CN106054195A - Turbulence spectrum width estimation method based on space-time optimal processor - Google Patents

Abstract

The invention discloses a turbulence spectrum width estimation method based on a space-time optimal processor. The method comprises steps: modeling is carried out on turbulence echoes for an airborne pulse Doppler weather radar under a phased array system, and thus, weather radar echo data for a turbulence field are acquired; a generalized space steering vector and a generalized time steering vector applied to the turbulence field are built respectively to obtain the space-time steering vector; according to the space-time steering vector obtained in the second step, the space-time optimal processor is built, the radar echo data are processed, interference generated by a non-weather factor is inhibited, unchanged radar echo power caused by a turbulence target can be ensured, and the turbulence spectrum width is estimated; and echo data for all distance units in a radar working range are processed sequentially, and a speed spectrum width estimation result for each distance unit is estimated and obtained. The method of the invention can effectively inhibit the spectrum spread interference generated by the non weather factor, the turbulence spectrum width can be accurately estimated, and a simulation experiment verifies the effectiveness of the method.

Description

Based on the turbulent flow spectrum width method of estimation of optimal processor during sky

Technical field

The invention belongs to Radar Signal Processing Technology field, particularly relate to a kind of based on the turbulent flow of optimal processor during sky Spectrum width method of estimation.

Technical background

Turbulent flow refers to the continuous random pulsation being superimposed upon on average wind, is a kind of atmospheric perturbation being frequently encountered by flight course Phenomenon, flowing generally quick by air, irregular causes.This turbulent flow easily makes aircraft generation jolt, and even makes it significantly Degree standoff, therefore totally unfavorable to flight safety.Airborne weather radar can be certain with front, explorer vehicle air route In sector including turbulent flow, wind shear, thunderstorm etc. including dangerous weather region, provide the orientation of harm weather and strong to pilot The information such as degree, using as early warning and the reference of avoidance deathtrap.

For airborne weather radar, turbulent flow is the meteorological target that a kind of particle speed deviation is bigger.Velocity deviation can be managed Solving the fluctuation range for speed or spectrum width, spectrum width is the biggest, and turbulence intensity is the biggest.Turbulent flow detection at present generally utilizes estimates echo spectrum Wide and with detection threshold contrast method realizes, it can be seen that, the whether accurate of spectrum width estimated result can directly affect detection The quality of energy.Therefore, the accuracy improving the estimation of turbulent flow spectrum width as far as possible has the dangerous meteorological of turbulent flow to effectively detection and early warning Region is the most necessary.

During turbulent flow detects, the motion of turbulent flow target not causes the single factor of spread spectrum, antenna azimuth Also can cause video stretching with non-meteorological factors such as antenna beamwidths, thus affect the accurate of true turbulent flow spectrum width estimated result Degree.

At present, it is usually used in the method that spectrum width is estimated and turbulent flow detects and mainly has pulse based on time-domain analysis to method (Pulse-pair Processing, PPP) and fast Fourier transform (Fast Fourier based on frequency-domain analysis Transformation, FFT) method etc..These methods, while calculating is simple and better performances under the conditions of high s/n ratio, but work as When existence interference or signal to noise ratio are relatively low, its spectrum width estimates that performance drastically declines, and these methods all do not consider by the non-meteorological factor The spectrum width extension caused, easily causes spectrum width true to turbulent flow and crosses estimation.

Compared with traditional single antenna system, the antenna array of the Pulse Doppler Weather Radar of phased array system is by multiple Array element forms, and the phase place of each array element is controlled, and beam position is flexible, and its echo-signal comprises the spatial sampling information of target.Logical Cross spatial domain and the time-domain information making full use of signal, the spread spectrum caused can be carried out self adaptation and press down during radar scanning System, it is possible to accurate detection to target is better achieved.

Summary of the invention

In order to solve the problems referred to above, it is an object of the invention to provide a kind of based on the turbulent flow spectrum width of optimal processor during sky Method of estimation.

In order to achieve the above object, what the present invention provided includes based on the turbulent flow spectrum width method of estimation of optimal processor during sky The following step carried out in order:

1) the turbulent flow echo of the airborne pulse Doppler weather radar under phased array system is modeled, thus obtains rapids The meteorological radar echo data in flow field；

2) spatially and temporally distribution character based on turbulent flow target, structure is applicable to the generalized space guiding of field of turbulent flow respectively Vector Generalized Time steering vector, thus obtain steering vector during its sky；

3) utilize space-time adaptive handling principle, integrating step 2) in structure empty time steering vector, optimum when structure is empty Processor, process step 1) in radar return data, suppression the non-meteorological factor produce interference ensure by turbulent flow target simultaneously The power of the radar return caused is constant, and estimates turbulent flow spectrum width；

4) step 2 is repeated) to step 3), in processing radar operating range successively, the echo data of all distance unit, estimates Meter obtains the speed spectrum width estimated result of each distance unit.

It is the most according to claim 1 based on the turbulent flow spectrum width method of estimation of optimal processor during sky, it is characterised in that: In step 3) in, described utilizes space-time adaptive handling principle, integrating step 2) in structure empty time steering vector, structure sky Time optimal processor, process step 1) in radar return data, suppression the non-meteorological factor produce interference ensure by rapids simultaneously The power of the radar return that stream target causes is constant, and the method estimating turbulent flow spectrum width is: analyzes and causes turbulent flow echo spectrum width The factor of extension and the impact of spectrum width estimated result true on turbulent flow thereof, structure is applicable to optimal processing during turbulent flow target empty Device, is filtered radar return processing, and the spectrum width extension finally caused the non-meteorological factor suppresses, and ensures by rapids simultaneously The power of the radar return that stream target causes is constant, and utilizes the non-coupled characteristic estimating of Doppler frequency and spectrum width to go out normal-moveout spectrum Wide.

What the present invention provided is the machine for phased array system based on the turbulent flow spectrum width method of estimation of optimal processor during sky Airborne weather radar, distributed meteorological target property based on turbulent flow, utilize space-time adaptive handling principle to construct optimal processor, Estimate turbulent flow spectrum width.The inventive method can suppress the interference that the non-meteorological factor produces, and relatively accurately estimates turbulent flow spectrum width, emulation The experimental verification effectiveness of this method.

Accompanying drawing explanation

Fig. 1 is the geometry observation figure of turbulent flow.

Fig. 2 (a), (b) are respectively the space-time two-dimensional spectrogram of radar turbulence signal, and wherein Fig. 2 (a) is top view, Fig. 2 (b) For 3-D view.

The Space-time domain response diagram of steering vector when Fig. 3 is point target empty.

The Space-time domain response diagram of steering vector when Fig. 4 is distributed meteorological target empty.

Fig. 5 is the space-time two-dimensional spectrogram of No. 75 distance unit turbulent flow wind field echo.

Fig. 6 is the inventive method and the traditional pulse spectrum width estimated result comparison diagram to method.

Specific implementation method

Enter based on the turbulent flow spectrum width method of estimation of optimal processor during sky below by what the present invention was provided by instantiation Row describes in detail.

The present invention provide based on the turbulent flow spectrum width method of estimation of optimal processor during sky include carrying out in order following Step:

1) the turbulent flow echo of the airborne pulse Doppler weather radar under phased array system is modeled, thus obtains rapids The meteorological radar echo data in flow field；

The flight speed assuming airborne pulse Doppler weather radar (hereinafter referred to as radar) is V_a, along course vertical direction Placing N unit even linear array, pulse recurrence frequency is f_r, Coherent processing umber of pulse is K, launches a length of λ of impulse wave.

In the present invention, x_lRepresent l (l=1,2 ..., L) NK × 1 dimension of individual distance unit empty time fast beat of data, its table Reach formula as follows:

x_l=s_l+n_l (1)

Wherein, s_l、n_lSnap and noise when representing the turbulent flow sky of the l distance unit respectively, it is assumed that noise is additive Gaussian White noise.

For the field of turbulent flow in the l distance unit, its sampled data can be write as the matrix of a N × K by radar S_l.Wherein, matrix S_lLine n, kth column element s_l(n, k) represent radar n-th (n=1,2 ... N) individual array element, kth (k=1, 2 ... K) the individual pulse sampled data to the l distance unit, total Q in the range of the beam of radar in this distance unit During individual meteorological scattering particles, its expression is as follows:

WhereinWithRepresent respectively q (q=1,2 ..., Q) individual The Space Angle frequency of meteorological scattering particles and time angular frequency, θ_q、Represent that this meteorology scattering particles is relative to radar respectively Azimuth and the angle of pitch, R_qFor the oblique distance of q-th meteorology scattering particles with the aircraft arranging radar,For radar antenna Receiving pattern, v_qRepresent the q-th meteorology scattering particles radial velocity relative to radar.

By matrix S above_lLaunch to become NK × 1 dimensional vector, snap s when being field of turbulent flow sky_l.Then radar full distance Echo-signal in unit can be expressed as:

X=[x₁ x₂ … x_L]^T (3)

Doppler velocity spectrum width is to characterize different size of doppler velocity in radar beam range of exposures to deviate it average The degree of value, actually it is to be had different radial velocities by scattering particles to cause, radial velocity v_qDisperse is in a certain Near heart speed, it it is the principal element affecting speed spectrum width.But, during radar scanning, when scanning angle exists certain During broadening, will also result in spread spectrum, discounting for the video stretching thereby resulted in, the mistake of spectrum width true to turbulent flow can be caused Estimate.

As it is shown in figure 1, radar is with constant flight speed V_aAlong X-axis with rectilinear flight, radar antenna azimuth is α a, that Scattering particles J static for some in the range of beam, it is v relative to the radial velocity of radar_q=V_a, how general Le frequency displacement is:

$F\left(\mathrm{\α}\right)=\frac{2V_a}{\mathrm{\λ}}cos\mathrm{\α}---\left(4\right)$

Wherein, α is the azimuth of scattering particles J, and λ is for launching pulse wavelength.By radar antenna beam angle Δ α, radar Antenna azimuth α_aThe spread spectrum caused can be expressed as:

${\mathrm{\ΔF}}_{ant}=F({\mathrm{\α}}_a-\frac{\mathrm{\Δ}\mathrm{\α}}{2})-F({\mathrm{\α}}_a+\frac{\mathrm{\Δ}\mathrm{\α}}{2})\≈\frac{2V_a\mathrm{\Δ}\mathrm{\α}}{\mathrm{\λ}}{\mathrm{sin\α}}_a---\left(5\right)$

Use σ_aRepresent corresponding speed spectrum width, then have:

${\mathrm{\σ}}_a={\mathrm{\ΔF}}_{ant}\frac{\mathrm{\λ}}{2}=V_a{\mathrm{\Δ\αsin\α}}_a---\left(6\right)$

If by radar antenna beam angle Δ α, radar antenna azimuth angle alpha_aDeng the non-meteorological factor with turbulent flow to echo wave speed The contribution approximation of spectrum width is regarded as separate, then the speed spectrum width σ of field of turbulent flow radar return_vIt is represented by:

${\mathrm{\σ}}_v\≈\sqrt{{{\mathrm{\σ}}_T}^2+{{\mathrm{\σ}}_a}^2}---\left(7\right)$

Wherein, σ_T ²Represent the normal-moveout spectrum variance of turbulent flow.When echo-signal is processed, discounting for radar scanning By radar antenna beam angle Δ α, radar antenna azimuth angle alpha in journey_aThe video stretching caused Deng the non-meteorological factor, returns radar The speed spectrum width σ of ripple_vEstimated value regard turbulent flow spectrum width as, then work as σ_v>σ_TTime, spectrum width true to turbulent flow can be caused to cross estimation. When carrying out turbulent flow detection, it is easily caused the generation of false-alarm.Therefore, it is necessary to consider the above-mentioned interference factor shadow to estimated result Ring.

2) spatially and temporally distribution character based on turbulent flow target, structure is applicable to the generalized space guiding of field of turbulent flow respectively Vector Generalized Time steering vector, thus obtain steering vector during its sky；

A) using the width of radar main lobe as the prior information of field of turbulent flow in the range of radar illumination, set up and be applicable to turbulent flow etc. The generalized space steering vector of distributed object.

When radar main lobe direction center hold angle is θ_i, the center angle of pitch isTime, if field of turbulent flow is wide in its range of exposures Justice steric direction vector isIts expression formula is as follows

Wherein,Steric direction vector for point target；For definitiveness angle signal density function, By turbulent flow target at center hold angle θ in the present invention_iWith the center angle of pitchOn extension be expressed as:

Wherein,σ_θ、Represent center hold angle θ respectively_i, center The angle of pitchAngle spread on direction.

B) Gaussian distribution feature based on weather echo, structure is applicable to the Generalized Time of turbulent flow distributed target and guides Vector.

The radar return of turbulent flow is formed by stacking by substantial amounts of scattering particles echo, and each scattering particles has random phase Position, and between scattering particles, there is relative motion, therefore there is spread spectrum in radar return.From central limit theorem, greatly The superposition of amount scattering particles scattering electric field can get a Gaussian statistics signal.Therefore, typically by the merit of the weather echos such as turbulent flow Rate spectrum is modeled as Gaussian spectrum, and the signal that power spectrum is Gauss distribution can decline by introducing Gauss in time domain Doppler signal Subtract and obtain.Thus can obtain the Generalized Time steering vector that can describe field of turbulent flow distributed meteorology target:

s_t(f_d,σ_f)_K×1=v_t(f_d)⊙g_t(σ_f) (10)

Wherein, f_d=2v/ λ represents Doppler frequency, v_t(f_d) represent that the time of the point target that radial velocity is v guides arrow Amount；σ_fRepresent the Doppler width of signal, g_t(σ_f) representing frequency spreading function, can be expressed as follows respectively:

$\begin{array}{c}{v}_t\left(f_d\right)={\left[\begin{array}{cccc}1& e^{j2\mathrm{\π}\·\frac{f_d}{f_r}}& \mathrm{...}& e^{j2\mathrm{\π}(K-1)\·\frac{f_d}{f_r}}\end{array}\right]}_{K\×1}^T\\ g_t\left({\mathrm{\σ}}_f\right)={\left[\begin{array}{cccc}1& e^{-2{\mathrm{\π}}^2\·{\mathrm{\σ}}_f^2}& \mathrm{...}& e^{-2{\mathrm{\π}}^2\·{(K-1)}^2\·{\mathrm{\σ}}_f^2}\end{array}\right]}_{K\×1}^T\end{array}---\left(11\right)$

Further the generalized space steering vector of the field of turbulent flow of gained and Generalized Time steering vector are Kronecker Long-pending, steering vector when can obtain its sky:

3) utilize space-time adaptive handling principle, integrating step 2) in structure empty time steering vector, optimum when structure is empty Processor, process step 1) in radar return data, suppression the non-meteorological factor produce interference ensure by turbulent flow target simultaneously The power of the radar return caused is constant, and estimates turbulent flow spectrum width；

Definition power factor is Z, and its expression formula is as follows:

$Z\left(w\right)=\frac{w^{H}R(f_d,{\mathrm{\σ}}_f)w}{w^{H}R(f_d,0)w}---\left(13\right)$

Wherein, w represents the weight vector of optimal processor；w^HR(f_d,0)w、w^HR(f_d,σ_f) w represents the Doppler of signal respectively Spectrum width σ_fThe output of optimal processor, R (f during different value_d,σ_f) represent radar return theoretical covariance matrix.Signal Doppler width σ_fWhen=0, R (f_d, 0) only with Doppler frequency f_dRelevant, spread spectrum now is due to radar scanning In journey, the common effect of the non-meteorological factor causes, by minimizing the output w of this partial echo^HR(f_d, 0) and w, permissible Suppression is by radar antenna beam angle Δ α and radar antenna azimuth angle alpha_aCommon effect turbulent flow spectrum width estimated result is produced Interference.R(f_d,σ_f) can be obtained by following formula:

R(f_d,σ_f)=S (f_d,σ_f)S^H(f_d,σ_f) (14)

Find the weight vector w of optimal processor, in the case of the output ensureing turbulent flow target echo is constant, minimum Changing the spectrum width extension caused by the non-meteorological factor, be equivalent to so that power factor Z maximizes, now this optimal processor can be with such as Lower mathematical optimization problem describes:

$\left\{\begin{array}{c}min{w}^{H}R(f_d,0)w\\ st.w^{H}R(f_d,{\mathrm{\σ}}_f)w=1\end{array}\right.---\left(15\right)$

According to broad sense CAPON criterion, solve the weight vector obtaining optimal processor:

W=p{R^-1(f_d,0)R(f_d,σ_f)} (16)

Wherein, p{ } represent solution matrix eigenvalue of maximum characteristic of correspondence vector.Use x_iRepresent distance unit to be detected Field of turbulent flow receive data, then the output signal of optimal processor is:

Y=w^Hx_i (17)

By the Doppler width σ of signal_fBe converted to speed spectrum width σ_v=σ_f, then there is w=p{R λ/2^-1(f_d,0)R(f_d, σ_f)}.Owing to Doppler's average frequency does not couples with Doppler width, the estimation of average frequency can be entered independent of spectrum width OK, and spectrum width estimate must carry out in conjunction with the valuation of average frequency.According to this thought, solving the weight vector of optimal processor During w, can first fix any spectrum width(C is the constant more than zero, unit: m/s), estimating Doppler frequency, then estimate Doppler width, to reduce computational complexity.

Obtain the average Doppler frequency estimation of distance unit to be detectedAfterwards, can estimating Doppler spectrum width.Work as merit When rate factor Z maximizes, represent that optimal processor is optimal to the suppression of interference factor and the matching effect of turbulence signal, try to achieve The weight vector w of optimal processor, output signal w^Hx_iSpectrum width corresponding to power maximum point be turbulent flow in distance unit to be detected The Doppler width estimated value of signal, its estimated result is:

${\widehat{\mathrm{\σ}}}_{v,i}=\underset{{\mathrm{\σ}}_{v,i}}{\arg \max }|w^H{x}_i{|}_{f={\hat{f}}_i}---\left(18\right)$

4) step 2 is repeated) to step 3), in processing radar operating range successively, the echo data of all distance unit, estimates Meter obtains the speed spectrum width estimated result of each distance unit.

What the present invention provided can be by following imitative based on the effect of the turbulent flow spectrum width method of estimation of optimal processor during sky True result further illustrates.

Simulation parameter is arranged: arrange aircraft flight speed V of radar_a=200m/s, flying height H=8000m, field of turbulent flow Being distributed at the 9-21km of radar front, radar antenna is array number N=8, the desired homogeneous linear array of array element distance d=λ/2, radar Operation wavelength λ=0.032m, Coherent processing umber of pulse K=16, pulse recurrence frequency is f_r=1500Hz, azimuth is 60 °, bows The elevation angle is 0 °, and beam angle is 3 °, minimum distinguishable distance 150m, signal to noise ratio 20dB.

Fig. 2 (a), (b) are respectively the space-time two-dimensional spectrogram (overlooking and 3-D view) of radar turbulent flow echo.Turbulent flow is distribution Formula target, the scattering particles disperse in field of turbulent flow is in bigger spatial dimension, figure it is seen that its echo-signal is at sky Between distribution on there is certain extension；Simultaneously as field of turbulent flow inscattering number of particles is more, and scattering particles does irregular fortune Dynamic, velocity attitude changes drastically, and the fluctuation range of velocity magnitude is relatively big, and its echo-signal exists bigger expansion in frequency distribution Exhibition, thus cause Doppler frequency spectrum broadening.

The Space-time domain response diagram of steering vector when Fig. 3 is point target empty；Fig. 4 is the sky of turbulent flow distributed meteorology target Time steering vector Space-time domain response diagram；Fig. 5 is the space-time two-dimensional spectrogram of No. 75 distance unit turbulent flow echo.It can be seen that The present invention carried for steering vector during the distributed meteorological target of turbulent flow empty can the preferably actual turbulence signal of matching, make The steering vector mismatch error become is less.

Fig. 6 is the inventive method and the traditional pulse spectrum width estimated result comparison diagram to method.Under equal conditions, tradition arteries and veins Punching does not considers the spectrum width caused in radar antenna scanning process due to radar antenna beam angle, radar antenna azimuth etc. to method Extension, estimated result and spectrum width true value have relatively large deviation (average deviation is about 0.55m/s).The inventive method is then being composed The spectrum width extension inhibiting the non-meteorological factor to cause before wide estimation, to the doppler velocity spectrum width estimated result of each range gate relatively Accurately, less with true value deviation (average deviation is about 0.05m/s), so being better than traditional method.

Claims (2)

1. one kind based on the turbulent flow spectrum width method of estimation of optimal processor during sky, it is characterised in that described spectrum width method of estimation The following step including carrying out in order:

1) the turbulent flow echo of the airborne pulse Doppler weather radar under phased array system is modeled, thus obtains field of turbulent flow Meteorological radar echo data；

2) spatially and temporally distribution character based on turbulent flow target, structure is applicable to the generalized space steering vector of field of turbulent flow respectively With Generalized Time steering vector, thus obtain steering vector during its sky；

3) utilize space-time adaptive handling principle, integrating step 2) in structure empty time steering vector, optimal processing when structure is empty Device, process step 1) in radar return data, suppression the non-meteorological factor produce interference ensure to be caused by turbulent flow target simultaneously The power of radar return constant, and estimate turbulent flow spectrum width；

4) step 2 is repeated) to step 3), in processing radar operating range successively, the echo data of all distance unit, estimates Speed spectrum width estimated result to each distance unit.

It is the most according to claim 1 based on the turbulent flow spectrum width method of estimation of optimal processor during sky, it is characterised in that: in step Rapid 3), in, described utilizes space-time adaptive handling principle, integrating step 2) in structure empty time steering vector, when structure is empty Excellent processor, process step 1) in radar return data, suppression the non-meteorological factor produce interference ensure by turbulent flow mesh simultaneously The power of the radar return that mark causes is constant, and the method estimating turbulent flow spectrum width is: analyzes and causes turbulent flow echo spectrum width to extend Factor and the impact of spectrum width estimated result true on turbulent flow, structure is applicable to optimal processor during turbulent flow target empty, right Radar return is filtered processing, and the spectrum width extension finally caused the non-meteorological factor suppresses, and ensures by turbulent flow mesh simultaneously The power of the radar return that mark causes is constant, and utilizes the non-coupled characteristic estimating of Doppler frequency and spectrum width to go out speed spectrum width.