A radar signal sorting method based on convex clustering

Abstract

The present application relates to a kind of radar signal sorting methods based on convex clustering, comprising the following steps: obtaining the pulse signal from different radar signal sources received by receiver interception, the pulse signal parameters of different radar intercepted are standardized, signal sorting corresponding convex optimization function is constructed, the optimization objective function of convex optimization is solved using alternate minimization method, the above steps are iterated until the change of convex objective function is less than the preset threshold, the global optimal solution of the convex optimization function is obtained, the Euclidean distance between optimal solution each vector is compared, when the distance is less than the preset threshold, the pulse signal represented by similar vector is classified into the same cluster, and the final signal sorting result is obtained, in short, the present application has the advantages of high sorting method precision, good stability and noise robustness.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of electronic countermeasure radar signal processing, and particularly relates to a radar signal sorting method based on convex clustering. BACKGROUND

[0002] As one of the key technologies in radar signal processing, the main task of signal sorting is to separate and identify radar signals of different radiation sources from a complex electromagnetic environment. With the development of modern electronic warfare, the radar signal environment is becoming increasingly complex, and the signal modulation parameters are changing, which poses a severe challenge to radar signal sorting technology. Radar signal sorting needs to separate each radar pulse train from complex and interlaced pulse streams, and then analyze the signal parameters, identify the radiation source type, and judge the threat level of each pulse train, so that the signal sorting technology not only affects the accuracy and efficiency of radar signal detection and processing, but also is the core step of electronic warfare situation awareness, which has important military significance and practical value.

[0003] Currently, the clustering method based on multiple parameters is widely used in radar signal sorting tasks. As an unsupervised classification method, the clustering algorithm can discover the potential rules and patterns in the data by mining the correlation between the input data, and is suitable for solving the radar signal sorting problem lacking prior knowledge. The main method is to use the parameters such as carrier frequency, pulse width, and angle of arrival in the signal pulse description word for clustering, and to perform correlation analysis on the radar signal to ensure the accuracy of radar signal sorting.

[0004] The existing signal sorting algorithms mainly improve several mainstream methods such as K-means clustering algorithm and fuzzy C-means clustering algorithm. First, the K-means clustering algorithm is widely used in radar signal sorting due to its simple algorithm, fast convergence speed, and the advantage of not requiring prior knowledge. However, the traditional K-means algorithm needs to manually preset the number of clusters, and is sensitive to the cluster center, which affects the final signal sorting accuracy. At the same time, these existing clustering algorithms and their improvements are based on the optimization of non-convex objective functions, which are very easy to fall into local optimal solutions and sensitive to sample input order, and it is difficult to guarantee the stability of the solution and performance. Therefore, there is an urgent need for a radar signal sorting method based on convex clustering to solve the problems of low accuracy and unstable results of existing clustering algorithms. SUMMARY

[0005] Therefore, the application provides a radar signal sorting method based on convex clustering, which is applied to the technical field of electronic countermeasure radar signal processing and can solve the technical problems of low accuracy, dependence on manual preset cluster number, sensitivity to signal input order, and unstable results of existing clustering algorithms.

[0006] In order to achieve the above technical purposes, the specific technical scheme adopted by the present application is:

[0007] A radar signal sorting method based on convex clustering, comprising the following steps:

[0008] S1, obtaining the pulse signals from different radar signal sources received by the receiver;

[0009] S2, standardizing the pulse signal parameters of different radars received;

[0010] S3, constructing a signal sorting corresponding convex optimization function;

[0011] S4, using the alternating minimization method to optimize and solve the convex optimization objective function;

[0012] S5, iterating the above step S4 until the change of the convex objective function is less than the preset threshold, obtaining the global optimal solution of the convex optimization function;

[0013] S6, comparing the Euclidean distance between the optimal solution vectors, when the distance is less than the preset threshold, the pulse signals represented by the similar vectors are classified into the same cluster, and the final signal sorting result is obtained.

[0014] Further, in step S2, the pulse signal parameters are distributed in the same [0, 1] interval, and participate in the radar signal sorting task with the same dimension, the pulse signal parameters include carrier frequency, pulse width and arrival angle, and the pulse signal parameter matrix is:

[0015]

[0016] Where, cf i ,pw i ,doa i ,i=1,...,N are the frequency, pulse width and arrival direction of the i-th pulse, and N is the total number of pulse signals.

[0017] Further, in step S2, the mean and standard deviation of the k-th dimension data of the pulse signal parameter matrix are:

[0018]

[0019]

[0020] Where, 1≤k≤3 is the dimension of the pulse signal parameter matrix; the standardized value of the sample pulse vector is obtained:

[0021]

[0022] The extreme value normalization formula is used to distribute the normalized value of the pulse signal matrix in the interval [0, 1], and the formula is as follows:

[0023]

[0024] Wherein, pdw' kmin and pdw' kmax are the minimum and maximum values in pdw' ik .

[0025] Further, in step S3, the convex optimization function includes a loss function and a regularization term, and the optimization objective function is:

[0026]

[0027] Wherein, the former is the loss function, and the latter is the regularization term and generally takes the l1 norm. pdw" i is the normalized pulse signal parameter, u i is the cluster center of pdw" i , the weight parameter ω ij is non-negative and ω ij = ω ji , and γ is a non-negative tuning constant.

[0028] Further, in step S4, the convex clustering problem is redefined as the following equivalent constraint problem:

[0029]

[0030]

[0031] According to the constraint solving method, the augmented Lagrange expression form is as follows:

[0032]

[0033] Wherein, l is the constructed centroid pair l=(l1, l2) and l1 l < l2, ε = {l=(l1, l2): ω l > 0} is the edge set defined on the non-zero weight, and a new variable is introduced to define the difference between the centroids.

[0034] Further, in step S4, the alternating minimization method is used to solve the minimum value of the convex optimization function, and the convex optimization objective function is optimized and solved, and the iteration process of this m times is as follows:

[0035]

[0036] Wherein, is the mapping of z to the closed convex set C l . Cl = {λ l :||λ l || + ≤γω l}, is the dual norm of, and υ is the iteration step size,

[0037] With the above technical solution, the application can also bring the following beneficial effects:

[0038] 1. The radar signal sorting method based on convex clustering can obtain a global optimal solution of signal sorting by fusing cluster centers, more accurately identify and separate different radar signal sources, and improve the accuracy of signal sorting, and can also make the radar signal sorting effect not affected by the signal input order, has higher stability, even in the case of complex and changeable signal characteristics and large noise interference, still can maintain stable clustering effect, has high noise robustness, and can also fully utilize the multi-dimensional characteristics of radar signals for clustering analysis, has the advantages of comprehensively considering multiple dimensions, fully describing radar signal characteristics, improving the accuracy of signal sorting and providing effective support for electronic countermeasure systems.

[0039] 2. The radar signal sorting method based on convex clustering can efficiently and accurately separate different signal sources from complex electromagnetic environments, provide strong support for military decision-making and intelligence analysis, be applied to electronic reconnaissance tasks that require accurate identification of target radars, be a key step for extracting valuable intelligence, and also be used in electronic countermeasure scenes under low signal-to-noise ratio conditions, has good algorithm stability and noise robustness, and has great popularization space. BRIEF DESCRIPTION OF DRAWINGS

[0040] In order to more clearly illustrate the technical solutions of the embodiments of the application, the following will briefly introduce the drawings needed to be used in the embodiments. Obviously, the drawings in the following description are only some embodiments of the application, and for those skilled in the art, other drawings can also be obtained from these drawings without creative labor.

[0041] Figure 1 The flowchart of the radar signal sorting method based on convex clustering in the specific embodiment of the application is shown in the figure;

[0042] Figure 2 The radar pulse signal parameter information data graph in the specific embodiment of the application is shown in the figure;

[0043] Figure 3Figure for signal sorting accuracy information data in the embodiment of the present application;

[0044] Figure 4 Figure for comparison with stability of K-means in the embodiment of the present application;

[0045] Figure 5 Figure for comparison of sorting accuracy of K-means under different signal-to-noise ratios in the embodiment of the present application. DETAILED DESCRIPTION

[0046] The embodiments of the present application will be described in detail below with reference to the drawings.

[0047] The above and other aspects of the present application will become more apparent by describing in detail the embodiments thereof with reference to the attached drawings in which:

[0048] It is to be understood that the foregoing description is that of certain specific embodiments of the application. Numerous modifications, substitutions, changes, and equivalents will now occur to those of ordinary skill in the art. It is, therefore, to be understood that appended claims are intended to cover all such modifications and changes as fall within the true spirit and scope of the application. In the drawings:

[0049] It is also to be understood that the following description is only illustrative of the aspects of the present application and that changes can be made to adapt the teachings of the present application to other implementations and applications without departing from the spirit and scope thereof.

[0050] Also in the following description, specific details are given to provide thorough understanding of examples. One skilled in the relevant art will understand, however, that the aspects can be practiced without

[0051] In one embodiment of the present application, as shown in Figure 1 The present application refers to a radar signal sorting method based on convex clustering, comprising the following steps:

[0052] In step S1, the pulse signals from each signal source received by the receiver are acquired;

[0053] In step S2, the parameters of the received pulse signals are standardized to highlight the essential features of the data, including carrier frequency, pulse width and angle of arrival;

[0054] Specifically, for N pulse signals received by the receiver, the corresponding pulse signal parameter matrix is constructed:

[0055]

[0056] Where cf i ,pw i ,doa i , i = 1,...,N are the frequency, pulse width and direction of arrival of the i-th pulse, and N is the total number of pulse signals. The mean and standard deviation of the k-th dimension data of the pulse signal parameter matrix are:

[0057]

[0058]

[0059] Where 1 ≤ k ≤ 3 is the dimension of the pulse signal parameter matrix. The standardized value of the sample pulse vector is thus obtained:

[0060]

[0061] The extreme value standardization formula is used to distribute the standardized value of the pulse signal matrix within the [0,1] interval:

[0062]

[0063] Where pdw′ kmin and pdw′ kmax are the minimum and maximum values in pdw′ ik

[0064] In step S3, the corresponding convex optimization function for signal sorting is constructed, including a loss function and a regularization term.

[0065] ​Specifically, the signal sorting convex optimization objective function is:

[0066]

[0067] wherein the former is a loss function, the latter is a regularization term and generally takes an l1 norm. pdw" i is a normalized pulse signal parameter, u i is a cluster center of pdw" i , and the weight parameter ω ij is non-negative and ω ij = ω ji , and γ is a non-negative tuning constant.

[0068] In step S4, the minimum value of the convex optimization function is solved using an alternating minimization method.

[0069] Specifically, the convex clustering problem is first redefined as an equivalent constrained problem as follows:

[0070]

[0071] wherein l is a constructed centroid pair l=(l1, l2) and l1 l < 0} is an edge set defined on non-zero weights, and a new variable is introduced to define the difference between centroids. The purpose of variable splitting is to simplify the optimization with respect to the regularization term. According to the constraint solving method, the augmented Lagrangian expression form is given as follows:

[0072]

[0073] The convex optimization objective function is optimized and solved using an alternating minimization method, and the iteration process in the mth time is as follows:

[0074]

[0075] wherein is a mapping of z to a closed convex set C l , C l = {λ l : ||λ l || + ≤ γω l}, is a dual norm of , and υ is an iteration step size,

[0076] In step S5, the step S4 is iterated until the change of the convex objective function is less than a preset threshold, and the global optimal solution of the convex optimization function is obtained.

[0077] In step S6, the Euclidean distances between the optimal solution vectors are compared, and when the distance is less than a preset threshold, the pulse signals represented by the similar vectors are classified into the same cluster, and the final signal sorting result is obtained.

[0078] As shown in Figure 2 The simulation data of 6 radar pulse signal parameters are set in the present application, and a certain Gaussian measurement error is added to each signal parameter simulation, and the K-means clustering method is compared. The basic idea of the K-means algorithm is to divide the data into K different clusters by iteration, so that the sum of the distances between each data point and the center of its own cluster is minimized. The execution process of the K-means algorithm usually includes specifying the number of clusters K, selecting the initial clustering center, iteratively calculating the distance between sample points, updating the clustering center of each cluster, and outputting the final cluster. The K-means algorithm needs to manually preset the number of clusters, and is sensitive to the clustering center, which affects the final signal sorting accuracy and stability.

[0079] The present application sorts the above-mentioned guideline data, and detects the sorting accuracy of the corresponding pulse signal. The mixed signals of N targets in the environment are sorted, and the total number of real signal pulses of the i-th sorted signal is n i The number of correctly sorted pulses is m i The sorting accuracy of the signal pulse of this class is:

[0080]

[0081] When the ratio value is greater than a preset threshold (such as 90%), it is determined that the target i is successfully sorted. Therefore, the overall clustering sorting accuracy is the ratio of the number of successfully sorted pulses to the total number of real pulses, and the calculation formula is as follows:

[0082]

[0083] As shown in Figure 3 The sorting accuracy of different simulation radar signals can be obtained, and by comparing the sorting results of each simulation signal, the signal sorting algorithm based on convex clustering proposed in the present application can still ensure correct sorting results in the case of overlapping signal source parameters.

[0084] The stability of the radar signal sorting algorithm based on convex clustering proposed in the present application is compared with the stability of the K-means algorithm, and the number of experimental repetitions is 50 times, as shown in Figure 4As shown, the experimental results prove that when the signal input sequence is not fixed, the clustering stability of the K-means algorithm is poor, the volatility is large, and the average sorting accuracy is 0.9103. The radar signal sorting algorithm based on convex clustering mentioned in the application is not affected by the signal input sequence, the signal sorting accuracy is 0.9758, and the stability is strong.

[0085] Different signal-to-noise ratio simulation data are constructed by adding Gaussian white noise to the simulation data, such as Figure 5 As shown, the change curve of the sorting accuracy under different signal-to-noise ratios is obtained, 50 experiments are performed each time, and the average accuracy is taken. It can be seen that as the signal-to-noise ratio decreases, the signal sorting performance decreases, but the sorting accuracy of the application under different signal-to-noise ratios is better than that of K-means. And when the signal-to-noise ratio is higher than 10dB, a signal sorting accuracy of more than 85% can be obtained, which has high noise robustness, indicating that the sorting effect of the application is not affected by the signal input sequence, and the stability is higher. And under different signal-to-noise ratios, it still has good noise robustness, and is very suitable for use in electronic reconnaissance tasks that require accurate identification of target radars.

[0086] The above is only a specific embodiment of the application, but the protection scope of the application is not limited thereto, any person skilled in the art can easily think of changes or replacements within the technical range disclosed by the application, which should be covered within the protection scope of the application. Therefore, the protection scope of the application should be subject to the protection scope of the claims.

Claims

1. A convex clustering based radar signal sorting method, characterized in that, The method comprises the following steps: S1, obtaining pulse signals from different radar signal sources received by a receiver; S2, performing standardization processing on pulse signal parameters of different radars; S3, constructing a convex optimization function corresponding to signal sorting; S4, using an alternating minimization method to optimize and solve the convex optimization objective function; S5, iterating the step S4 until the change of the convex objective function is less than a preset threshold, and obtaining a global optimal solution of the convex optimization function; S6, comparing the Euclidean distances between optimal solution vectors, when the distance is less than a preset threshold, pulse signals represented by similar vectors are classified into the same cluster, and a final signal sorting result is obtained; In the step S2, the pulse signal parameters, including carrier frequency, pulse width and angle of arrival, which are intercepted are distributed in the same interval, and participate in the radar signal sorting task with the same dimension. The pulse signal parameter matrix is: wherein, are the frequency, pulse width and direction of arrival, respectively, of the first pulse, is the total number of pulses. The first column of the pulse signal parameter matrix The mean and standard deviation of the data in dimension wherein, is the dimension of the pulse signal parameter matrix; and the normalized value of the sample pulse vector is obtained: The extreme value normalization formula is used to distribute the normalized values of the pulse signal matrix in the interval [0, 1], and the formula is as follows: The extreme value normalization formula is used to distribute the normalized values of the pulse signal matrix in the interval [0, 1], and the formula is as follows wherein and are the minimum and maximum values, respectively, in ; In the step S3, the convex optimization function comprises a loss function and a regularization term, and the optimization objective function is: where the first term is the loss function, and the second term is the regularization term and generally takes the form norm, is the normalized pulse signal parameter, is the cluster center of the class, and the weight parameter is non-negative and , is a non-negative tuning constant; In the step S4, the convex clustering problem is redefined as an equivalent constraint problem as follows: According to the constraint solving method, the augmented Lagrange expression form is as follows: wherein, is the centroid pair of the configuration and , is the set of edges defined on non-zero weights, while introducing new variables to define the difference between centroids; The minimum value of the convex optimization function is solved using an alternating minimization method, and the optimization of the convex optimization objective function is solved, and the first The iteration process is as follows: wherein is a mapping to a closed convex set , , is the dual norm of is the iteration step size .

Images