CN109946668A - Target Secondary Identification Method Based on Multi-beamforming - Google Patents

Abstract

The invention discloses a kind of secondary discriminating method of the target based on Multibeam synthesis, mainly solve the problems, such as that prior art operand is big and detection efficiency is low, implementation step are as follows: 1) determine the object element in clutter secondary lobe area；2) the land clutter azimuth for having identical Doppler frequency with object element is calculated；3) azimuth coverage for determining multi-beam calculates the weight vector of multi-beam；4) difference of the average output power of multi-beam and the output power of main beam is judged whether into false-alarm compared with secondary detection thresholding；5) all object elements are traversed, real object element is saved, complete secondary detection.The present invention forms multiple wave beams within the scope of special angle, according to the otherness of main beam and the output power of multi-beam, judge whether it is false-alarm, significantly reduce false-alarm number, improve the detection performance to moving target, operand is reduced, can be used for detection of the airborne early warning radar to ground moving object under non-homogeneous environment.

Description

Target Secondary Identification Method Based on Multi-beamforming

technical field

The invention belongs to the technical field of radar, in particular to a secondary target identification method, which can be used to reduce the false alarm number of the CFAR technology detection method.

Background technique

The main task of airborne early warning radar is to detect and track moving targets such as aircraft, missiles, and ships. In the detection of moving targets in airborne pulse Doppler radar, the motion of the radar causes the stationary objects on the ground to generate radial velocity relative to the radar, resulting in Doppler spread of ground echoes. The energy of the main beam of the radar antenna is very strong, and the ground echo energy in the irradiation area is also very strong. The area occupied by the distance Doppler plane is called the main lobe clutter area. The area occupied by the ground echo generated by the antenna side lobe beam irradiation is called the side lobe clutter area, and the power of the side lobe clutter area is relatively weak. The area other than this is called the noise area or the clear area. Because the energy of the main lobe clutter region is very strong, it is difficult to distinguish the target when it falls in this region. Therefore, usually only the targets falling outside the main lobe clutter region are detected. The radar moving target detection is only for the target illuminated by the main beam of the radar. Although the target in the main beam has the same azimuth angle as the ground in the illuminated area, when the speed of the target is large enough, the speed of the target itself causes it to be relative to the target. The radial velocity of the carrier and the ground clutter are different from the radial velocity of the carrier. In the Doppler frequency domain, the moving target will contain the Doppler frequency shift generated by its own motion. This frequency difference is in the range Doppler frequency. It is reflected on the Le diagram that the target falls in the non-main lobe clutter region. At this time, by using the obvious difference between the target power and the clutter and noise power, and by setting the detection threshold reasonably and effectively, the target in the non-main lobe clutter region can be detected.

In order to obtain predictable, stable detection performance, radar system designers generally tend to design detectors with a constant probability of false alarms. To achieve this, the actual interfering noise power level must be estimated from the data in real time, and then the detection threshold should be adaptively adjusted to obtain a good detection probability and maintain a tolerable false alarm probability. A detector processor that maintains a constant false alarm probability is called a constant false alarm rate CFAR processor. The approach used in CFAR processing is based on the following two assumptions:

1) The clutter uniformity of adjacent units is good, and its statistical characteristics are in good consistency with the unit to be detected;

2) Adjacent cells contain only interfering noise and no target.

Under the above conditions, the statistical characteristics of interference and clutter of the unit to be detected can be estimated from the data of adjacent units.

When the unit to be detected contains discrete clutter points, the statistical characteristics of the adjacent distance units are inconsistent with the statistical characteristics of the unit to be detected. At this time, the echo strength of the unit to be detected is often higher than the threshold level estimated and set by the reference unit, thereby causing a false alarm. In this special case, it is impossible to reduce the false alarm rate by reasonably selecting the reference unit, but discrete strong clutter points are ubiquitous in the working environment of the radar, such as man-made railways, power lines and other strong scatterers will produce discrete strong clutter. polka dots.

Gerlach K. et al. used the FRACTA algorithm to achieve robust STAP processing in non-uniform environments. Although the FRACTA algorithm can provide robust detection performance in the presence of outliers, non-uniform clutter and interference in the data, the processing process is fixed and cumbersome, and the engineering versatility is not strong. The repeated checking of the RC algorithm and the adaptive power residual APR algorithm is too complicated, the amount of calculation is large, and the detection efficiency is low.

The adaptive coherent detector ACE can determine whether the echo contains the target by measuring the square of the cosine of the angle between the whitened data vector and the whitened steering vector, and comparing this value with the detection threshold. The method has CFAR characteristics and good detection performance. However, in the whitening process of ACE, it involves the inversion of the interference-plus-noise autocorrelation matrix. Usually, the space-time dimension of the matrix ranges from several hundred to several thousand, the number of samples is large, and the inversion operation is complicated, which requires more computing power. high.

SUMMARY OF THE INVENTION

The purpose of the present invention is to solve the above-mentioned deficiencies of the prior art, and propose a target secondary identification method based on multi-beam forming, which can reduce the number of false alarms and improve the detection efficiency with a small amount of calculation.

In order to achieve the above object, the technical scheme of the present invention comprises the following steps:

(1) Obtain the distance Ri and the normalized Doppler frequency of the _ith target unit M is the total number of target units;

(2) According to the normalized Doppler frequency The size of judging whether the i-th target unit is located in the side lobe clutter area: if located in the side lobe clutter area, then execute (3); otherwise, directly save the unit and jump to (9);

(3) According to the i-th target unit information R _i and Calculate the pitch angle φ _i of the unit and the same as the target unit The ground clutter azimuth θ _i ;

(4) Obtain the azimuth angle θ ₀ of the main beam, calculate the main beam width Δθ, and obtain the azimuth angle Φ ₀ of the target unit and the angle Φ _c of the ground clutter respectively:

Φ ₀ =(θ ₀ -Δθ/2,θ ₀ +Δθ/2)

Φ _c =(θ _i -Δθ/2,θ _i +Δθ/2)

where θ _i is the azimuth of the ground clutter;

(5) Form K pointing beams in the angle area other than the angle Φ _c of the ground clutter and the azimuth angle Φ ₀ of the target unit, and calculate the weight vector of the K beams:

w=[w ₁ ,...,w _j ,...,w _K ]

Among them, w _j is the weight vector of the j-th beam, j=1,2,...,K;

(6) Obtain the original echo data received by the radar at the distance R _i right Perform the Fourier transform of the pulse dimension to obtain the array element domain data x _i corresponding to the normalized Doppler frequency f _d, i, N is the number of array elements;

(7) According to the results of (5) and (6), calculate the output power of the K beams:

y=[y ₁ ,…,y _j ,…,y _K ]

in, is the output power of the j-th beam, (·) ^H represents the conjugate transpose of the elements in the brackets, and |(·)| ² represents the square of the absolute value of the elements in the brackets;

(8) According to the result of (6), calculate the output power of the main beam:

Among them, w ₀ is the weight vector corresponding to the main beam azimuth angle θ ₀ ;

(9) According to the results of (7) and (8), calculate the absolute value Δy of the difference between the output power of the main beam and the average output power of the K beams, and convert Δy into a decibel value Δy':

Δy'=10log10(Δy)

(10) Given the secondary detection threshold ζ=15, compare Δy' and ζ, and judge whether the target unit is a false alarm: if Δy'≥ζ, then judge as a real target, save the target unit, and execute (11); Otherwise, it is determined as a false alarm, and (11) is executed;

(11) Determine whether all target units have been detected: if the detection is completed, end the secondary target identification process; otherwise, obtain the next target unit information, and repeat (2) to (10).

The present invention has the following advantages:

Aiming at the false alarm points caused by discrete strong clutter points, the invention forms a plurality of beams with different directions in the angular area outside the azimuth angle range of ground clutter and the azimuth angle range of the target of interest, and compares the average output of these beam formations. The power and the output power of the main beam direction are used to determine whether it is a false alarm, which reduces the amount of calculation, effectively reduces the number of false alarms, improves the detection performance of the radar for moving targets, and is easy to implement in engineering.

Description of drawings

Fig. 1 is the general flow chart of the present invention;

Fig. 2 is the range Doppler map after PD processing used for simulation of the present invention;

Fig. 3 is the CFAR detection result diagram that the simulation of the present invention uses;

FIG. 4 is a result diagram of the secondary detection of the CFAR detection result using the present invention.

Detailed ways

Embodiments of the present invention and results are described in further detail below in conjunction with the accompanying drawings:

1, the implementation steps of the present invention are as follows:

Step 1, obtain target unit parameters.

Obtain the total number M of target units, obtain the distance R _i of the i-th target unit and the normalized Doppler frequency

Step 2: Determine the positional relationship between the target unit and the sidelobe clutter region.

2a) Calculate the maximum normalized Doppler frequency of the sidelobe clutter:

Among them, v=[v _x , v _y , v _z ] is the velocity vector of the carrier aircraft, v _x , v _y , v _z respectively represent the three components of velocity v on the body coordinate axis; λ ₀ is the radar operating wavelength, _fr is the pulse repetition frequency;

2b) Determine the normalized Doppler frequency range of the sidelobe clutter region:

f _d,c =[-f _{d,c_max} ,f _{d,c_max} ],

2c) Judging the normalized Doppler frequency and the normalized Doppler range f _d,c of the sidelobe clutter region:

like but The corresponding target unit is located in the side lobe clutter area, and step 3 is performed;

Otherwise, the target unit is located in the clear area, save the target unit, and jump to step 11.

Step 3, calculate the pitch angle φ _i of the target unit and the same as the target unit The ground clutter azimuth θ _i .

3a) Calculate the pitch angle φ _i of the ith target unit:

Among them, ha is the flight height of the carrier aircraft, a _e is the equivalent radius of the earth, and R _i is the distance of the _ith target unit;

3b) The calculation is the same as the target unit The ground clutter azimuth θ _i :

Among them, A=v _x cosφ _i , B=v _y cosφ _i , φ _i is the pitch angle of the ith target unit, v _x , v _y , v _z respectively represent the three components of the carrier aircraft’s velocity vector on the body coordinate axis, λ ₀ is the radar operating wavelength, and f _r is the pulse repetition frequency .

Step 4: Calculate the azimuth range of the target unit and the azimuth range of ground clutter.

4a) Calculate the main beam width Δθ:

Δθ=θ _3dB =57.3°×0.886λ ₀ /N/d

Among them, N is the number of array elements, and d is the distance between array elements.

4b) Obtain the azimuth angle θ ₀ of the main beam, and obtain the target unit azimuth angle Φ ₀ and the ground clutter azimuth angle Φ _c respectively:

Φ ₀ =(θ ₀ -Δθ/2,θ ₀ +Δθ/2)

Φ _c =(θ _i -Δθ/2,θ _i +Δθ/2)

where θ _i is the azimuth angle of the ground clutter and Δθ is the main beam width.

Step 5: Calculate the weight vectors of the K beams.

5a) Calculate the range of the azimuth angle θ _j of the K beams:

θ _j ∈(-π/2,π/2)-(Φ ₀ ∪Φ _c ),j=1,…,K

Among them, Φ ₀ is the azimuth angle range of the target unit, and Φ _c is the azimuth angle range of the ground clutter;

5b) Calculate the weight vector of the jth beam pointing according to the range of θ _j :

w _j =[w _j,0 ,…,w _j,n ,…,w _j,N-1 ] ^H

Among them, w _j,n =exp(j2πk _j ·d _n ), k _j =[sinθ _j cosφ _i cosθ _j cosφ _i sinφ _i ] is the line of sight vector of multiple beams, θ _j is the azimuth angle of the jth beam; φ _i is the pitch angle of the i-th target unit; d _n =[d _x,n ,d _y,n ,d _z,n ] is the position vector of the n-th array element, d _x,n , _dy,n , d _{z, n} represent the three components of the array element on the coordinate axis of the array, respectively.

5c) According to the weight vector w _j of the j-th beam, obtain the weight vector of the K beams:

w=[w ₁ ,...,w _j ,...,w _K ].

Step 6: Obtain the array element domain data xi corresponding to the normalized Doppler frequency f _{d, i} .

Get the raw echo data received by the radar at distance R _i right Perform the Fourier transform of the pulse dimension to obtain the array element domain data x _i corresponding to the normalized Doppler frequency f _d, i, N is the number of array elements.

Step 7: Calculate the output power of the multi-beam.

7a) Calculate the output power of the jth beam according to the results of steps 5 and 6:

Among them, w _j is the weight vector of the jth beam, x _i is the array element domain data corresponding to the normalized Doppler frequency f _d,i , (·) ^H represents the conjugate transpose of the elements in the brackets, | ( )| ² means taking the square of the absolute value of the elements in parentheses.

7b) According to the output power y _j of the jth beam, obtain the output power of the K beams:

y=[y ₁ ,...,y _j ,...,y _K ].

Step 8: Calculate the output power of the main beam.

8a) Calculate the weight vector w _{0 of the main beam azimuth θ 0 :}

w ₀ =[w _0,0 ,…,w _0,n ,…,w _0,N-1 ] ^H ,

where w _0,n =exp(j2πk ₀ ·d _n ), k ₀ =[sinθ ₀ cosφ _i ,cosθ ₀ cosφ _i ,sinφ _i ] is the line-of-sight vector of the main beam, θ ₀ is the main beam azimuth, φ _i is the pitch angle of the i-th target unit; d _n =[d _x,n ,d _y,n ,d _z,n ] is the position vector of the n-th array element, d _x,n , _dy,n ,d _{z and n} respectively represent the three components of the array element on the coordinate axis of the array.

8b) According to the result of step 6, calculate the output power of the main beam:

Among them, w ₀ is the weight vector corresponding to the azimuth angle θ ₀ of the main beam, and x _i is the array element domain data corresponding to the normalized Doppler frequency f _d,i .

Step 9: Calculate the absolute value of the difference between the output power of the main beam and the average output power of the multi-beam, and convert it into a decibel value.

9a) According to the results of step 7 and step 8, calculate the absolute value Δy of the difference between the output power of the main beam and the average output power of the K beams:

9b) Convert Δy to decibel value Δy':

Δy'=10log10(Δy).

Step 10, according to the secondary detection threshold, determine whether the target unit is a false alarm.

Given the secondary detection threshold ζ=15, compare Δy' and ζ to determine whether the target unit is a false alarm:

If Δy'≥ζ, it is determined as a real target, save the target unit, and execute step 11;

Otherwise, it is determined as a false alarm, and step 11 is executed;

Step 11: Determine whether all target units have been detected.

Determine whether all target units have been detected:

If the detection is completed, end the target secondary screening process;

Otherwise, obtain the next target unit information, and repeat steps 2 to 10.

The effect of the present invention can be further illustrated by the following simulation experiments:

1. Simulation experimental parameters

In this experiment, the flight height of the carrier aircraft is 8000m, the front antenna adopts a 16×1 uniform linear array, the array element spacing is 0.1092m, the carrier plane’s speed direction is parallel to the front axis, that is, the front side view array, and the main beam is directed to the front. The included angle of the axial direction is 90°, and the pitch angle of the main beam is 0°. The carrier frequency f _c is 1.24 GHz, the wavelength is 0.2419 m, the range sampling frequency f _s used by the radar is 1.25 MHz, the pulse repetition frequency is 1500 Hz, and the number of pulses is 64. We add a real target to the No. 300 range gate and No. 20 Doppler channel, and at the same time, randomly add 5 strong clutter points in the sidelobe clutter area.

2. Simulation content and result analysis

Set the simulation parameters, and compare and explain through three figures.

First, obtain the range Doppler map after PD processing, as shown in Figure 2, in which "+" marks the side lobe strong clutter point, and "☆" marks the real target point;

In simulation 1, CFAR detection is performed on the range Doppler map in Figure 2, and the CFAR detection result map is obtained. The result is shown in Figure 3, where "○" marks the target points detected by CFAR. It can be seen from Figure 3 that 4 of the 5 strong sidelobe clutter points are determined as targets by the CFAR detector. At the same time, the clutter near the height line is also judged as the target;

Simulation 2, using the present invention to perform secondary detection on the detection result graph of CFAR, the result is shown in Figure 4, in which "X" marks the point determined by the present invention as a false alarm. It can be seen from Figure 4 that, except for the real target point, other false alarm points caused by strong clutter are eliminated.

Through the above analysis, it can be concluded that the present invention can effectively eliminate false alarms caused by strong clutter points in the sidelobe region.

Claims (6)

1. A target secondary screening method based on multi-beam forming comprises the following steps:

(1) obtaining the distance R of the ith target unit_iAnd normalized Doppler frequencyM is the total number of target units;

(2) according to normalized Doppler frequencyJudging whether the ith target unit is positioned in a side lobe clutter area: if the position is in the side lobe clutter area, executing (3); otherwise, directly saving the unit and jumping to (9);

(3) according to ith target unit information R_iAndcalculating the pitch angle phi of the unit_iAnd have the same as the target unitAzimuth theta of ground clutter_i；

(4) Obtaining the azimuth theta of the main beam₀Calculating the main beam width delta theta to respectively obtain the azimuth angle phi of the target unit₀Angle phi of ground clutter_c：

Φ₀＝(θ₀-Δθ/2,θ₀+Δθ/2)

Φ_c＝(θ_i-Δθ/2,θ_i+Δθ/2)

Wherein, theta_iIs the azimuth of the ground clutter;

(5) angle phi of ground clutter_cAnd the azimuth angle phi of the target unit₀Forming K pointed beams in the angle areas except the angle areas, and calculating weight vectors of the K beams:

w＝[w₁,…,w_j,…,w_K]

wherein, w_jIs the weight vector of the jth beam, j ═ 1,2, …, K;

(6) obtaining a distance R_iTo raw echo data received by radarTo pairFourier transform of pulse dimension is carried out to obtain normalized Doppler frequency f_d,iCorresponding array element field data x_i，N is the number of array elements;

(7) calculating the output power of K beams according to the results of (5) and (6):

y＝[y₁,…,y_j,…,y_K]

wherein,is the output power of the jth beam, (-)^HIndicating the conjugate transpose of an element within brackets, | (. phi) & gtnon-visual cells²Represents the square of the absolute value of the element in parentheses;

(8) calculating the output power of the main beam according to the result of (6):

wherein, w₀Is the main beam azimuth angle theta₀A corresponding weight vector;

(9) calculating an absolute value deltay of a difference between the output power of the main beam and the average output power of the K beams according to the results of (7) and (8), and converting deltay into a decibel value deltay':

Δy'＝10log10(Δy)

(10) and (3) giving a secondary detection threshold zeta as 15, comparing the delta y with the zeta, and judging whether the target unit is a false alarm: if delta y is larger than or equal to zeta, the real target is judged, the target unit is stored, and the step (11) is executed; otherwise, judging as a false alarm, and executing (11);

(11) judging whether all target units are detected completely: if the detection is finished, finishing the secondary target discrimination process; otherwise, acquiring next target unit information, and repeating (2) to (10).

2. The method of claim 1,(2) in accordance with normalized Doppler frequencySize judgment ofWhether the corresponding target unit is located in the side lobe clutter zone or not is realized as follows:

2a) calculating the maximum normalized Doppler frequency of the sidelobe clutter:

wherein v ═ v_xv_yv_z]^TIs the velocity vector of the carrier, v_x，v_y，v_zRespectively representing 3 components of the speed v on the coordinate axis of the machine body; lambda [ alpha ]₀For the operating wavelength of the radar, f_rIs the pulse repetition frequency;

2b) determining a normalized Doppler frequency range of the side lobe clutter zone:

f_d,c＝[-f_{d,c_max},f_{d,c_max}]，

2c) determining normalized Doppler frequencyNormalized Doppler range f for sum side lobe clutter region_d,cThe relationship of (1): if it isThenThe corresponding target unit is positioned in a lobe area beside the clutter; otherwise, the target unit is located in the clear zone.

3. The method of claim 1, wherein the distance R in (3) is determined according to the ith target unit_iSum normalized dopplerLe frequencyCalculating the pitch angle phi of the target unit_iAnd have the same as the target unitAzimuth angle theta of ground clutter_iIt is implemented as follows:

3a) calculating the pitch angle phi of the ith target unit_i：

Wherein h is_aIs the flying height of the aircraft, a_eIs the earth equivalent radius;

3b) the calculation is the same as that of the target unitAzimuth theta of ground clutter_i：

Wherein A ═ v_xcosφ_i，B＝v_ycosφ_i，φ_iIs the pitch angle, v, of the ith target unit_x，v_y，v_zRespectively representing 3 components of the speed vector of the carrier on the coordinate axis of the body, lambda₀For the operating wavelength of the radar, f_rIs the pulse repetition frequency.

4. The method of claim 1, wherein the main beam width Δ θ is calculated in (4) as follows:

Δθ＝θ_3dB＝57.3°×0.886λ₀/N/d

wherein, N is the array element number, and d is the array element interval.

5. The method of claim 1, wherein the weight vector for the k-th beam direction is calculated in (5) as follows:

w_j＝[w_j,0,…,w_j,n,…w_j,N-1]^H，

wherein, w_j,n＝exp(j2πk_j·d_n)，k_j＝[sinθ_jcosφ_icosθ_jcosφ_isinφ_i]Line of sight vector of multi-beam, theta_jIs the azimuth angle of the jth beam, in the range of θ_j∈(-π/2,π/2)-(Φ₀∪Φ_c)；φ_iIs the pitch angle of the ith target unit; d_n＝[d_x,nd_y,nd_z,n]Is the position vector of the nth array element, d_x,n，d_y,n，d_z,nEach representing 3 components of the array element on the axis of the wavefront coordinate.

6. The method of claim 1, wherein the main beam azimuth θ is calculated in (8)₀Weight vector w of₀The formula is as follows:

w₀＝[w_0,0,…,w_0,n,…,w_0,N-1]^H，

wherein, w_0,n＝exp(j2πk₀·d_n)，k₀＝[sinθ₀cosφ_icosθ₀cosφ_isinφ_i]Is the line-of-sight vector of the main beam, θ₀Is the main beam azimuth angle phi_iIs the pitch angle of the ith target unit; d_n＝[d_x,nd_y,nd_z,n]Is the position vector of the nth array element, d_x,n，d_y,n，d_z,nEach representing 3 components of the array element on the axis of the wavefront coordinate.