CN105259540A - Optimization method for confronting active deception jamming by multi-station radar - Google Patents

Abstract

The invention discloses an optimization method for multi-station radar anti-active spoofing interference. The idea is: establish a multi-station radar system, including N node radars, and select a node radar to perform matching filtering and target detection on received signals, Get K+M point targets, take the first node radar as the reference radar, and perform time alignment operation on K+M point targets and other N-1 node radars, and then get K+M point targets The respective corresponding amplitude ratio feature vectors; set the range of the number of clusters to [1,K+M], cluster the corresponding amplitude ratio feature vectors of K+M point targets, and obtain K+M cluster numbers respectively The corresponding clustering results are evaluated by the similarity and distance index, and the number of clusters corresponding to the maximum value of the similarity and distance index is selected as the optimal number of clusters, and the clustering result corresponding to the optimal number of clusters is used as the final Clustering result: setting the judgment threshold ε of the false point target, and obtaining the false point target contained in the final clustering result of the point target amplitude ratio feature vector accordingly.

Description

An optimization method for anti-active deceptive jamming of multi-station radar

technical field

The invention relates to the technical field of radar anti-jamming, in particular to an optimization method for multi-station radar anti-active spoofing interference, which is suitable for effectively identifying and eliminating deceptive false targets in a networked radar system data fusion center to realize a multi-station radar system Combat deceptive interference.

Background technique

Electronic deception technology is dedicated to deceiving the victim radar in terms of information such as direction, position, and tracking starting point, or creating so many false targets around the real target echo that the real target information cannot be extracted. An effective category of electronic spoofing technology is deceptive electronic spoofing technology. The deceptive purpose of this deceptive electronic spoofing technology is to mislead the amplitude, phase and other information of the radar received echo through modulated transmission or forwarding, especially digital radio frequency memory. (DRFM), that is, the emergence of advanced transponder jammers makes deceptive jamming technology more mature, and is widely used in self-defense jamming and team jamming; in addition, deceptive jamming will occupy a lot of system resources and seriously affect the detection of radar systems performance and tracking performance.

For deceptive false target interference, single-station radar is difficult to counteract it due to its single viewing angle, while multi-station radar can use the method of dot trace correlation to distinguish the true and false of the detected target and eliminate the false target, so as to realize Countering deceptive jamming. However, since each node radar in the multi-station radar will be subject to deceptive interference, dense false targets will lead to a high error rate in the correlation test between the measurement values of each node radar, and the location of the multi-station radar is not ideal. It also affects the ability of multi-station radars to counteract spoofing jamming.

Most of the existing multi-station radars use data-level fusion to counter deceptive jamming. In the process of multi-station radar measurement of targets, only the point track information or track information of the target is used, so that the data-level fusion anti-jamming The method can't give full play to its anti-interference ability, and then can't make full use of the advantages of multi-station radar.

The existing anti-spoofing jamming method is a signal-level fusion method. Although it can make full use of various echo information, it also has many limitations and deficiencies. This signal-level fusion method utilizes the real target echoes in different radar stations. The envelopes are independent of each other and the interfering signal complex envelope is correlated to distinguish true and false targets. The identification effect depends on the number of pulse repetition period (PRT) in the slow time complex envelope sequence. In the actual radar working environment, it can be The number of PRTs used is very limited, and even only one pulse repetition period (PRT) can be used, so that the method of signal correlation detection will be completely invalid at this time. At the same time, it is very likely to identify a slow-fluctuating true target as a false target. A previously proposed multi-station radar anti-active spoofing jamming method overcomes the above shortcomings, but requires each false target to have the same jamming-to-noise power ratio (JNR).

Contents of the invention

Aiming at the deficiencies in the above-mentioned prior art, the purpose of the present invention is to propose an optimization method for multi-station radar anti-active deceptive jamming. effective target identification.

In order to achieve the above object, the present invention adopts the following technical solutions to achieve.

An optimization method for multi-station radar anti-active deceptive jamming, comprising the following steps:

Step 1, establish a multi-station radar system, the multi-station radar system includes N node radars, and the N node radars receive signals respectively, N is a natural number, and N≥2, and within a pulse repetition time, all After any one of the N node radars performs matched filtering and target detection on the received signal, K+M point targets are obtained; where K represents the number of real point targets in the received signal, and M represents the number of real point targets in the received signal. The number of false point targets that exist;

Step 2, take the first node radar as the reference radar, and perform time alignment operation on K+M point targets and other N-1 node radars, and obtain the point targets and reference points detected by the N-1 node radars The echo amplitude correspondence between the point targets detected by the radar, and then select the echo amplitude ε _p,n of the p-th point target in the n-th node radar from the echoes of the n-th node radar, and according to Calculate the amplitude ratio feature vector Ω _p of the p-th point target, and then obtain the corresponding amplitude ratio feature vectors of K+M point targets; where, p∈{1,2,...,K+M}, n∈{ 2,...,N};

Step 3, set the range of the number of clusters to [1, K+M], and accordingly perform clustering operations on the magnitude ratio feature vectors corresponding to the K+M point targets, and obtain K within the range of the set number of clusters The clustering results of the magnitude ratio feature vectors of the point targets corresponding to the +M clustering numbers;

Step 4: Evaluate the similarity and distance index for the clustering results of the point target amplitude ratio feature vectors corresponding to the K+M cluster numbers within the set cluster number range, and calculate the K+M cluster numbers respectively. Corresponding similarity and distance index values, and select the similarity distance index maximum value in the similarity distance index values corresponding to the K+M clustering numbers respectively, and then cluster the corresponding similarity distance index maximum value The number is used as the optimal clustering number, and the clustering result of the amplitude ratio feature vector of the point target corresponding to the optimal clustering number is used as the final clustering result of the point target amplitude ratio feature vector;

Step 5: Set the judgment threshold ε of the false point target, and obtain the real point target and the false point target included in the final clustering result of the point target amplitude ratio feature vector; wherein, ε is a natural number.

Compared with the prior art, the present invention has the following advantages: First, compared with the existing method, the present invention utilizes the discrete distribution of the amplitude ratio of the real target echo at each radar station, and the amplitude ratio of the false target echo is approximately the same , and the method of systematic cluster analysis can more effectively fight against various deceptive interference;

Second, the present invention does not rely on long-term data accumulation, and only needs one pulse repetition period (PRT) to complete the identification of true and false targets, with higher efficiency and stronger practicability;

Third, the present invention does not require each false target to have the same JNR, so it can fight against deceptive dense false targets, and can also effectively identify false targets produced by different deceptive interferences.

Description of drawings

The present invention will be described in further detail below in conjunction with the accompanying drawings and specific embodiments.

Fig. 1 is the realization flowchart of the optimization method of a kind of multi-station radar anti-active deceptive jamming of the present invention;

Figure 2 is a schematic diagram of the distribution of real targets and false targets in the amplitude ratio feature space;

Figure 3 is a curve diagram of the correct identification probability P _PT of the real target as a function of the interference and noise power ratio (JNR) under the three station deployment methods, where the abscissa is the interference and noise power ratio (JNR), and the ordinate is the real target The correct identification probability P _PT ;

Fig. 4 is a curve diagram of the correct identification probability _P'FT of false targets with the interference and noise power ratio (JNR) under the three station deployment methods, where the abscissa is the interference and noise power ratio (JNR), and the ordinate is the false target Probability of correct identification of the target P' _FT .

detailed description

With reference to Fig. 1, it is the realization flowchart of the optimization method of a kind of multi-station radar anti-active deceptive jamming of the present invention, the optimization method of this kind multi-station radar anti-active deceptive jamming, comprises the following steps:

Step 1, establish a multi-station radar system, the multi-station radar system includes N node radars, and the N node radars receive signals respectively, N is a natural number, and N≥2, and within a pulse repetition time (PRT) Inside, after any node radar in the N node radars performs matched filtering and target detection on the received signal, K+M point targets are obtained; wherein, K represents the number of real point targets existing in the received signal, and M represents The number of false point targets in the received signal.

Step 2, take the first node radar as the reference radar, and perform time alignment operation on K+M point targets and other N-1 node radars, and obtain the point targets and reference points detected by the N-1 node radars The echo amplitude correspondence between the point targets detected by the radar, and then select the echo amplitude ε _p,n of the p-th point target in the n-th node radar from the echoes of the n-th node radar, and according to Calculate the amplitude ratio feature vector Ω _p of the p-th point target, and then obtain the corresponding amplitude ratio feature vectors of K+M point targets; where, p∈{1,2,...,K+M}, n∈{ 2,...,N}.

Specifically, the first node radar is used as a reference radar, and K+M point targets and other N-1 node radars are time aligned to obtain the point targets detected by the N-1 node radars and the reference radar The echo amplitude correspondence between the detected point targets, and then select the echo amplitude ε _p,n of the p-th point target in the n-th node radar from the echoes of the n-th node radar, and calculate accordingly Get the amplitude ratio feature vector Ω _p of the pth point target, its expression is:

${\mathrm{\Ω}}_p=\[{\mathrm{\η}}_{12}^p,{\mathrm{\η}}_{13}^p,\mathrm{...},{\mathrm{\η}}_{1\mathrm{no}}^p,\mathrm{...},{\mathrm{\η}}_{1N}^p\],$ p∈{1,2,…,K+M}

${\mathrm{\η}}_{1\mathrm{no}}^p=\frac{\left|{\mathrm{\ϵ}}_{p,1}\right|}{\left|{\mathrm{\ϵ}}_{p,\mathrm{no}}\right|},$ n∈{2,…,N}

Among them, ε _p,1 represents the echo amplitude of the p-th point target on the reference radar, ε _p,n represents the echo amplitude of the p-th point target on the n-th node radar, and K represents the The number of real point targets, M represents the number of false point targets in the received signal.

According to the amplitude ratio feature vector Ω _p of the p-th point target, the corresponding amplitude ratio feature vectors of the K+M point targets are obtained.

Step 3: Set the range of the number of clusters to [1, K+M], use the method of systematic cluster analysis to perform clustering operations on the amplitude ratio feature vectors corresponding to K+M point targets, and obtain the set clusters The clustering results of the amplitude ratio feature vectors of the point targets corresponding to the K+M cluster numbers within the number range.

The sub-steps of step 3 are:

3.1 Set the range of the number of clusters to [1, K+M], and classify the magnitude ratio feature vector Ω _p of the pth point target into the pth class, and then get K+M different classes, p∈ {1,2,...,K+M}, the K+M different classes, that is, the clustering result of the amplitude ratio feature vector of the point target corresponding to the cluster number K+M.

3.2 Get the Euclidean distance value between any two classes in K+M different classes, where the Euclidean distance value between two classes in K+M different classes is its corresponding two Euclidean distance values between elements of the class;

Specifically, if the two classes each contain only one element, that is, when they are separate classes, the Euclidean distance value between the two classes is the Euclidean distance value between the elements of the corresponding two classes, and is calculated accordingly to get Euclidean distance values between every two classes in K+M different classes, and then obtain D Euclidean distance values, select the smallest Euclidean distance value in the D Euclidean distance values and perform optimization class operations;

Merge the two classes corresponding to the smallest Euclidean distance value among the D Euclidean distance values into the first optimized class, so that K+M different classes become K+M-1 different The K+M-1 different classes are the clustering results of the amplitude ratio feature vectors of the point targets corresponding to the cluster number K+M-1.

3.3 Select any two of the K+M-1 different classes, if at least one of the two classes contains two or more elements, the Euclidean distance between the two classes The value is the maximum Euclidean distance value between the elements contained in the corresponding two classes, and then calculate the Euclidean distance values of K+M-1 different classes, and select the K+M-1 each The smallest Euclidean distance value among the Euclidean distance values of different classes and optimize the class operation;

Merge the two classes corresponding to the smallest Euclidean distance values of the K+M-1 different classes into the second optimized class, and accordingly the K+M-1 different classes become K+M-2 different classes, the K+M-2 different classes are the clustering results of the amplitude ratio feature vectors of the point targets corresponding to the clustering number K+M-2.

3.4 Repeat the optimization operation to obtain the clustering results of the amplitude ratio feature vectors of the corresponding point targets when the number of clusters is K+M-3, K+M-4, ..., 1, and stop the optimization operation. Furthermore, the clustering results of the amplitude ratio feature vectors of the point objects corresponding to the K+M cluster numbers within the set cluster number range are obtained.

Step 4: Evaluate the similarity and distance index for the clustering results of the point target amplitude ratio feature vectors corresponding to the K+M cluster numbers within the set cluster number range, and calculate the K+M cluster numbers respectively. Corresponding similarity and distance index values, and select the similarity distance index maximum value in the similarity distance index values corresponding to the K+M clustering numbers respectively, and then cluster the corresponding similarity distance index maximum value The number is used as the optimal clustering number, and the clustering result of the amplitude ratio feature vector of the point target corresponding to the optimal clustering number is used as the final clustering result of the point target amplitude ratio feature vector.

Specifically, the similarity and distance index evaluation is performed on the clustering results corresponding to K+M cluster numbers within the set cluster number range. First, each of the K+M cluster numbers within the set cluster number range In the clustering results of the corresponding point target amplitude ratio feature vectors, select the clustering results of the point target amplitude ratio feature vectors corresponding to the qth clustering number, and compare the point target amplitude ratios corresponding to the qth clustering number The clustering results of the feature vectors are evaluated by the similarity and distance index, and the similarity and distance index value HS(q) of the clustering result corresponding to the qth cluster number is calculated, and its expression is:

HS(q)＝|Hom(q)-Sep(q)|

$\mathrm{<span\; class="notranslate">\; h</span>}\mathrm{<span\; class="notranslate">\; o</span>}\mathrm{<span\; class="notranslate">\; m</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 2</span>}}{\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; q</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; (</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; -</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; )</span>}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}}{\overset{\mathrm{<span\; class="notranslate">\; q</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}\underset{\mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; t</span>}}{\underset{\mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; \&Element;</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>$

$\mathrm{<span\; class="notranslate">\; S</span>}\mathrm{<span\; class="notranslate">\; e</span>}\mathrm{<span\; class="notranslate">\; p</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; q</span>}<span\; class="notranslate">\; )</span><span\; class="notranslate">\; =</span>\frac{\mathrm{<span\; class="notranslate">\; 1</span>}}{\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}<span\; class="notranslate">\; ;</span>\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; j</span>}}{\overset{\mathrm{<span\; class="notranslate">\; q</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}<span\; class="notranslate">\; \&times;</span>{\mathrm{<span\; class="notranslate">\; no</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; <</span>\mathrm{<span\; class="notranslate">\; j</span>}}{\overset{\mathrm{<span\; class="notranslate">\; q</span>}}{\underset{\mathrm{<span\; class="notranslate">\; i</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 1</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; j</span>}<span\; class="notranslate">\; =</span>\mathrm{<span\; class="notranslate">\; 2</span>}}{<span\; class="notranslate">\; \&Sigma;</span>}}}\underset{\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; \&Element;</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; j</span>}}}{\underset{\mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; \&Element;</span>{\mathrm{<span\; class="notranslate">\; C</span>}}_{\mathrm{<span\; class="notranslate">\; i</span>}}}{<span\; class="notranslate">\; \&Sigma;</span>}}\mathrm{<span\; class="notranslate">\; R</span>}<span\; class="notranslate">\; (</span>\mathrm{<span\; class="notranslate">\; the\; s</span>}<span\; class="notranslate">\; ,</span>\mathrm{<span\; class="notranslate">\; t</span>}<span\; class="notranslate">\; )</span>$

Among them, i∈{1,2,…,q}, j∈{1,2,…,q}, q∈[1,K+M], n _i represents the number of elements contained in the i-th individual class C _i n _j represents the number of elements contained in the j-th individual class C _j , R(s,t) represents the amplitude ratio feature vector Ω _s of the s-th point target and the amplitude-ratio feature vector Ω _t of the t-th point target The correlation coefficient between, s∈{1,2,...,K+M}, K represents the number of real point targets in the received signal, M represents the number of false point targets in the received signal, t∈{1 ,2,...,K+M}; when s=t, R(s,t)=1; when s≠t, ||·|| ₂ represents the 2 norm, Ω _m represents the amplitude ratio feature vector of the m-th point target, Ω _q represents the amplitude ratio feature vector of the q-th point target, m∈{1,2,…,K+ M}, q∈{1,2,...,K+M}.

According to the similarity and distance index value HS(q) of the clustering results corresponding to the qth cluster number, the similarity and distance index values corresponding to the K+M cluster numbers are obtained, and the K+M cluster numbers Select the maximum value of the similarity distance index from the similarity distance index values corresponding to the number of classes, and then use the cluster number corresponding to the maximum value of the similarity distance index as the optimal cluster number, and use the optimal cluster number corresponding to The clustering result of the amplitude ratio feature vector of the point target is used as the final clustering result of the point target amplitude ratio feature vector.

Step 5: Set the judgment threshold ε of the false point target, and obtain the real point target and the false point target included in the final clustering result of the point target amplitude ratio feature vector; wherein, ε is a natural number.

Specifically, in the amplitude ratio feature space, the amplitude ratios of the real point target echoes of each node radar in the multi-station radar system are discretely distributed, and the amplitude ratios of the false point target echoes are approximately the same, and the distribution is concentrated, After performing cluster analysis on the amplitude ratio feature vectors of K+M point targets, each real point target will become a separate class, that is, a separate class; while false targets will be concentrated into the same class due to the amplitude ratio distribution, according to This sets the judgment threshold ε of false point objects. If the number of point objects contained in any class in the final clustering result is less than ε, the point objects contained in the corresponding class correspond to real point objects.

If any class in the final clustering result contains ε or more point objects, it is determined that the point objects included in the class correspond to false point objects.

Wherein, the judgment threshold ε of the false point target is generally taken as 2; if the number of false point targets included in the final clustering result of the feature vector of the amplitude ratio of the point target is greater than the number of real point targets , ε can be taken as an integer greater than 2.

The ability of the present invention to resist deceptive interference can be further verified by the following simulation.

(1) Simulation parameters

Taking the multi-station radar system composed of four node radars as an example, the simulation experiment is carried out. The working mode of the first node radar is the transmit-receive mode, and the other three node radars are respectively in the receive mode, and then the four node radars are The composed multi-station radar system detects the same space area. In the same space area detected, there are five real targets, one of which carries a self-defense jammer, and produces 30 active deceptive false targets.

(2) Experimental content and result analysis

Experiment 1: In the above test scenario, the four node radar stations are [0,0], [-300,0], [300,0], and [600,0], and the first node radar is The reference point establishes a Cartesian coordinate system, and the size of the five real targets is 15m. The initial state of the five real targets is as follows:

When the noise in the multi-station radar system is Gaussian white noise, the distribution of true and false targets in the amplitude ratio feature space is shown in Figure 2, and Figure 2 is the distribution of real targets and false targets in the amplitude ratio feature space respectively Schematic diagram; Among them, the false alarm rate is the false alarm rate of the set false target judged as the real target.

It can be seen from Figure 2 that in the amplitude ratio feature space, the five real targets are randomly distributed, while the false targets are concentrated. This is due to the existence of noise in the system, so that the amplitude ratios of the false targets are not exactly the same, but very close.

Then the method proposed by the present invention is adopted to obtain the final identification result: the targets whose serial numbers are 23, 26, 28, 30, and 32 are real targets, and the remaining targets are false targets. From the above results and analysis, it can be seen that the proposed method of the present invention is effective.

Experiment 2: Analyze the influence of radar geometric layout on the identification performance of this method. The multi-station radar system composed of four node radars in Experiment 1 is set as station arrangement mode 1, except for the radar position, the other settings remain unchanged, and the other two station arrangement methods are:

Layout station 2: [0,0],[-200,0],[200,0],[400,0];

Layout station 3: [0,0],[-300,0],[300,0],[600,0],[-600,0].

Assuming that the jamming noise power ratio (JNR) of the first node radar is used as a variable, each increment is increased by 5dB, and the jamming noise power ratio (JNR) of the other node radars can be obtained by the bistatic radar equation. According to each fixed interference-to-noise power ratio (JNR), the active false target identification method proposed by the present invention is carried out 10 ⁴ Monte Carlo experiments, and the statistics obtain the correct identification probability P _PT of the real target and the correct identification probability P' of the false false target _FT , as shown in Fig. 3 and Fig. 4 respectively; Fig. 3 is a curve diagram of the correct identification probability P _PT of the real target with the interference-to-noise power ratio (JNR) under the three station layout modes, where the abscissa is the interference Noise power ratio (JNR), the ordinate is the correct identification probability P _PT of the real target; Figure 4 shows the variation curves of the correct identification probability P' _FT of the false target with the interference noise power ratio (JNR) under the three station deployment methods In the figure, the abscissa is the interference noise power ratio (JNR), and the ordinate is the correct identification probability _P'FT of the false target; the variation range of the interference noise power ratio (JNR) in Fig. 3 and Fig. 4 is 20-60.

It can be seen from Fig. 3 and Fig. 4 that with the increase of the interference-to-noise power ratio (JNR), the correct identification probability P _PT of the real target and the correct identification probability P' _FT of the false target increase respectively, especially when the interference noise After the power ratio (JNR) > 30, respectively have a fairly high probability of identification. In fact, in order to obtain better deception performance, the jamming-to-noise power ratio (JNR) is usually larger, and therefore the method proposed by the present invention can more effectively identify false targets and retain real targets.

It can also be seen from Fig. 3 and Fig. 4 that under the three station layout modes, the method proposed by the present invention has very satisfactory correct identification probability of real targets and correct identification probability of false targets respectively, and the performance is very close, indicating that the radar Station layout has little influence on the identification performance of the method proposed in the present invention. However, comparing the results of the first and second station deployment methods, the first station arrangement has better discrimination performance, because the first station arrangement has a longer baseline between radar stations, the longer the baseline, the real target The smaller the correlation of , the greater the randomness of the discrete division in the amplitude ratio feature space, and the better the discriminative performance. Comparing the results of the first and third station layout methods, it can be concluded that the more receiving stations there are, the better the identification performance will be. This is because the more receiving stations, the more information can be provided to build a multi-dimensional feature space, which greatly increases the separability between real targets and false targets.

In summary, the simulation experiment has verified the correctness, effectiveness and reliability of the present invention.

Obviously, those skilled in the art can carry out various modifications and variations to the present invention without departing from the spirit and scope of the present invention; Like this, if these modifications and variations of the present invention belong to the scope of the claims of the present invention and equivalent technologies thereof, It is intended that the present invention also encompasses such changes and modifications.

Claims (5)

1. An optimization method for resisting active deception jamming of a multi-station radar is characterized by comprising the following steps:

step 1, establishing a multi-station radar system, wherein the multi-station radar system comprises N node radars, the N node radars respectively receive signals, N is a natural number and is more than or equal to 2, and any one of the N node radars performs matched filtering and target detection on the received signals within a pulse repetition time to obtain K + M point targets; wherein, K represents the number of real point targets existing in the received signal, and M represents the number of false point targets existing in the received signal;

step 2, taking a first node radar as a reference radar, performing time alignment operation on the K + M point targets and other N-1 node radars to obtain the echo amplitude corresponding relation between the point targets detected by the N-1 node radars and the point targets detected by the reference radar, and then selecting the echo amplitude of the p point target in the N node radar from the echo of the N node radar_p,nAnd calculating the amplitude ratio characteristic vector omega of the p point target according to the amplitude ratio characteristic vector omega_pObtaining amplitude bit feature vectors corresponding to the K + M point targets respectively, wherein p ∈ {1,2, …, K + M }, N ∈ {2, …, N };

step 3, setting a clustering number range to be [1, K + M ], and carrying out clustering operation on the amplitude ratio feature vectors corresponding to the K + M point targets respectively according to the range to obtain clustering results of the amplitude ratio feature vectors of the point targets corresponding to the K + M clustering numbers in the set clustering number range;

step 4, performing similar phase separation index evaluation on the clustering results of the point target amplitude ratio feature vectors corresponding to K + M clustering numbers in the set clustering number range, respectively calculating to obtain similar phase separation index values corresponding to the K + M clustering numbers, selecting a maximum value of the similar phase separation index from the similar phase separation index values corresponding to the K + M clustering numbers, taking the clustering number corresponding to the maximum value of the similar phase separation index as an optimal clustering number, and taking the clustering result of the amplitude ratio feature vector of the point target corresponding to the optimal clustering number as a final clustering result of the point target amplitude ratio feature vector;

step 5, setting a judgment threshold value of the false point target, and accordingly obtaining a real point target and a false point target which are contained in a final clustering result of the point target amplitude ratio feature vector; wherein, it is a natural number.

2. The method as claimed in claim 1, wherein in step 2, the amplitude ratio feature vector Ω of the p-th point target is obtained_pThe expression is as follows:

${\mathrm{\&Omega;}}_p=\&lsqb;{\mathrm{\&eta;}}_{12}^p,{\mathrm{\&eta;}}_{13}^p,...,{\mathrm{\&eta;}}_{1n}^p,...,{\mathrm{\&eta;}}_{1N}^p\&rsqb;,p\&Element;\{1,2,...,K+M\}$

${\mathrm{\&eta;}}_{1n}^p=\frac{\left|{\mathrm{\&epsiv;}}_{p,1}\right|}{\left|{\mathrm{\&epsiv;}}_{p,n}\right|},n\&Element;\{2,\mathrm{...},N\}$

wherein,_p,1representing the echo amplitude of the p-th point target on the reference radar,_p,nthe echo amplitude of the p point target on the nth node radar is represented, K represents the number of real point targets existing in the received signal, and M represents the number of false point targets existing in the received signal.

3. The method as claimed in claim 1, wherein in step 2, the obtained clustering result of the amplitude ratio feature vector of the point target corresponding to each of K + M cluster numbers in the set cluster number range is obtained by the following process:

3.1 set the clustering number range to [1, K + M]And comparing the amplitude ratio feature vector omega of the p point target_pClassifying the obtained data into a pth class to further obtain K + M different classes, namely p ∈ {1,2, …, K + M }, wherein the K + M different classes are the clustering results of the amplitude ratio feature vectors of the point targets corresponding to the clustering number K + M;

3.2, acquiring Euclidean distance values between any two classes in K + M different classes, wherein the Euclidean distance values between the two classes in the K + M different classes are Euclidean distance values between elements of the two corresponding classes;

specifically, if two classes respectively only contain one element, namely, the two classes are independent classes, the Euclidean distance value between the two classes is the Euclidean distance value between the elements of the corresponding two classes, the Euclidean distance value between every two classes in K + M different classes is obtained through calculation according to the Euclidean distance value, D Euclidean distance values are further obtained, and the minimum Euclidean distance value in the D Euclidean distance values is selected to perform optimized class operation;

combining two classes corresponding to the minimum Euclidean distance value in the D Euclidean distance values into a first optimization class, wherein K + M different classes are changed into K + M-1 different classes according to the first optimization class, and the K + M-1 different classes are clustering results of the amplitude ratio feature vectors of the point targets corresponding to the clustering number K + M-1;

3.3 selecting any two classes of the K + M-1 different classes, if at least one of the two classes contains two or more elements, the Euclidean distance value between the two classes is the maximum Euclidean distance value between the elements contained in the corresponding two classes, further calculating to obtain the Euclidean distance values of the K + M-1 different classes according to the maximum Euclidean distance value, and selecting the minimum Euclidean distance value in the Euclidean distance values of the K + M-1 different classes and carrying out optimized class operation;

merging two classes corresponding to the minimum Euclidean distance values of the K + M-1 different classes into a second optimization class, wherein the K + M-1 different classes are changed into K + M-2 different classes, and the K + M-2 different classes are clustering results of the amplitude ratio feature vectors of the point targets corresponding to the clustering number K + M-2;

3.4 repeating the optimization operation to obtain the clustering results of the amplitude ratio feature vectors of the point targets corresponding to the clustering numbers of K + M-3, K + M-4, …, 1 respectively, and stopping the optimization operation to further obtain the clustering results of the amplitude ratio feature vectors of the point targets corresponding to the K + M clustering numbers in the set clustering number range.

4. The method as claimed in claim 1, wherein in step 4, the calculation procedure of the similar phase separation index values corresponding to the K + M cluster numbers is as follows:

firstly, selecting a clustering result of a point target amplitude ratio feature vector corresponding to a qth clustering number from clustering results of point target amplitude ratio feature vectors corresponding to K + M clustering numbers in a set clustering number range, carrying out similar phase separation index evaluation on the clustering result of the point target amplitude ratio feature vector corresponding to the qth clustering number, and calculating to obtain a similar phase separation index value HS (q) of the clustering result corresponding to the qth clustering number, wherein the expression is as follows:

HS(q)＝|Hom(q)-Sep(q)|

$Hom\left(q\right)=\frac{2}{\underset{i=1}{\overset{q}{\&Sigma;}}{n}_i(n_i-1)}\underset{i=1}{\overset{q}{\&Sigma;}}\underset{s<t}{\underset{s,t\&Element;C_i}{\&Sigma;}}R(s,t)$

$Sep\left(q\right)=\frac{1}{\underset{i=1,j=2;i<j}{\overset{q}{\&Sigma;}}{n}_i\&times;n_j}\underset{i<j}{\overset{q}{\underset{i=1,j=2}{\&Sigma;}}}\underset{t\&Element;C_j}{\underset{s\&Element;C_i}{\&Sigma;}}R(s,t)$

wherein, i ∈ {1,2, …, q }, j ∈ {1,2, …, q }, q ∈ [1, K + M }, and the like]，n_iRepresents the ith individual class C_iNumber of elements contained, n_jDenotes the jth individual class C_jThe number of elements contained, R (s, t) represents the amplitude ratio characteristic vector omega of the s-th point target_sAnd the amplitude ratio feature vector omega of the t-th point target_tThe correlation coefficient between s ∈ {1,2, …, K + M }, K representing the number of true point targets present in the received signal, M representing the number of false point targets present in the received signal, t ∈ {1,2, …, K + M }, where R (s, t) is 1 when s is t, and where s is not equal to t,||·||₂represents a 2 norm, Ω_mThe amplitude-bitfeature vector, Ω, representing the mth point target_qAnd the amplitude bit feature vector representing the target at the q-th point, M ∈ {1,2, …, K + M }, q ∈ {1,2, …, K + M }.

5. The method as claimed in claim 1, wherein in step 5, the threshold value for determining the false point target is set, wherein the threshold value is an integer greater than or equal to 2.