CN102841337A - Method for removing non-linear phase pollution from sky wave OTHR (over-the-horizon radar) echo signal - Google Patents

Abstract

The invention provides a method for removing non-linear phase pollution from a sky wave OTHR (over-the-horizon) radar echo signal, which relates to the field of the method for removing the non-linear phase pollution from the sky wave OTHR (over-the-horizon radar) echo signal and aims at solving the problems that the existing non-linear phase pollution has an effect on the sky wave OTHR echo signal and further reduces detecting performance under all kinds of conditions, especially under the lower signal-to-noise ratio. The method provided by the invention comprises the following steps that a data acquisition module collects to-be-processed sky wave OTHR time domain data; a data selecting module converts the to-be-processed sky wave OTHR time domain data; a pollution function computing module obtains a non-linear phase pollution function; and a signal correcting module obtains the sky wave OTHR echo signal after being subjected to removing the non-linear phase pollution. The method provided by the invention solves the problem of influence caused by the non-linear phase pollution in the sky wave OTHR radar echo signal, obviously reduces errors and has better effect under the lower signal-to-noise ratio. The method provided by the invention is used for processing the sky wave OTHR radar echo signal.

Description

A kind of sky-wave OTH radar echoed signal is removed the nonlinear phase pollution method

Technical field

A kind of sky-wave OTH radar echoed signal of the present invention is removed the nonlinear phase pollution method, relates to sky-wave OTH radar echoed signal decontamination method field.

Background technology

High frequency sky-wave OTH radar (OTHR) utilizes ionosphere that electromagnetic reflex is come the detection of a target, and its maximum detectable range can reach more than the 3000km.OTHR has that monitoring range is very big, pre-warning time is long, can prevent the low latitude ultra-low altitude penetration, can find advantage such as stealthy target.But OTHR still exists some defectives and problem needs to solve.Because the ionosphere medium receives various such environmental effects, be a unsettled propagation medium, so it has very large influence to the final result of detection of OTHR.The ionosphere problem of OTHR major concern is the influence that ionosphere produces echo spectrum, and the electromagnetic wave that its influence is mainly reflected in through ionospheric propagation may produce the variation on the phase place, and phase place is polluted and is divided into linear phase pollution and nonlinear phase pollution.The former generally is present under the situation of coherent accumulation in short-term, mainly causes the doppler spectral frequency shift (FS), can not influence the target detection performance; And the latter mainly occurs under the situation of long period coherent accumulation, has a strong impact on the coherent accumulation effect of information, causes dopplerbroadening, is easy to make weak echo signal to be influenced even flood by strong noise signal.To being used for the sky-wave OTH radar of Ship Target detection at a slow speed, generally all need carry out long coherent accumulation in order to improve signal to noise ratio (S/N ratio), but this generation that has also caused nonlinear phase to pollute, and then have a strong impact on detection performance.

The correction that nonlinear phase pollutes mainly contains three types: 1, the nonlinear phase pollution is converted into linear phase and pollutes the method for handling again, it mainly is represented as characteristic value decomposition method and maximum entropy method (MEM); 2, the instantaneous frequency of estimated signal is carried out integration again to obtain the method that phase place is polluted, and it mainly is represented as pseudo-distribution and the phase gradient method (PGA) received of tieing up; 3, use the undetermined coefficient polynomial expression to come analogue phase to pollute the method for function, it mainly is represented as minimum entropy search procedure and polynomial expression modeling.These methods have suppressed the ionosphere phase place to a certain extent to be polluted, but same some problems that exist, and for example when signal to noise ratio (SCR) was lower, result can degenerate even algorithm lost efficacy.

Summary of the invention

The purpose of this invention is to provide a kind of sky-wave OTH radar echoed signal and remove the nonlinear phase pollution method; To solve under the situation that especially signal to noise ratio is lower under the various situation; Owing to exist nonlinear phase to pollute the sky-wave OTH radar echoed signal is exerted an influence, and then influence the problem of detection performance.

The present invention realizes through following steps:

Step 1: the sky-wave OTH radar echoed signal is converted into pending sky-wave OTH radar time domain data by data acquisition module;

Step 2: the pending sky-wave OTH radar time domain data that step 1 collects is converted into corresponding frequency domain data through FFT by the data decimation module; Frequency domain data is chosen data window and intercepting is carried out in the zone of concentration of energy in the frequency spectrum, the frequency domain data after the intercepting is carried out IFFT obtain new sky-wave OTH radar time domain data and use data as handling;

Step 3: utilize the processing that obtains in the step 2 to pass through structure time domain correlation matrix with data by polluting the function calculation module; Take out its phasing matrix; Utilize Two-dimensional FFT to handle and draw echo frequency, and then draw nonlinear phase pollution function

Step 4: the nonlinear phase pollution function

by the utilization of signal correction module is drawn is proofreaied and correct pending sky-wave OTH radar time domain data; Obtain to remove the pending sky-wave OTH radar time domain data that nonlinear phase pollutes, and then obtain to go the sky-wave OTH radar echoed signal after nonlinear phase pollutes.

The beneficial effect of the invention

Solved the problem that the nonlinear phase pollution exerts an influence in the sky-wave OTH radar echoed signal, error obviously reduces, and this method should use better stability.

Description of drawings

Fig. 1 removes the process flow diagram of nonlinear phase pollution method for this sky-wave OTH radar echoed signal; Fig. 2 removes the comparison synoptic diagram of the power spectrum signal after the nonlinear phase pollution method depollutes for ideal power spectrum, pollution back power spectrum and this sky-wave OTH radar echoed signal in the embodiment two; Fig. 3 for pollution function that adds in the embodiment two and the pollution ratio of functions that draws than synoptic diagram; Fig. 4 goes nonlinear phase pollution method and the comparison diagram signal of PGA method under the condition of different pollution function frequencies for this sky-wave OTH radar echoed signal in the embodiment two; Fig. 5 removes nonlinear phase pollution method and the comparison synoptic diagram of PGA method under the condition of different signal to noise ratio (S/N ratio)s for this sky-wave OTH radar echoed signal in the embodiment two.

Embodiment

Below in conjunction with Fig. 1 to Fig. 5 embodiment of the present invention is described:

Embodiment one:

As shown in Figure 1, the sky-wave OTH radar echoed signal goes the nonlinear phase pollution method to realize through following steps:

Step 1: the sky-wave OTH radar echoed signal is converted into pending sky-wave OTH radar time domain data by data acquisition module;

Step 2: the pending sky-wave OTH radar time domain data that step 1 collects is converted into corresponding frequency domain data through FFT by the data decimation module; Frequency domain data is chosen data window and intercepting is carried out in the zone of concentration of energy in the frequency spectrum, the frequency domain data after the intercepting is carried out IFFT obtain new sky-wave OTH radar time domain data and use data as handling; At first, supposing to have a radar echo signal is that X is an echoed signal in a coherent accumulation cycle that does not receive ionosphere phase place pollution from certain resolution element.Then it is represented as follows through the corresponding echoed signal that produces ionosphere phase place pollution behind the ionospheric propagation:

pollutes function (rad) for phase place in the formula.

Convenience for research work; The definition of the radar return that exists the ionosphere phase place to pollute is following: as to suppose that processing comprises N pulse repetition time with data; And handle with data and contain unknown phase place pollution, handle with i in the data and be expressed as formula one from the sky-wave OTH radar time domain data of all irradiation areas constantly:

F in the formula---echoed signal frequency (Hz);

---first phase (rad);

γ _i---i phase place constantly pollutes (rad);

γ _iIn comprise small clutter and noise, and be reference value when establishing constantly i=1 and γ arranged ₁=0.

Step 3: utilize the processing that obtains in the step 2 to pass through structure time domain correlation matrix with data by polluting the function calculation module; Take out its phasing matrix; Utilize Two-dimensional FFT to handle and draw echo frequency, and then to draw the time domain correlation matrix that nonlinear phase pollutes function

structure be formula two:

R＝E[SS ^H]＝[r _m，k] _N×N

Here [] ^HThe expression conjugate transpose;

An interior signal that receives of coherent accumulation cycle can be expressed as formula three with matrix form:

S＝[S(1)，S(2)，...，S(N)] ^T

Here [] ^TThe expression transposition;

The capable k column element of the m of R is a formula four in the formula two:

r _m，k＝S(m)·S(k) ^H＝|S(m)||S(k)|·exp{j[2πf(m-k)+(γ _m-γ _k)]}

m＝1，2，...，N；k＝1，2，...，N；

Phasing matrix is:

If φ is the phasing matrix of time domain correlation matrix R, then φ _{M, k}Be r _{M, k}Phase place, formula five is then arranged

φ _m，k＝arg(r _m，k)＝2πf(m-k)+(r _m-r _k)

So first column element of φ is a formula six:

φ _m，1＝arg(r _m，1)＝2πf(m-1)+(r _m-r ₁)＝2πf(m-1)+r _m

Will produce a peak value in the matrix after data are handled through Two-dimensional FFT in the formula five; It is gone, column position is corresponding with phase function slope 2 π f; When this slope is drawn, phase place is polluted function and also just can be obtained by formula six, and the pollution function that then draws is a formula seven:

$\widehat{\mathrm{\γ}}={[{\widehat{\mathrm{\γ}}}_1,{\widehat{\mathrm{\γ}}}_2,...,{\widehat{\mathrm{\γ}}}_N]}^T$

Here use

and represent the i phase place constantly that draws to pollute.

Step 4: utilize the nonlinear phase of gained to pollute function

by the signal correction module and proofread and correct pending sky-wave OTH radar time domain data; Obtain to remove the pending sky-wave OTH radar time domain data that nonlinear phase pollutes, and then obtain to go the sky-wave OTH radar echoed signal after nonlinear phase pollutes.

Embodiment two:

In order to verify the validity of this method, carried out following emulation:

One, simulation analysis

Adopt simulation sky-wave OTH radar sea echo signal data.If it is formula eight that radar receives echo:

s(t)＝s ₁·exp(j2πf _bt)+s ₂·exp(-j2πf _bt)+w(t)

S in the formula ₁, s ₂---be respectively positive and negative single order sea clutter power magnitude;

f _b---the Bragg frequency;

W (t)---white Gaussian noise.

Eight corresponding power spectrum of formula are as shown in Figure 1.

Table 1 time domain related algorithm emulation experiment parameter

On the ideal signal basis shown in it, add phase place and pollute (it is generally acknowledged that polluting function is sinusoidal function), it is following to define phase place pollution function here:

γ(t)＝A(t)sin(2πf(t)·t)

A in the formula---pollute function amplitude;

F---pollute the function frequency.

It is as shown in table 1 that correction emulation experiment major parameter is polluted in the ionosphere of time domain related algorithm.

Dotted line is desirable power spectrum among Fig. 2, and dot-and-dash line is for adding the power spectrum after polluting, and solid line is the power spectrum that obtains after this sky-wave OTH radar echoed signal goes the nonlinear phase pollution method to handle; Dotted line pollutes function, the pollution function that solid line goes the nonlinear phase pollution method to obtain for this sky-wave OTH radar echoed signal for the phase place of adding among Fig. 3.Can find out that go power spectrum and original signal spectrum after the nonlinear phase pollution method is handled identical basically through this sky-wave OTH radar echoed signal, the pollution function consistance of the pollution function of interpolation and estimation is also relatively good, has reached the purpose of depolluting.

Two, this method and PGA performance are relatively

For further studying the performance of time domain related algorithm, from two aspects PGA and this method are compared analysis below.The used experiment parameter of PGA is identical with above-mentioned this method.

At first, analytical algorithm is polluted under the function different frequency situation phase place and is carried out process result.The RMSE (root-mean-square error) of the pollution function that utilization is here estimated and the pollution function of interpolation estimates its performance:

$\mathrm{RMSE}=\sqrt{\frac{1}{N}\underset{i=1}{\overset{N}{\mathrm{\Σ}}}{(\mathrm{\γ}-\widehat{\mathrm{\γ}})}^2}$

Through relation such as the table 2 and shown in Figure 4 between the frequency of root-mean-square error after time domain related algorithm and the processing of phase gradient method and phase place pollution function.

Table 2RMSE and the relation of polluting the function frequency

Can know according to the result; No matter this method still is the PGA algorithm is all satisfied following characteristic: hour RMSE is bigger when polluting the function frequency; At this moment because coherent only contains the pollution function that is less than one-period in integration time,, data pollute function information so only comprising part; RMSE is less when pollution function frequency is big, and this is that the pollution function information is complete in the data because coherent includes a plurality of pollution function cycles integration time at this moment; When pollution function frequency arrived to a certain degree greatly, RMSE tended towards stability.And there is certain difference in two kinds of algorithms, and comparatively speaking, the error that this method causes is slightly less than the error of PGA algorithm.

Next the property relationship between algorithm and the signal to noise ratio (S/N ratio) is studied.Here still adopt RMSE to compare analysis as evaluation index and employing PGA algorithm to algorithm performance.

Root-mean-square error after treatment and the relation between the signal to noise ratio (S/N ratio) such as table 3 are with shown in Figure 5.

The relation of table 3RMSE and signal to noise ratio (S/N ratio)

Can know that according to the result two kinds of algorithms all satisfy following characteristic: RMSE under the situation of low signal-to-noise ratio is bigger for algorithm, and the RMSE that under the situation of high s/n ratio, produces is less, and RMSE tends towards stability when signal to noise ratio (S/N ratio) is enough high.On this basis, there is certain difference in two kinds of algorithms, and this method RMSE is more level and smooth less than the curve of the relative PGA algorithm of relation curve of PGA algorithm and resulting RMSE of this method and signal to noise ratio (S/N ratio) comparatively speaking.

Visible by last emulation experiment, the broadening that this method can make the signal echo frequency spectrum produce is inhibited, and reach and proofread and correct the purpose that the ionosphere phase place is polluted, and its stability is superior to the PGA algorithm, and the total algorithm error is slightly less than the PGA algorithm.

Claims (3)

1. a sky-wave OTH radar echoed signal is removed the nonlinear phase pollution method, it is characterized in that: said method comprising the steps of:

Step 1: the sky-wave OTH radar echoed signal is converted into pending sky-wave OTH radar time domain data by data acquisition module;

Step 2: the pending sky-wave OTH radar time domain data that step 1 collects is converted into corresponding frequency domain data through FFT by the data decimation module; Frequency domain data is chosen data window and intercepting is carried out in the zone of concentration of energy in the frequency spectrum, the frequency domain data after the intercepting is carried out IFFT obtain new sky-wave OTH radar time domain data and use data as handling;

Step 3: utilize the processing that obtains in the step 2 to pass through structure time domain correlation matrix with data by polluting the function calculation module; Take out its phasing matrix; Utilize Two-dimensional FFT to handle and draw echo frequency, and then draw nonlinear phase pollution function

Step 4: utilize the nonlinear phase of gained to pollute function

by the signal correction module and proofread and correct pending sky-wave OTH radar time domain data; Obtain to remove the pending sky-wave OTH radar time domain data that nonlinear phase pollutes, and then obtain to go the sky-wave OTH radar echoed signal after nonlinear phase pollutes.

2. sky-wave OTH radar echoed signal as claimed in claim 1 is removed the nonlinear phase pollution method, it is characterized in that:

In the said step 2:

Suppose to handle and comprise N pulse repetition time, and handle and contain unknown phase place with data and pollute, handle with the i moment in the data and be expressed as formula one from the sky-wave OTH radar time domain data of all irradiation areas with data:

F in the formula---echoed signal frequency (Hz);

---first phase (rad);

γ _i---i phase place constantly pollutes (rad);

γ _iIn comprise small clutter and noise, be reference value during i=1 constantly and γ arranged ₁=0.

3. according to claim 1 or claim 2 sky-wave OTH radar echoed signal is removed the nonlinear phase pollution method, it is characterized in that: in the said step 3:

The time domain correlation matrix of structure is a formula two:

R＝E[SS ^H]＝[r _m，k] _N×N

Here [] ^HThe expression conjugate transpose;

An interior signal that receives of coherent accumulation cycle is expressed as formula three with matrix form:

S＝[S(1)，S(2)，...，S(N)] ^T

Here [] ^TThe expression transposition;

The capable k column element of the m of R is a formula four in the formula two:

r _m，k＝S(m)·S(k) ^H＝|S(m)||S(k)|·exp{j[2πf(m-k)+(γ _m-γ _k)]}

m＝1，2，...，N；k＝1，2，...，N；

Phasing matrix is:

If φ is the phasing matrix of time domain correlation matrix R, then φ _{M, k}Be r _{M, k}Phase place, formula five is then arranged

φ _m，k＝arg(r _m，k)＝2πf(m-k)+(r _m-r _k)

First column element of φ is a formula six:

φ _m，1＝arg(r _m，1)＝2πf(m-1)+(r _m-f ₁)＝2πf(m-1)+r _m

Will produce a peak value in the matrix after data are handled through Two-dimensional FFT in the formula five, it is gone, column position is corresponding with phase function slope 2 π f, and this slope is drawn, and phase place is polluted function and obtained by formula six, and the pollution function that then draws is a formula seven:

$\widehat{\mathrm{\γ}}={[{\widehat{\mathrm{\γ}}}_1,{\widehat{\mathrm{\γ}}}_2,...,{\widehat{\mathrm{\γ}}}_N]}^T$

Here use

and represent the i phase place constantly that draws to pollute.

Images