CN108152796A - A kind of main lobe based on grey-kalman filtering moves interference elimination method - Google Patents

Abstract

The invention discloses a kind of main lobes based on grey-kalman filtering to move interference elimination method.The proportionality coefficient of cancellation interchannel can accurately be predicted, and then can utilize and mobile interference is effectively eliminated based on the major lobe suppression removing method with poor channel when interference source is in mobile status using the present invention.The present invention is with the ratio between channel and gun parallax channel, ratio between trim channel and the poor channel of difference, ratio between channel and trim channel, and the ratio between gun parallax channel and the poor channel of difference is used as and the proportionality coefficient of difference interchannel interference cancellation, construct four groups of proportionality coefficient sequences, and then subsequent time proportionality coefficient is predicted using the method for Kalman filtering, so as to obtain the more accurate and poor channel proportionality coefficient at each moment, and then realize that is interfered when interference source is in mobile status effectively eliminates, improve angle measurement performance of the radar system to target.

Description

A Main Lobe Mobile Interference Elimination Method Based on Gray Kalman Filter

technical field

The invention relates to the technical field of signal processing, in particular to a gray Kalman filter-based main lobe mobile interference elimination method.

Background technique

In communication and radar systems, interference is an important limiting factor affecting information transmission and target detection. When there is interference in the space environment, in order to ensure the normal performance of the radar system such as angle measurement and ranging, it is necessary to suppress and eliminate the interference signal. When the interference falls into the main lobe area, the traditional adaptive beamforming will produce nulls in the main lobe, resulting in distortion of the antenna pattern and greatly degrading the detection performance of the radar. Therefore, in today's complex electromagnetic environment, research on the main lobe anti-jamming technology based on monopulse not only has important theoretical significance, but also has great engineering application value.

In recent years, scholars have studied the main lobe anti-jamming method. In order to solve the problem that the performance of the array radar is seriously degraded when the main lobe interference exists, a main lobe interference suppression algorithm based on the auxiliary array is proposed. Based on the blocking matrix preprocessing ( BMP) method, method based on eigenprojection preprocessing (EMP) and main lobe interference elimination method based on sum and difference channel, etc.

The essence of the main lobe interference suppression algorithm based on the auxiliary array is to add an auxiliary array near the main antenna, so that the interference signal originally located in the main lobe falls into the side lobe area of the entire array, and then the more mature side lobe can be used to interfere with the adaptive digital beam Algorithms are formed to suppress. However, when the interference signal falls into the main lobe area of the radar, the traditional adaptive digital beamforming technology will bring a series of problems such as antenna pattern distortion and side lobe level increase, which will lead to the performance of the adaptive digital beamforming algorithm severe decline or even failure.

BMP and EMP algorithms can improve the problem of beam distortion after eliminating the main lobe interference, but the BMP algorithm has a large amount of calculation, and the performance of the algorithm is greatly affected by the estimated value of the main lobe interference azimuth; the EMP algorithm needs to perform eigenvalue decomposition on the covariance matrix , and matrix inversion is required, the calculation is complicated. At the same time, there is a problem of peak offset in the EMP algorithm.

The main lobe anti-interference technology based on the sum and difference channel utilizes the separable nature of the antenna pattern in the azimuth and elevation directions, and eliminates the main lobe interference along one direction while keeping the sum and difference beam antenna pattern in the other direction undistorted , so as to derive an undistorted monopulse ratio for angle estimation. The key to this technique is to accurately determine the ratio coefficient between interference cancellation channels. The existing scaling coefficient determination method is to use the sampling value of each sum and difference channel as the input data of the Wiener filter, and use the Wiener filtering method to predict the scaling coefficient, and the final scaling coefficient is the cross-correlation coefficient and The ratio of autocorrelation coefficients. However, Wiener filtering is only suitable for the estimation of stationary random signals, that is, this method is only effective when the interference source is stationary. When the interference source is in the process of moving, the interference signal is a non-stationary random signal. At this time, the method of Wiener filtering will lose its effect.

Contents of the invention

In view of this, the present invention provides a main lobe mobile interference elimination method based on gray Kalman filter, which can accurately predict the proportional coefficient between the cancellation channels when the interference source is in a moving state, and then can use the sum and difference based The channel's main lobe interference cancellation method effectively eliminates mobile interference.

The main lobe mobile interference elimination method based on gray Kalman filtering of the present invention comprises the following steps:

Step 1, interference signal sampling: under the condition that the radar transmitter is not turned on and there is a mobile interference source, the interference signal received by the four original channels of the radar receiver antenna is sampled, and the radar transmitter starts to work after the sampling is completed;

Step 2, use the interference sampling signals of the four original channels obtained in step 1 to calculate the values of the sum channel, pitch difference channel, azimuth difference channel, and difference channel at each sampling time when the transmitter is not turned on; then calculate the sum of each sampling time Difference channel proportional coefficient; where the proportional coefficient of the sum difference channel is: the ratio between the sum channel and the azimuth difference channel, the ratio between the pitch difference channel and the difference difference channel, the ratio between the sum channel and the pitch difference channel, and the azimuth difference The ratio between channel and differential channel;

Step 3, for the four groups of proportional coefficients obtained in step 2, carry out gray Kalman filter respectively to the proportional coefficients of the sampling moments of each group, obtain the predicted value of the next moment of each group of proportional coefficients;

Step 4: Sample the echo signals received by the four original channels at the next moment, wherein the echo signals at the next moment include target signals and interference signals; calculate the sum channel, pitch difference channel, and azimuth at the next moment Values of difference channel and difference channel;

Step 5, according to the predicted values at the next moment of each group of proportional coefficients obtained in step 3, use the main lobe anti-interference method based on the sum and difference channel to perform interference elimination on the value of the sum and difference channel obtained in step 4.

Further, in step 2, the interference sampling signals of the four original channels obtained in step 1 are first screened and replaced, and then the value of the sum and difference channel at each sampling moment is calculated by using the value after screening and replacement; wherein, the filtered and replaced The method is: only when the sampling values of the interference signals of the four original channels at the same sampling time are greater than the threshold value Th _x set for each channel, the sampling data of the interference signals at the sampling time are retained; At the sampling moment of the data, the sampling values of the interference signals of the four original channels are replaced by the interpolation determined by linear interpolation using the previous retained interference signal sampling data and the latter reserved interference signal sampling data of each channel.

Further, the threshold value Th _x is 10 times of the mean value of the floor noise of the channel to which it belongs.

Further, the mean value of the channel floor noise is obtained by the following method: under the condition that the radar transmitter is not turned on and there is no interference source, the floor noise of the four original channels a, b, c, and d of the radar receiver antenna Sampling, and then calculate and obtain the mean values na, n _{b , nc} , _nd of the noise floor sampling values of each original channel respectively.

Further, in the step 5, the pitch ratio η _E and the azimuth monopulse ratio η _A after eliminating the interference are:

in, is the predicted value of the four proportional coefficients at the next moment obtained in step 3; r _Σ , r _ΔE , r _ΔA , and r _ΔΔ are the value of the sum channel, the value of the pitch difference channel, and the azimuth at the next moment obtained in step 4 respectively The value of the difference channel and the value of the difference channel.

Beneficial effect:

The present invention uses the ratio between the sum channel and the azimuth difference channel, the ratio between the pitch difference channel and the difference channel, the ratio between the sum channel and the pitch difference channel, and the ratio between the azimuth difference channel and the difference channel as The proportional coefficients of the interference cancellation between the sum and difference channels are constructed to construct four sets of proportional coefficient sequences, and then the proportional coefficients of the next time are predicted by the method of Kalman filtering, so as to obtain the more accurate proportional coefficients of the sum and difference channels at each time, and then Realize the effective elimination of interference when the interference source is in a moving state, and improve the angle measurement performance of the radar system to the target.

Description of drawings

Fig. 1 is a schematic flow sheet of the present invention;

Fig. 2 is a schematic diagram of the original channel of the antenna in the present invention;

Fig. 3 is a schematic diagram of an antenna and a difference channel in the present invention;

Fig. 4 is the contrast figure of monopulse ratio in mobile interference elimination of Kalman filtering method of the present invention and traditional Wiener filtering method;

Fig. 5 is a comparison diagram of the angle measurement error between the Kalman filtering method of the present invention and the traditional Wiener filtering method in mobile interference elimination.

Detailed ways

The present invention will be described in detail below with reference to the accompanying drawings and examples.

The present invention provides a kind of main lobe mobile interference elimination method based on gray Kalman filtering, comprising the following steps:

Step 1, interference signal sampling:

Under the condition that the radar transmitter is not turned on and there is a mobile interference source, the interference signals received by the four original channels a, b, c, and d of the radar receiver antenna are sampled, and the sampling values of the four original channel interference signals are obtained, respectively. Expressed as r _a,i , r _b,i , r _c,i , r _d,i , i=1,2,..., N, N is the total number of sampling points; the radar transmitter starts to work after the sampling is completed.

Then, the proportional coefficient at time N+1 is estimated by using the sampling values of the original channel interference signal at time 1 to time N.

Step 2, take the ratio between the sum channel and the azimuth difference channel, the ratio between the pitch difference channel and the difference channel, the ratio between the sum channel and the pitch difference channel, and the ratio between the azimuth difference channel and the difference channel The 4 ratios are used as the sum and difference channel proportional coefficient in the main lobe anti-jamming method based on the sum and difference channel; calculate the sum and difference channel proportional coefficient at each sampling moment;

First, using the sampling values of the four sets of original channel interference signals, calculate the values of the sum and difference channels (namely, the sum channel r _Σi , the pitch difference channel r _ΔEi , the azimuth difference channel r _ΔAi and the difference channel r _ΔΔi ) at each time from 1 to N , and then calculate the sum-difference channel proportional coefficient at each moment;

Among them, the values of the sum and difference channels (sum channel r _Σi , pitch difference channel r _ΔEi , azimuth difference channel r _ΔAi , and difference channel r _ΔΔi ) at each moment are:

r _Σi = r _a,i +r _b,i +r _c,i +r _d,i (1)

r _ΔEi = r _a,i +r _b,i -r _c,i -r _d,i (2)

r _ΔAi = r _a,i -r _b,i -r _c,i +r _d,i (3)

r _ΔΔi = r _a,i -r _b,i +r _c,i -r _d,i (4)

The proportional coefficient w _e1,i of the sum channel and the azimuth difference channel, the proportional coefficient w _e2,i of the pitch difference channel and the difference channel, the proportional coefficient w _a1,i of the sum channel and the pitch difference channel, the azimuth difference channel and the difference channel The proportional coefficients w _{a2, i} are respectively:

Considering that the noise amplitude in the antenna channel at a certain sampling moment may be far greater than the interference signal amplitude entering the antenna channel, causing the proportionality coefficient at the sampling moment to deviate from the reasonable range too much, the present invention is for the four groups obtained in step 1 The sampling value of the original channel interference signal is screened and replaced. The specific screening method is as follows:

First measure the mean value of the noise floor: under the condition that the radar transmitter is not turned on and there is no interference source, sample the floor noise of the four original channels a, b, c, and d of the radar receiver antenna, and calculate the floor noise of each channel respectively. Mean values n _a , n _b , n _c , n _d of noise sampling values;

Then take 10 times the mean value of the base noise as the screening threshold value Th _x , that is, Th _x = 10n _x , filter and replace the sampling values of the four groups of original channel interference signals obtained in step 1: only when the four channels at the same time When the sampling values of the interference signal are greater than the threshold value _Thx , that is, when formula (9) is satisfied, the sampling data of the interference signal at this moment is retained:

r _{a, i} ≥ Th _a ; r _{b, i} ≥ Th _b ; r _{c, i} ≥ Th _c ; r _{d, i} ≥ Th _d ; (9)

For the time k when no sampling data is retained (i.e., the time of elimination), the sampling values of the interference signals of the four original channels are replaced by: using the previous retained interference signal sampling data and the latter reserved interference signal sampling data of each channel The interpolation is determined using the method of linear interpolation.

Thus, each of the four channels gets a set of filtered effective sampling values This effective value has a large interference-to-noise ratio, which can reduce the influence of noise on the sampled interference signal, making the calculated proportional coefficient more reasonable.

Then use four sets of filtered effective sampling values Calculate the value of the sum-difference channel (sum channel r _Σi , pitch difference channel r _ΔEi , azimuth difference channel r _ΔAi , and difference channel r _ΔΔi ) at the corresponding moment, namely:

Step 3, use the gray Kalman filter to predict the proportional coefficient at the next moment:

The four sets of proportional coefficients obtained in step 2 are respectively passed through the Kalman filter constructed by the gray model to predict the proportional coefficient at time N+1 to obtain the predicted value of the proportional coefficient at time N+1

The specific process is:

Step 31: model assumption, for the first M (1<M<N) proportional coefficients of each channel, assume that the process model of the four channel proportional coefficients is s(n)=as(n-1)+w(n ), the measurement model is x(n)=cs(n)+v(n), where a is the process transfer coefficient, c is the measurement coefficient, the process noise Q=E(w ² [n]), the measurement noise R=E (v ² [n]).

Step 32: Forecast, in the first M moments, get the predicted value at time n (1<n<M) according to the estimated value of time n-1:

At this time, a and c are both set to 1, and the initial condition ξ(0) can be set to any value, and Q and R are set according to empirical values; after M time, the prediction method adopts the gray model composed of the first M Kalman filter estimated values at the current time n (M<n≤N) Predict the predicted value at n time.

Specifically, the prediction process using the gray model is:

Step 321: Construct a gray prediction input sequence, using M Kalman filter estimated values before time n As a gray prediction input sequence X ⁽⁰⁾ = [x(1), x(2),..x(N ₁ )];

Step 322: Construct the 1-AGO sequence X ^{( 1) of X (0)} , X ⁽¹⁾ = [x ⁽¹⁾ (1), x ⁽¹⁾ (2),...x ⁽¹⁾ (N ₁ ) ], the constructor is:

Step 323, constructing a GM (1,1) mean value formal model x(k)+pz ⁽¹⁾ (k)=q, where parameter vector It can be obtained by the method of least squares, and the solution formula is as follows:

in:

Step 324, calculate the predicted value at time n

Step 33: Calculate the minimum predicted mean square error P(n)

P(n)＝a ² ξ(n-1)+Q (16)

Step 34: Calculate the Kalman gain G(n)

Step 35: Correct the predicted value to obtain the estimated value at time n

Step 36: Calculate the minimum mean square error ξ(n) of the estimated value at time n

ξ(n)=(1-cG(n))P(n) (19)

Perform steps 31 to 36 for the four sets of proportional coefficients to obtain the predicted values of the four proportional coefficients at time N+1

Step 4, calculate the sum and difference channel values:

The echo signals received by the four channels are sampled once at the time N+1, wherein the echo signal at the time N+1 includes the target signal and the interference signal; calculate the value r _Σ and the channel at the time N+1 The value of pitch difference channel r _ΔE , the value of azimuth difference channel r _ΔA and the value of difference channel r _ΔΔ ;

Step 5, according to the predicted value of the proportional coefficient at the time N+1 obtained in step 3, adopt the existing main lobe anti-jamming method based on the sum and difference channel to the sum and difference channel value obtained in step 4 (the value r _Σ of the sum channel, pitch The value of the difference channel r _ΔE , the value of the azimuth difference channel r _ΔA and the value of the difference channel r _ΔΔ ) to eliminate interference, and obtain the pitch monopulse ratio η _E and the azimuth monopulse ratio η _A after the interference is eliminated, so as to obtain The angle of pitch and azimuth of the target at time N+1.

The calculation formula is:

Repeating steps 1 to 5, the target angle can be obtained every N sampling times, and appropriately reducing the sampling time can improve the measurement performance of the target angle.

The following is a simulation of using the above method to eliminate the simulated mobile interference source, and comparing it with the angle measurement performance under the traditional Wiener filter method for interference elimination.

Simulation parameter settings:

Can obtain Fig. 3, Fig. 4 simulation result by above-mentioned simulation parameter, can find out from the simulation result that the single-pulse ratio curve obtained under the method of the present invention is more consistent with the single-pulse ratio curve under the non-interference, and angle measurement error is smaller, From this, it can be concluded that the method of the present invention has obvious advantages in angle measurement performance compared with the traditional method using Wiener filtering for interference elimination.

In summary, the above are only preferred embodiments of the present invention, and are not intended to limit the protection scope of the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims (5)

1. a main lobe mobile interference elimination method based on gray Kalman filter, is characterized in that, comprises the steps: Step 1, interference signal sampling: under the condition that the radar transmitter is not turned on and there is a mobile interference source, the interference signals received by the four original channels of the radar receiver antenna are sampled, and the radar transmitter starts to work after the sampling is completed; Step 2, use the interference sampling signals of the four original channels obtained in step 1 to calculate the values of the sum channel, pitch difference channel, azimuth difference channel, and difference channel at each sampling time when the transmitter is not turned on; then calculate the sum of each sampling time Difference channel proportional coefficient; where the proportional coefficient of the sum difference channel is: the ratio between the sum channel and the azimuth difference channel, the ratio between the pitch difference channel and the difference difference channel, the ratio between the sum channel and the pitch difference channel, and the azimuth difference The ratio between channel and differential channel; Step 3, for the four groups of proportional coefficients obtained in step 2, gray Kalman filtering is performed on the proportional coefficients of the sampling moments of each group respectively, to obtain the predicted value of the next moment of each group of proportional coefficients; Step 4: Sample the echo signals received by the four original channels at the next moment, wherein the echo signals at the next moment include target signals and interference signals; calculate the sum channel, pitch difference channel, and azimuth at the next moment Values of difference channel and difference channel; Step 5, according to the predicted value of each group of proportional coefficients obtained in step 3 at the next moment, the value of the sum and difference channel obtained in step 4 is eliminated by using the main lobe anti-interference method based on the sum and difference channel. 2. the main lobe mobile interference elimination method based on gray Kalman filter as claimed in claim 1, is characterized in that, in described step 2, at first the interference sampling signal of four original channels obtained in step 1 is screened and replaced, Then use the value after filtering and replacing to calculate the value of the sum and difference channel at each sampling time; wherein, the method of filtering and replacing is: only when the sampling values of the interference signals of the four original channels at the same sampling time are greater than the set gate of each channel When the limit value Th _x , the sampling data of the interference signal at the sampling time is reserved; for the sampling time without sampling data, the sampling values of the interference signals of the four original channels are replaced by: using the previous retained interference of each channel The interpolation value determined by the method of linear interpolation is adopted for the signal sampling data and the latter preserved interference signal sampling data. 3. the main lobe mobile interference elimination method based on gray Kalman filter as claimed in claim 2, is characterized in that, described threshold value Th _x is 10 times of the base noise mean value of belonging channel. 4. the main lobe mobile interference elimination method based on gray Kalman filter as claimed in claim 3, is characterized in that, described passage base noise mean value adopts following method to obtain: in the condition that radar transmitter is not turned on, and there is no interference source Next, sample the noise floor of the four original channels a, b, c, and d of the radar receiver antenna, and then calculate the mean values of the noise floor sampling values of each original channel n _a , n _b , n _c , n _d . 5. the main lobe mobile interference elimination method based on gray Kalman filtering according to any one of claims 1 to 4, wherein in the step 5, the pitch to monopulse ratio η _E and azimuth after interference elimination The single pulse ratio η _A is: in, is the predicted value of the four proportional coefficients at the next moment obtained in step 3; r _Σ , r _ΔE , r _ΔA , and r _ΔΔ are the value of the sum channel, the value of the pitch difference channel, and the azimuth at the next moment obtained in step 4 respectively The value of the difference channel and the value of the difference channel.