CN101599932A - A kind of ultra-low side lobe pulse compression method - Google Patents

Abstract

The invention provides a kind of ultra-low side lobe pulse compression method, it is all to satisfy ULTRA-LOW SIDE LOBES waveform after the compression that radar system requires at every index, ask weighting windows in frequency domain inverse, can realize utilizing multiphase filtering to reduce the principle that data transfer rate is finished at frequency domain according to quadrature interpolation and pulse compression, propose based on multiphase filtering quadrature interpolation and pulse compression composition algorithm structure with quadrature interpolation.Adopt the present invention, under to the less condition of other index influences of pulse pressure, improved the major-minor ratio of pulse pressure, thereby can effectively reduce sidelobe level, reach the purpose that reduces the radar false alarm rate, the present invention can be widely used in the pulse-compression radars.

Description

A kind of ultra-low side lobe pulse compression method

Technical field

The invention belongs to the signal processing technology field, it relates to the ULTRA-LOW SIDE LOBES pulse compression technique.

Background technology

The range resolution of radar depends on signal bandwidth, and in the normal pulsed radar, the time wide bandwidth of radar signal is long-pending to be a constant (being about 1), therefore can not take into account range resolution and two indexs of speed resoluting force.Modern radar system extensively adopts pulse compression technique, it adopts the broad pulse emission to improve the average power of emission, guarantee enough maximum operating ranges, and when receiving, adopt corresponding pulse compression method to obtain burst pulse, improving range resolution, thereby can solve the contradiction between operating distance and the resolution capability preferably.

The outstanding feature of this system is:

(1) its broad pulse that transmits and adopt carrier frequency to change according to certain rules makes its pulse duration and effective product B τ＞＞1 of carrier frequency spectral width, and these two signal parameters are independently basically, thereby can be selected to satisfy the tactics requirement respectively.Under the limited condition of transmitter peak power, it has improved the average power P of transmitter _Av, strengthened the energy that transmits, therefore enlarged detection range.

(2) compression network that is complementary with the frequency spectrum that transmits is set in receiver, makes transmit (it is generally acknowledged it also is the echo-signal of receiver inlet) of broad pulse become burst pulse, therefore kept good range resolution.This processing procedure is referred to as " pulse compression ".

(3) help improving the antijamming capability of system.Concerning active noise disturbs, because signal bandwidth is very big, force jammer emission broadband noise, thereby reduced the spectrum density of disturbing., in the delay of signal, amplification, repeating process, can produce bigger distortion, thereby obtain certain inhibition also owing to adopted complicated pulse internal modulation repeater jamming.Disturb as for passiveness, then owing to improved the resolution capability of system, interference free performance also has some improvement.

Because these advantages of pulse compression system, the pulse compression system has become a kind of system of radar extensive use in modern age.

Common pulse compression system is to add weighting windows to form best mismatch network in the matched filter back, guarantees high as far as possible output signal-to-noise ratio when reaching the compression of pulse duration.Traditional weighting windows has rectangular function, Taylor's weighting, hamming weighting etc., though these classical weighting windows at compression ratio greater than 200 o'clock, sidelobe level also can only be compressed to-40dB about.So high secondary lobe might make strongly disturbing echo flood the echo of target fully, causes the inaccurate of range finding.

In recent years, the theory of digit pulse compress technique and algorithm become one research focus, various digit pulse compress algorithm emerge in an endless stream, this advanced greatly the digit pulse compress technique development.These algorithm major parts all center on improves pulse compression RMS (major-minor ratio), but does not have a kind of universal method that effectively is suitable for various pulse pressure signals to obtain big RMS so far.N.Vincent for example, J.Richard, N.Suinot has proposed a kind of nonlinear frequency modulation signal in " Verylow side-lobe level pulse compression for rain radar " literary composition, its RMS can reach more than the 60dB, RMS than conventional pulse pressure algorithm greatly improves, but it does not point out the improvement method of RMS in the pulse compression prevailingly.In " a kind of pulse compression signal side lobe suppression method " literary composition, proposed the new method of a kind of RMS of improvement again such as Xu Qing, Xu Jilin, yellow spiciness, but its effect not very remarkable.

In the conventional pulse compression radar receiver, adopt analog frequency mixing to orthogonally transform into base band, carry out digitlization again at intermediate frequency.The shortcoming of this system is to be difficult to guarantee the strict orthogonal of I, Q binary channels simulation local oscillator, disturbs thereby introduce very big image frequency in the orthogonal transform process.Along with the raising of A/D technology, carry out The digital quadrature transformation at intermediate frequency and become possibility.Common its signal processing front end of pulse-compression radars carries out the intermediate frequency digital sample earlier at present, how with the sample rate sampling greater than 4B (B is a signal bandwidth), carries out Digital Down Convert afterwards and obtains twin-channel baseband signal, carries out pulse pressure afterwards and handles.This processing structure, no matter pulse pressure is to carry out at frequency domain or time domain, and total operand is all bigger, and bigger operand has caused the complexity of system configuration and the raising of cost.By the research for digital quadrature-sampling and pulse compression, Su Tao, Johnson ﹠ Johnson are refined, Wu Shunjun has proposed a kind of composition algorithm efficiently in " highly effective algorithm of digital quadrature-sampling and pulse pressure and realization " literary composition.

Summary of the invention

In order to force down sidelobe level as far as possible, so that sidelobe level and noise level are similar, a kind of ultra-low side lobe pulse compression method of the present invention can effectively reduce sidelobe level, thereby reduces the false alarm rate of radar.

Content of the present invention for convenience of description, at first make following term definition:

Define 1 intermediate frequency pulse pressure signal x (t)=Re (f (t) * exp (j*2 π * f ₀)), f ₀Be resting frequence, f (t) is the base band reference signal of emission, and Fourier transform is expressed as F (ω), Re[] expression gets real part.

Define 2 bandpass samplings

The bandpass sampling rate is f _s=4f ₀/ 2m+1, and f _s＞2B, m are integer, and B is a signal bandwidth, f ₀Be resting frequence.

Define 3 multiphase filtering structures

If filter transfer function is H (z), impulse response is h (n), then $H\left(z\right)=\underset{k=0}{\overset{D-1}{\mathrm{\Σ}}}{z}^{-k}{E}_k\left(z^D\right),$

E _k(z ^D) be its branching filter, E so _k(z)=h (nD+k) z ^-n, wherein D is a branches.

Define 4 The digital quadrature transformation

In fact so-called The digital quadrature transformation is exactly earlier analog signal x (t) to be passed through to form Serial No. x (n) after the A/D sampled digitalization, then with two orthogonal local oscillation sequence cos (ω ₀N), sin (ω ₀N) (ω multiplies each other ₀=2 π f ₀/ f _s), obtain baseband complex signal by digital low-pass filtering again.

Define 5 time delay filtering

In the The digital quadrature transformation based on multiphase filtering, digital intermediate frequency signal is as follows:

$=x_{\mathrm{BI}}\left(n\right)\cos \left(\frac{2m+1}{2}\mathrm{\πn}\right)-x_{\mathrm{BQ}}\left(n\right)\sin \left(\frac{2m+1}{2}\mathrm{\πn}\right)---\left(1\right)$

Can get by following formula

x(2n)＝x _BI(2n)·(-1) ⁿ，x(2n+1)＝x _BQ(2n+1)·(-1) ⁿ (2)

Make x ' _BI(n)=x _BI(2n), x ' _BQ(n)=x _BQ(2n+1)

Then

x′ _BI(e ^jω)＝1/2x _BI(e ^jω/2)，x′ _BQ(e ^jω)＝1/2x _BQ(e ^jω/2)·e ^jω/2 (3)

Both digital spectral differ a delay factor e as can be known ^{J ω/2}So, carry out time delay filtering to sequence.

Defining 6 conjugation symmetry decomposes with antisymmetry

X (k) is done the conjugation symmetry to be decomposed as follows with antisymmetry:

X _ep(k)＝[X(k)+X ^*(-k)]/2，X _op(k)＝[X(k)-X ^*(-k)]/2

X wherein _Ep(k) be conjugate symmetric component, X _Op(k) be conjugate antisymmetric component.

It is f (t) that reference signal is established in definition 7, and its Fourier transform is expressed as F (ω)

It is H (ω)=kF that the matched filter frequency domain response is established in definition 8 ^*(ω) exp (j ω t ₀), F (ω) is the reference signal Fourier transform, t ₀For realizing matched filtering needed time of delay

It is W (ω) that the weighting function Fourier transform is established in definition 9

The definition 10 pulse pressure frequency domains of establishing target echo are output as Y (ω), and time domain is y (t)

The present invention proposes a kind of ultra-low side lobe pulse compression method, it comprises following steps:

The digitlization of step 1 echo signal of intermediate frequency

Echo signal of intermediate frequency to input carries out bandpass sampling, and divides two-way to extract the Serial No. of bandpass sampling output, obtains the two-way sequence, and as shown in Figure 2, the one tunnel is called I road sequence, and one the tunnel is called Q road sequence;

The preliminary treatment of step 2 Serial No.

The I that step 1 is obtained, the symbol of Q two-way Serial No. are revised, and as shown in Figure 2, wherein when m is even number, then select (1) ^P+1, m selects (1) during for odd number ^p, p=0,1,2,3 ..., m is an arbitrary integer, carries out bandpass sampling, obtains pretreated I way word sequence and Q way word sequence; Wherein I represents in-phase component, and Q represents quadrature component;

Step 3 time delay adjustment obtains the frequency spectrum behind the echo-signal shift frequency

Pretreated I, Q two-way Serial No. that step 2 is obtained carry out plural FFT, decomposite the corresponding respectively frequency spectrum of I, Q two-way Serial No., and Q way word sequence is carried out filtering wave by prolonging time obtain through Q way word sequence spectrum behind the filtering wave by prolonging time, with the frequency spectrum of I way word sequence with through the Q way word sequence spectrum addition behind the filtering wave by prolonging time, obtain the frequency spectrum behind the echo-signal shift frequency.Decomposite the corresponding respectively frequency spectrum of I, Q two-way Serial No. and Q way word sequence spectrum done the process of filtering wave by prolonging time as follows:

X′(k)＝DFT(x′ _BI(n)+j*x′ _BQ(n)) (6)

The X ' that obtains (k) is carried out following conjugation symmetry to be decomposed with antisymmetry

X′ _ep(k)＝[X′(k)+X′ ^*(-k)]/2，X′ _op(k)＝[X′(k)-X′ ^*(-k)]/2 (7)

Then

X′ _BI(k)＝DFT(x′ _Bl(n))＝X′ _ep(k)，X′ _BQ(k)＝DFT(x′ _BQ(n))＝-j*X′ _op(k) (8)

To X ' _BQ(k) carry out the frequency domain filtering wave by prolonging time, filter is H _Q(e ^{J ω/2})=e ^{-j ω/2}

Wherein X ' is (k) for I, frequency spectrum when Q two-way Serial No. is regarded sequence of complex numbers as; X ' ^*(k) be X ' conjugation (k); X ' _Ep(k) be X ' conjugate symmetric component (k), X ' _Op(k) be X ' conjugate antisymmetric component (k); X ' _BI(n) be I way word sequence, X ' _BI(k) be I way word sequence spectrum; X ' _BQ(n) be Q way word sequence, X ' _BQ(k) be Q way word sequence spectrum;

Step 4 is asked for ULTRA-LOW SIDE LOBES pulse compression frequency-domain result

Design a rectangular function, the length that makes rectangular function is less than 1/4 of the data sequence length of handling in the pulse pressure processing procedure, with this rectangular function and himself do convolution more than four times, computing obtains the result after the convolution, with the result after this convolution as ULTRA-LOW SIDE LOBES pulse compression frequency-domain result;

Step 5 designing filter

According to the ULTRA-LOW SIDE LOBES pulse compression frequency-domain result that obtains in the step 4, the anti-weighting function of asking, as shown in the formula:

$W\left(\mathrm{\ω}\right)=\frac{Y\left(\mathrm{\ω}\right)}{F\left(\mathrm{\ω}\right)*H\left(\mathrm{\ω}\right)}---\left(9\right)$

Wherein F (ω) is the reference signal frequency spectrum, and H (ω) is a matched filter, and then W (ω) promptly is required frequency domain weighting function.H (ω) W (ω) then constitutes the pulse pressure filter, and the low pass filter that provides in the actual engineering and pulse pressure filter are merged low pass filter after obtaining merging;

Step 6 frequency domain filtering

Utilize the low pass filter after the merging that obtains in the step 5,, obtain the frequency-domain result of pulse compression on the frequency domain frequency spectrum behind the step 3 resultant echo signal shift frequency being carried out filtering;

Step 7 pulse pressure result

The pulse compression frequency-domain result that obtains in the step 6 is carried out the frequency spectrum stack, afterwards the frequency spectrum after the stack is carried out IFFT, obtain final pulse compression result with ULTRA-LOW SIDE LOBES effect.

Need to prove:

When dividing two-way to extract in the step 1, wherein extract after mobile half sampling instant of one tunnel sequence, the sequence after the extraction is called Q road sequence again.

In the step 3 during, the frequency spectrum of Q road correspondence be taken advantage of addition again behind the j, be that is to say and regard Q road sequence as pure imaginary number the frequency spectrum addition of I, Q two-way Serial No..

The principle of determining Y (ω) in the step 4 is explained as follows: be-13.2dB according to the sidelobe level of Singh's function, then a plurality of Singh's functions multiply each other then that sidelobe level can reduce.A plurality of Singh's functions multiply each other and are equivalent to the convolution of corresponding rectangular function on frequency domain on the time domain.The selection of rectangular function length will be selected according to desired sidelobe level, and multiple convolution result (Shu Chu frequency spectrum just) will mainly concentrate in the signal bandwidth simultaneously, to reduce the snr loss.In order to satisfy the ultralow major-minor ratio and the main lobe broadening of pulse pressure simultaneously, must improve sample rate.In fact this method is exactly to make full use of the purpose that big frequency domain width reaches the compression secondary lobe from another perspective.Emulation proves, during less than 2 times of signal bandwidths, can not satisfy ULTRA-LOW SIDE LOBES and little main lobe broadening in sample rate simultaneously.More than 2 times the time, can satisfy ULTRA-LOW SIDE LOBES and little main lobe broadening greater than signal bandwidth when sample rate, the specific targets selection of can compromising.Choose the principle of the every index of pulse pressure according to compromise,, carry out the multiple convolution computing, obtain the Fourier transform Y (ω) of needed low secondary lobe pulse pressure output waveform rectangular function.

When designing rectangular function in the step 4, rectangular function length can not be too little, otherwise can cause the main lobe broadening of pulse pressure too big, do not satisfy requirement of engineering, and concrete length need by emulation experiment relatively screening repeatedly to be selected.

In the step 4, when the pulse pressure signal was phase-coded signal, the pulse pressure frequency spectrum function had zero point, can not directly retrodict and ask for weighting function, need do certain correction to spectral null, with zero point frequency spectrum replace with a less numerical value.The weighting function of retrodicting and trying to achieve because mismatch is comparatively serious, may cause the deterioration of output signal-to-noise ratio.When signal-to-noise ratio degradation is big, need carry out certain correction to weighting function, the Frequency point numerical value of having introduced big noise is increased.

The process of step 7 intermediate frequency spectrum stack is equivalent to extract on time domain, and the data transfer rate after therefore should guaranteeing to extract satisfies sampling thheorem.

Principle of the present invention is:

Classical weighting windows all is directly filtering to be carried out in the output of matched filter, and the weighting windows of the signal of different parameters is all identical.A kind of every index that proposes this algorithm earlier all satisfies the waveform after the compression that radar system requires, the counter again weighting windows of asking, and therefore to the signal of different parameters, its weighting windows is with incomplete same.The pulse pressure frequency domain output when if Y (ω) expression has target echo, the Fourier that a kind of every index that is promptly proposed all satisfies the waveform after the compression that radar system requires changes, the Fourier of F (ω) expression reference waveform changes, the frequency response of H (ω) expression matched filter, then the weighting windows function can be tried to achieve by following formula: $W\left(\mathrm{\ω}\right)=\frac{Y\left(\mathrm{\ω}\right)}{F\left(\mathrm{\ω}\right)*H\left(\mathrm{\ω}\right)}.$ Make the secondary lobe of y (t) very little in order to reach, the bandwidth requirement of Y (ω) is very high, then must improve sample rate, therefore can consider this algorithm application in the if sampling system.

When carrying out if digitization, establish input signal and be with bandpass sample theory:

F wherein ₀Be resting frequence, a (t) is an amplitude function, is generally rectangular pulse,

The expression signal phase.

Sample rate satisfies bandpass sample theory, for:

$f_s=\frac{4f_0}{2m+1}(m=\mathrm{0,1,2},...)---\left(11\right)$

Obtaining sample sequence is:

$=x_I\left(n\right)\cos \left(\frac{2m+1}{2}\mathrm{\πn}\right)-x_Q\left(n\right)\sin \left(\frac{2m+1}{2}\mathrm{\πn}\right)$

Can get by following formula, work as n=2p, p=0,1,2 ... the time

x(2p)＝x _I(2p)cos[(2m+1)πp]＝x _I(2p)(-1) ^p (13)

Work as n=2p+1, p=0,1,2 ... the time

Order:

Then: x ' _I(p)=x _I(2p), x ' _Q(p)=x _Q(2n+1)

Know by extracting principle, if x _I(p) and x _Q(p) digital spectral width is less than pi/2, and then its 2 times are extracted sequence x ' _I(p) and x ' _Q(p) can represent former sequence undistortedly.X ' _I(p) and x ' _Q(p) digital spectral is:

${x^{\′}}_I\left(e^{\mathrm{j\ω}}\right)=\frac{1}{2}{x}_I\left(e^{j\frac{\mathrm{\ω}}{2}}\right),{x^{\′}}_Q\left(e^{\mathrm{j\ω}}\right)=\frac{1}{2}{x}_Q\left(e^{j\frac{\mathrm{\ω}}{2}}\right)\·e^{j\frac{\mathrm{\ω}}{2}}---\left(15\right)$

The digital spectral that is to say both differs a delay factor exp (jw/2).Can adopt filter that its frequency spectrum is proofreaied and correct, the correcting filter frequency response of I, Q two-way (is respectively H _I(e ^{J ω}) and H _Q(e ^{J ω})) should satisfy:

$\frac{H_Q\left(e^{\mathrm{j\ω}}\right)}{H_I\left(e^{\mathrm{j\ω}}\right)}=e^{-j\frac{\mathrm{\ω}}{2}},$ And | H _Q(e ^{J ω}) |=| H _I(e ^{J ω}) |=1 (16)

Therefore quadrature interpolation can utilize multiphase filtering to realize, combines again because quadrature interpolation can realize in frequency domain all that with pulse compression and frequency domain realizes that operand greatly reduces, so with two parts and finishes, and operand further reduces.

Innovative point of the present invention is:

Propose the ULTRA-LOW SIDE LOBES waveform after a kind of every index all satisfies the compression that radar system requires, ask weighting windows in frequency domain inverse again.Because its frequency domain bandwidth of ULTRA-LOW SIDE LOBES waveform that this every index all satisfies after the compression that radar system requires is very big, so require big sample rate, obviously in baseband sampling, should must cause big data transfer rate by algorithm, improve hardware cost, so we are applied in the system of if sampling, can realize at frequency domain according to quadrature interpolation and pulse compression etc., and quadrature interpolation can utilize multiphase filtering to reduce the principle that data transfer rate is finished, quadrature interpolation and pulse compression composition algorithm structure have been designed based on multiphase filtering, this composition algorithm structure can be used the ULTRA-LOW SIDE LOBES pulse pressure algorithm that proposes previously easily, and its operand is littler than conventional method, and performance is better.

The technical problem that the present invention solves:

In the pulse compression system of routine, its major-minor generally only is about 40dB than (RMS), the RMS of the signal of small reduction ratio even be reduced to 30dB, so high secondary lobe can make strong target echo in the detections of radar flood weak target echo or cause the raising of false alarm rate, use the present invention and can improve the major-minor of pulse pressure greatly than (RMS), and less, thereby improved the performance of pulse pressure greatly to the influence of signal to noise ratio and main lobe width.

Advantage of the present invention is:

Under to the less condition of other index influences of pulse pressure, the major-minor that has improved pulse pressure is than (RMS), thereby the reduction sidelobe level reaches the purpose that reduces the radar false alarm rate.The present invention can be widely used in the pulse-compression radars.

Description of drawings

Fig. 1 the inventive method FB(flow block)

Fig. 2 the inventive method principle scheme block diagram.

F among the figure _sBe sample rate, f ₀Be resting frequence, m is an arbitrary integer, and z represents that digital signal moves a chronomere, 2 times of extractions of ↓ 2 expressions, p=0,1,2 ...If m is when being odd number, then select (1) among the figure ^P+1, otherwise select (1) ^p, A/D represents D/A converter module.

Embodiment

Institute of the present invention in steps, conclusion all on MATLAB 7.0 checking correct.Concrete implementation step is as follows:

The digitlization of step 1 echo signal of intermediate frequency

Intermediate frequency pulse pressure signal is carried out The digital quadrature transformation based on multiphase filtering.

The medium-frequency pulse echo parameter is: f ₀=15MHz widely during pulse is: T=50us, and bandwidth is: B=2MHz, then the intermediate frequency echo impulse can be expressed as:

X (t)=Rect (t/T) cos (2 π (f ₀+ f _d) t+kt ²), f _dBe Doppler frequency shift.

As shown in Figure 2, above-mentioned signal being carried out sample rate is f _sThe bandpass sampling of=20Mhz is decomposed into I, Q two-way Serial No. with sampled data;

The preliminary treatment of step 2 Serial No.

I, Q two-way Serial No. that step 1 is obtained carry out 2 times of extractions, and the symbol of two-way Serial No. is revised, and as Fig. 2, wherein for the Q Serial No., when m is even number, then take advantage of (1) ^P+1, when m is odd number, take advantage of (1) ^pI way word sequence is always taken advantage of (1) ^pP is a natural number;

Step 3 time delay adjustment obtains the frequency spectrum behind the echo-signal shift frequency

I, Q two-way Serial No. are carried out plural FFT, decomposite the corresponding respectively frequency spectrum (according to formula 6,7) in I road, Q road, and Q way word sequence carried out filtering wave by prolonging time at frequency domain, and take advantage of j to add I road frequency spectrum Q road frequency spectrum afterwards, obtain the frequency spectrum behind the echo-signal shift frequency;

Step 4 is asked for ULTRA-LOW SIDE LOBES pulse compression frequency-domain result Y (ω)

Design rectangular function W1 length is 65 points, and the ULTRA-LOW SIDE LOBES pulse compression frequency-domain result Y (ω) that then wishes to get is:

Y＝conv(conv(conv(conv(w1，w1)，conv(w1，w1))，conv(conv(w1，w1)，conv(w1，w1)))，conv(conv(conv(w1，w1)，conv(w1，w1))，conv(conv(w1，w1)，conv(w1，w1))))；

Wherein Conv () represents convolution algorithm;

Step 5 designing filter

The ULTRA-LOW SIDE LOBES pulse compression result's who obtains according to step 4 counter the pushing away of Fourier transform Y (ω) obtains the Sidelobe Suppression weighting function, and method is w=Y./(H.*conj (H)), and then the pulse pressure filter is H.*W;

The low pass filter that provides in the actual engineering and the pulse pressure filter of trying to achieve are above merged low pass filter after obtaining merging;

Step 6 frequency domain filtering

The echo-signal frequency spectrum that utilizes the low pass filter after the merging of step 5 design on frequency domain step 3 to be obtained carries out filtering, obtains the pulse compression result on frequency domain;

Step 7 is tried to achieve the pulse pressure result

The frequency-region signal that obtains in the step 6 is done the frequency spectrum stack, be equivalent on time domain, do extraction, carry out IFFT afterwards, obtain the pulse pressure result.

The frequency-domain result 1-256 point, 513-768 point, 1025-1280 point, the 1537-1792 point that obtain are superposed to 256 point data, with remaining 1024 aliasings is other 256 point data, in that two parts frequency spectrum is formed a complete frequency spectrum, carry out 512 IFFT, this moment, data transfer rate became 2.5MHz.

The invention has the advantages that and reduced the sidelobe level in the pulse compression.

It is right by ultra-low side lobe pulse compression method proposed by the invention and traditional impulse compression method are done Ratio can be found out, the sidelobe level after traditional impulse compression method pulse pressure can only be suppressed to-40dB about, And the sidelobe level after the ultra-low side lobe pulse compression method pulse pressure proposed by the invention can be suppressed to-52dB or Lower, concrete numerical value depends on the relative size of sample rate. Therefore, compare with traditional impulse compression method, Adopt technology of the present invention can effectively reduce the false alarm rate of radar.

Claims (1)

1, a kind of ultra-low side lobe pulse compression method is characterized in that it comprises following steps:

The digitlization of step 1 echo signal of intermediate frequency

Echo signal of intermediate frequency to input carries out bandpass sampling, and divides two-way to extract the Serial No. of bandpass sampling output, obtains the two-way sequence, and one the tunnel is called I road sequence, and one the tunnel is called Q road sequence;

The preliminary treatment of step 2 Serial No.

The I that step 1 is obtained, the symbol of Q two-way Serial No. are revised, and wherein when m is even number, then select (1) ^P+1, m selects (1) during for odd number ^p, p=0,1,2,3 ..., m is an arbitrary integer, carries out bandpass sampling, obtains pretreated I way word sequence and Q way word sequence; Wherein I represents in-phase component, and Q represents quadrature component;

Step 3 time delay adjustment obtains the frequency spectrum behind the echo-signal shift frequency

Pretreated I, Q two-way Serial No. that step 2 is obtained carry out plural FFT, decomposite the corresponding respectively frequency spectrum of I, Q two-way Serial No., and Q way word sequence is carried out filtering wave by prolonging time obtain through Q way word sequence spectrum behind the filtering wave by prolonging time, with the frequency spectrum of I way word sequence with through the Q way word sequence spectrum addition behind the filtering wave by prolonging time, obtain the frequency spectrum behind the echo-signal shift frequency; Decomposite the corresponding respectively frequency spectrum of I, Q two-way Serial No. and Q way word sequence spectrum done the process of filtering wave by prolonging time as follows:

X′(k)＝DFT(x′ _BI(n)+j*x′ _BQ(n)) (6)

The X ' that obtains (k) is carried out following conjugation symmetry to be decomposed with antisymmetry

X′ _ep(k)＝[X′(k)+X′ ^*(-k)]/2，X′ _op(k)＝[X′(k)-X′ ^*(-k)]/2 (7)

Then

X′ _BI(k)＝DFT(x′ _BI(n))＝X′ _ep(k)，X′ _BQ(k)＝DFT(x′ _BQ(n))＝-j*X′ _op(k)(8)

To X ' _BQ(k) carry out the frequency domain filtering wave by prolonging time, filter is H _Q(e ^{J ω/2})=e ^{-j ω/2}

Wherein X ' is (k) for I, frequency spectrum when Q two-way Serial No. is regarded sequence of complex numbers as; X ' ^*(k) be X ' conjugation (k); X ' _Ep(k) be X ' conjugate symmetric component (k), X ' _Op(k) be X ' conjugate antisymmetric component (k); X ' _BI(n) be I way word sequence, X ' _BI(k) be I way word sequence spectrum; X ' _BQ(n) be Q way word sequence, X ' _BQ(k) be Q way word sequence spectrum;

Step 4 is asked for ULTRA-LOW SIDE LOBES pulse compression frequency-domain result

Design a rectangular function, the length that makes rectangular function is less than 1/4 of the data sequence length of handling in the pulse pressure processing procedure, with this rectangular function and himself do convolution algorithm more than four times, obtain the result after the convolution, with the result after this convolution as ULTRA-LOW SIDE LOBES pulse compression frequency-domain result;

Step 5 designing filter

According to the ULTRA-LOW SIDE LOBES pulse compression frequency-domain result that obtains in the step 4, the anti-weighting function of asking, as shown in the formula:

$W\left(\mathrm{\ω}\right)=\frac{Y\left(\mathrm{\ω}\right)}{F\left(\mathrm{\ω}\right)*H\left(\mathrm{\ω}\right)}---\left(9\right)$

Wherein F (ω) is the reference signal frequency spectrum, and H (ω) is a matched filter, and then W (ω) promptly is required frequency domain weighting function; H (ω) W (ω) then constitutes the pulse pressure filter, and low pass filter and the pulse pressure filter that provides in the actual engineering merged the low pass filter after obtaining merging;

Step 6 frequency domain filtering

Utilize the low pass filter after the merging that obtains in the step 5,, obtain the frequency-domain result of pulse compression on the frequency domain frequency spectrum behind the step 3 resultant echo signal shift frequency being carried out filtering;

Step 7 pulse pressure result

The pulse compression frequency-domain result that obtains in the step 6 is carried out the frequency spectrum stack, afterwards the frequency spectrum after the stack is carried out IFFT, obtain final pulse compression result with ULTRA-LOW SIDE LOBES effect.

Images