CN104898132A - Navigation anti-interference algorithm combining threshold processing and space-frequency adaptive algorithm - Google Patents

Abstract

The invention discloses a navigation anti-interference algorithm combining threshold processing and a space-frequency adaptive algorithm, and belongs to the field of satellite navigation communication. The algorithm enables a threshold processing algorithm and the space-frequency adaptive algorithm to be combined together, carries out the threshold preprocessing of a signal converted into a frequency domain at first, and then carries out the space-frequency adaptive processing, thereby effectively improves the capability of resisting narrow-band strong interference of an anti-interference system. Meanwhile, the algorithm employs a method of two-channel data stack windowing for the anti-interference processing of the signal during the space-frequency adaptive processing, effectively reducing the impact on spectrum leakage from a conventional space-frequency adaptive algorithm, and improving the performance of a system.

Description

The navigation Anti-interference algorithm that a kind of threshold process and empty adaptive algorithm frequently combine

Technical field

The invention belongs to the satellite navigation communications field, in particular to the optimization of Anti-interference algorithm in satellite navigation system, that signal FFT is transformed to frequency domain concretely, first carry out pre-service at a kind of threshold algorithm of frequency domain, then realize empty Anti-interference algorithm frequently by a kind of structure of two paths of data windowing overlap.

Background technology

After the mankind enter information society, very general to the demand of geography information.Global position system has become the 3rd IT new growing point after communication, internet, and the application of global satellite positioning (GPS) is also increasingly extensive." GPS " (GPS) has played vital role in civilian and military domain, and its range of application is also in constantly expanding.

Navsat is generally designed to the weak signal satellite that emissive power only has a few milliwatt, and this reduces cost for satellite and increase the service life is needs.But because satellite-signal is weak, so be very easy to be interfered.Except the intentional interference of the jammer special in the face of enemy possible in war, the TV station that some frequencies are higher, aeronautical satellite communication and maneuvering satellite system terminal all may weaken navigation satellite signal, and some phenomenons that nature occurs also can cause signal disturbing.

Just may interrupt its use once satellite-signal is disturbed, positioning error be increased and cannot realize navigation feature even completely.At present, even the dependence of war to airmanship is increasing in social life, and the requirement of navigation neceiver interference free performance is also more and more higher, and the research specially for the Anti-Jamming Technique of navigation neceiver is just provided with realistic meaning.

Current is space-time adaptive signal transacting (STAP for the anti-interference main method taked of navigation, Space Time Adaptive Processing) way, the method after each array element, increases time-domain taps on the basis of spatial filter.The spatial filter that the same delay node of different array element is formed can to difference come to space interference form resolution characteristic, for each bay, the time delay tap number increased can form time domain filtering, can form resolution characteristic to the interference source of different frequency.STAP algorithm, owing to adding time domain degree of freedom, is compared simple airspace filter performance and is significantly improved, but the also corresponding increase of its computation complexity.M array element P is postponed to the STAP of node, often will carry out the matrix disposal of MP × MP dimension, operand is excessive.

Therefore, the empty Adaptive Signal Processing (SFAP, Space Frequency Adaptive Processing) (as shown in Figure 1) frequently of another algorithm becomes the algorithm that the anti-interference field of navigation begins one's study.

First SFAP algorithm is by the input data cutout K section of each array element, every segment data is carried out to FFT (Fast Fourier Transform (FFT)) computing of J point, obtain each array element respectively in the information of J frequency, finally obtain K group narrow band frequency component, often group has J point data.Then in this J frequency, carry out auto adapted filtering respectively by these K group data and suppress interference, finally the value of a filtered J frequency is carried out IFFT (Fast Fourier Transform Inverse (FFTI)), obtain output valve.

But, FFT conversion implies carries out periodic extension to input data, is not continuous print in the edge blocked after the input signal after blocking carries out periodic extension, and this discontinuous very unexpected, very serious spectral leakage can be caused like this, thus affect the spectrum estimation of other frequency.If directly carry out FFT computing to original input signal, be equivalent to add a rectangular window to original input signal, the first sidelobe level only 13.46dB lower than main lobe level of rectangular window, and general interfering signal power tens dBs higher than pilot signal power, its sidelobe level is more much larger than navigation signal level, has had a strong impact on the spectral line value of other frequency.So to estimate the power of each frequency accurately, just must reduce spectral leakage, common disposal route carries out windowing to input signal, makes the input signal after blocking become more level and smooth at edge after periodic extension.

If x (n) is input signal, w (n) is window function, then the signal expression after windowing is:

x _w(n)＝x(n)w(n) (1)

By can well be improved the impact of spectrum leakage after input signal windowing process, but due to the process of window function, input signal can partial distortion.

In order to improve the performance of SFAP when not increasing element number of array, before SFAP process, filtering can be carried out to arrowband interference.Process because SFAP needs first to convert the signal into frequency domain, so first carry out a pre-service at frequency domain to become a kind of selection before SFAP process.

The method of generally carrying out AF panel at frequency domain can use thresholding algorithm, and general thresholding algorithm has return-to-zero method and clamper method two kinds.Wherein return-to-zero method is the spectral line zero setting of amplitude higher than thresholding, and clamper method is that amplitude is dropped to threshold value higher than the spectral line value of thresholding.Jamproof snr loss of making zero can increase along with the bandwidth of undesired signal and increase; In order to retain useful signal as much as possible while suppression interference, can adopt clamper facture, its disposal route the spectral line amplitude be interfered is reduced and retains its phase propetry.

Thresholding algorithm only processes the spectral line that amplitude is greater than thresholding, when input signal is noiseless exist time, do not have the spectral line that amplitude is greater than threshold value, Threshold detection method can not be carried out any process to spectral line and directly be exported, so can not bring the impact of distortion to navigation signal.Because navigation undesired signal Relative Navigation signal is a kind of high reject signal, so choose suitable threshold value, strong arrowband interference can be filtered out well after threshold processing.

Summary of the invention

The present invention proposes the navigation Anti-interference algorithm of a kind of threshold process and the combination of empty adaptive algorithm frequently, first with thresholding algorithm, pre-service is carried out to signal, then carry out sky frequency adaptive algorithm by the method for two paths of data windowing overlap and realize.The present invention reduces the impact of spectrum leakage well, further increases the jamproof performance of this algorithm.

The technical scheme that the present invention takes is as follows:

A. the input signal of each array element is carried out piecemeal process as first via signal, namely choose at continuous J o'clock of input data as a data block process; By data delay half data block of first via signal, the signal after delay, as the second road signal, then carries out the piecemeal process identical with the first via.

B. the first via in a and the second road signal are carried out windowing process respectively, to be multiplied with window function the input data after obtaining windowing by the input data on the first via and the second tunnel.

C. the signal in b after windowing process is carried out FFT process, namely each data block is carried out to the FFT conversion of J point, obtain inputting the information of data at each frequency.

D. the signal obtained in c is carried out threshold processing, namely according to return-to-zero method or clamper method, the value rejecting or ream a certain frequency can be selected.

E. the signal after process in d is imported into empty immunity module frequently and carries out anti-interference process at each frequency, namely obtain weights at each frequency by Adaptive Anti-jamming algorithm, then weights are multiplied with input signal obtain anti-interference after the output of each frequency.

F. after the output valve of each frequency obtained in e being arranged in order, as a data block, import IFFT module into, obtain the output signal of time domain, by each data block each frequency anti-interference after output valve carry out J point IFFT, obtain the time domain output valve of this data block.

G. the first via time-domain signal obtained in f is outputed signal the time of delay half data block, then first via time-domain signal is added with the second road time-domain signal, obtain final time domain output signal.

The invention has the beneficial effects as follows: after FFT process is carried out to signal, first carry out threshold processing, and then carry out the mode of SFAP algorithm process, effectively improve the effect of system rejection to disturbance.Simultaneously by taking the method for two paths of data windowing overlap, while suppression spectrum leakage, decreasing the distorted signals that window function causes, thus further increasing the interference free performance of whole system.

Accompanying drawing explanation

Fig. 1 is the empty adaptive algorithm structured flowchart frequently of tradition;

Fig. 2 is the overlapping jamproof schematic diagram of two paths of signals windowing of the present invention.

Embodiment

Be illustrated in figure 2 the schematic diagram of two paths of signals windowing overlap of the present invention.

A. two paths of signals is carried out piecemeal process, and delay disposal is carried out to the second road signal.

If will carry out the FFT of J point to input signal, then input data are carried out piecemeal by every J data one piece, namely first via signal first blocks of data is:

x _in1(n)＝(x(n),x(n+1),…,x(n+J-1)) (2)

Wherein x (n)=(x ₁(n), x ₂(n) ... x _m(n)) ^t, represent the snap value of M array element in the n-th moment.Using first via signal delay 1/2 data block as the second road input signal, namely the second road signal first blocks of data is:

x _in2(n)＝(x(n+J/2),x(n+J/2+1),…,x(n+J/2+J-1)) (3)

B. input signal is carried out windowing process.

In fact exactly by data block, a window function coefficient is multiplied by data block windowing.Different window functions suppresses degree varies sample to the spectral leakage that FFT converts, this decides primarily of main lobe width and side lobe levels, the situation of not windowing is in fact equivalent to and adds rectangular window, and the signal to noise ratio (S/N ratio) decline degree of window functions different in addition to original signal is also different.

Consider spectral leakage to improve and signal to noise ratio (S/N ratio) decline these two aspects factor, the combination property of hamming window is best.Length is the mathematic(al) representation of the hamming window function of J

$w\left(n\right)=0.54-0.64\cos \left(2\mathrm{\π}\frac{n}{J-1}\right),0\≤n\≤J-1---\left(4\right)$

The window function coefficient of J point can be generated by Calling MATLAB function hamming (J), then by the window function multiplication of the first via and the second road input signal and generation, the input signal after windowing process can be obtained.

C. the signal after windowing process is sent into FFT module:

Data after the first via and the second tunnel windowing are sent into FFT module successively, if each road signal is fed into altogether K data block, carries out the FFT computing of J point, then can obtain K group data, often organize the information that data have J frequency.Then can obtain the value X that each road signal kth data block carries out J after FFT point _k(f ₁), X _k(f ₂) ..., X _k(f _j).

D. pre-service is carried out to the frequency-region signal obtained.

First choose suitable thresholding, because undesired signal is strong jamming relative to navigation signal, so thresholding is chosen less demanding, generally use single order moments method, the interference threshold computing formula of single order moments method is

$T_{\mathrm{th}}=\mathrm{\λ}\frac{1}{J}\underset{i=1}{\overset{J}{\mathrm{\Σ}}}a\left(i\right)---\left(5\right)$

Wherein a (i) is the amplitude of i-th spectral line.λ is thresholding Optimization Factor, and general value is between 2 to 5.The value of λ directly has influence on jamproof effect, specifically need determine according to actual conditions.

After determining thresholding, can check the amplitude of each frequency, the amplitude of the frequency being greater than thresholding is reduced to the value that certain is selected, and retain its phase propetry.Strong arrowband interference can be curbed through pretreated signal.

E. signal pretreated in d is carried out sky anti-interference process frequently.

Choose specific adaptive filter algorithm, by calculating the weight w (f of this frequency respectively at J frequency _j).Due to sky adaptive filter algorithm frequently, when each frequency carries out weight computing, only need the computing carrying out M × Metzler matrix, greatly reduce calculated amount, so use direct matrix in verse in Power-inversion algorithm to try to achieve weights.Establish in Power-inversion algorithm steering vector s=[1,0 ..., 0], then at the formula of a jth frequency calculating weights be:

$w\left(f_i\right)=\frac{R_{\mathrm{xx}}\left(f_j\right)s}{s^T{R}_{\mathrm{xx}}{\left(f_j\right)}^{-1}s}---\left(6\right)$

Wherein R _xx(f _j) be correlation matrix, can be drawn by the statistical average of K group data, namely

${\hat{R}}_{\mathrm{xx}}\left(f_j\right)=\frac{1}{K}\underset{k=1}{\overset{K}{\mathrm{\Σ}}}{X}_k\left(f_j\right)X_k^H\left(f_j\right)---\left(7\right)$

Wherein K is the fast umber of beats be averaging chosen, and the output valve finally obtaining each frequency is

Y(f _j)＝X(f _j)w ^*(f _j) (8)

F. the first via data obtained in e are carried out the delay of 1/2 data block, and then be added with the second circuit-switched data, obtain last output.

Carry out AF panel after the overall thought of windowing overlap goes up branch road windowing exactly, then carry out the delay of 1/2 data block; Lower branch road first carries out the delay of 1/2 data block, and then windowing, finally carries out AF panel.Upper branch road and lower branch road all carried out the upper and lower branch road of the delay guaranteed of 1/2 data block be added time in time time corresponding.If do not disturbed, AF panel module can not do any process to the output signal of FFT module, and the output signal of IFFT module is the same with the input signal of FFT module.

Because the envelope of window function on input signal has impact, the structure of windowing overlap can ensure the signal envelope after two paths of signals superposition and input signal envelope close, input signal can be reached in distortionless situation, carry out AF panel.

Claims (2)

1. a navigation Anti-interference algorithm for threshold process and the combination of empty adaptive algorithm frequently, is characterized in that comprising the following steps:

A. the input signal of each array element is carried out piecemeal process as first via signal, namely choose at continuous J o'clock of input data as a data block process; By data delay half data block of first via signal, the signal after delay, as the second road signal, then carries out the piecemeal process identical with the first via;

B. the first via in a and the second road signal are carried out windowing process respectively, to be multiplied with window function the input data after obtaining windowing by the input data on the first via and the second tunnel;

C. the signal in b after windowing process is carried out FFT process, namely each data block is carried out to the FFT conversion of J point, obtain inputting the information of data at each frequency;

D. the signal obtained in c is carried out threshold processing, select the value rejecting or ream a certain frequency;

E. the signal after process in d is imported into empty immunity module frequently and carries out anti-interference process at each frequency, namely obtain weights at each frequency by Adaptive Anti-jamming algorithm, then weights are multiplied with input signal obtain anti-interference after the output of each frequency;

F. after the output valve of each frequency obtained in e being arranged in order, as a data block, import IFFT module into, obtain the output signal of time domain, by each data block each frequency anti-interference after output valve carry out J point IFFT, obtain the time domain output valve of this data block;

G. the first via time-domain signal obtained in f is outputed signal the time of delay half data block, then first via time-domain signal is added with the second road time-domain signal, obtain final time domain output signal.

2. the navigation Anti-interference algorithm that combines of a kind of threshold process as claimed in claim 1 and empty adaptive algorithm frequently, is characterized in that: to tell the threshold processing method adopted in steps d be return-to-zero method or clamper method.