... a(f_N, β_m)]^T, m=1,2..., M, wherein a (f_n, β_m) it is frequency f_nUnder β_mThe dual station directivity factor in direction；

S=[s (β₁) s(β₂) ... s(β_M)]^T, it is β_mThe backscattering coefficient of direction target, s (β_m) it is abbreviated as s_m, m=1, 2 ..., M；

U=[u (f₁) u(f₂) ... u(f_N)]^TFor the noise of different frequency, x (f_n) it is abbreviated as x_n, n=1,2 ..., N；

Step S2：Subspace average treatment is carried out to original echoed signals, signal matrix is obtained：

Step S3：Autocorrelation matrix is asked for the signal matrix obtained in step S2：

Step S4：Eigenvalues Decomposition is carried out to the signal autocorrelation matrix obtained in step S3, characteristic value and characteristic vector is obtained Space, and the characteristic vector space is divided into signal subspace and noise subspace；

Step S5：The signal subspace and noise subspace that are obtained in step S4 is orthogonal, and construct FS-MUSIC puppet spectrums：

Step S6：Spectrum peak search is carried out to pseudo- spectrum obtained by step S5, it is exactly that target is returned to find out the corresponding angle of multiple maximum points Ripple direction.

2. it is described double in the object detection method according to claim 1 based on frequency scan antenna, wherein step S1 Stand directivity factor a (f_n, β_m) include amplitude or phase information.

3. the object detection method according to claim 1 based on frequency scan antenna, wherein step S2 includes：

Subspace average treatment is carried out to original echoed signals X：X=[x^{1} x^{2} ... x^{P}],

Wherein：

x^{1}=[x₁, x₂..., x_Q]^T,

x^{2}=[x₂, x₃..., x_Q+1]^T,

......

x^{P}=[x_P, x_P+1..., x_N]^T, P is subspace number, and its value is more than target number；Q is per sub-spaces element Number, and P+Q+1=M.

4. the object detection method according to claim 3 based on frequency scan antenna, wherein step S3 includes：

Autocorrelation matrix is asked for according to following formula to the signal matrix that is obtained in step S2：

$R_{xx}=\frac{1}{P}\underset{p=1}{\overset{P}{\mathrm{\Σ}}}{x}^{\left\{p\right\}}{x}^{\left\{p\right\}H};$

Wherein, P is subspace number, and H represents conjugate transposition.

5. the object detection method according to claim 4 based on frequency scan antenna, wherein step S4 includes：

Eigenvalues Decomposition is carried out to the signal autocorrelation matrix obtained in step S3, characteristic value is obtainedIt is empty with characteristic vector BetweenWhereinAnd λ₁≥λ₂≥...≥λ_Q, the corresponding characteristic vector space of characteristic valueWhereinIt is characterized value λ_i(i=1,2 ..., Q) corresponding characteristic vector；

Known target number is D, so only D maximum characteristic value is relevant with signal, is denoted as signal characteristic valueCharacteristic value equally related to noise is

Characteristic vector space is divided into signal subspace and noise subspace, wherein signal subspace Dimension is Q*D；Noise subspaceDimension is Q* (Q-D).

6. the object detection method according to claim 5 based on frequency scan antenna, wherein step S5 includes：

Array steering vector in step S1 is denoted as more common form：α (f, β)；

The signal subspace and noise subspace that are obtained in step S4 is orthogonal, it can thus be appreciated that array steering vector and noise Subspace is also orthogonal, i.e.,Wherein ε is a very small number；And

Construct FS-MUSIC puppet spectrums：

$P_{FS-MU}\left(\mathrm{\β}\right)=\frac{\mathrm{\α}{(f,\mathrm{\β})}^H\mathrm{\α}(f,\mathrm{\β})}{\mathrm{\α}{(f,\mathrm{\β})}^H{\hat{E}}_N\widehat{E_N^H}\mathrm{\α}(f,\mathrm{\β})};$

Wherein, β={ β₁, β₂..., β_M}。

7. the object detection method according to claim 6 based on frequency scan antenna, is constructed wherein described in step S5 The formula of FS-MUSIC puppet spectrums is the pseudo- spectrum formula with weighted type.

8. the object detection method according to claim 6 based on frequency scan antenna, wherein also being wrapped after step S6 Include：

Step S7：By the tranmitting frequency and scanning angle relation, the multiple target echo that step S6 is obtained of frequency scanning antenna Direction is converted into multiple frequencies；And

Step S8：The signal extraction of the obtained the multiple frequencies of step S7 and neighbouring frequency band is come out, pulse compression is carried out, obtains To the two dimensional image of target.

9. the object detection method according to claim 1 based on frequency scan antenna, wherein the object detection method Terahertz imaging, Area Surveillance RADAR, Non-Destructive Testing, target is applied to quickly to recognize and tracking field.