CN115034261A - Method and equipment for extracting features between pulses of radar radiation source signal and storage medium - Google Patents
Method and equipment for extracting features between pulses of radar radiation source signal and storage medium Download PDFInfo
- Publication number
- CN115034261A CN115034261A CN202210587752.7A CN202210587752A CN115034261A CN 115034261 A CN115034261 A CN 115034261A CN 202210587752 A CN202210587752 A CN 202210587752A CN 115034261 A CN115034261 A CN 115034261A
- Authority
- CN
- China
- Prior art keywords
- pulse
- sub
- networks
- network
- radiation source
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Images
Classifications
-
- G—PHYSICS
- G01—MEASURING; TESTING
- G01S—RADIO DIRECTION-FINDING; RADIO NAVIGATION; DETERMINING DISTANCE OR VELOCITY BY USE OF RADIO WAVES; LOCATING OR PRESENCE-DETECTING BY USE OF THE REFLECTION OR RERADIATION OF RADIO WAVES; ANALOGOUS ARRANGEMENTS USING OTHER WAVES
- G01S7/00—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00
- G01S7/02—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00
- G01S7/28—Details of pulse systems
- G01S7/285—Receivers
-
- Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
- Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
- Y02A—TECHNOLOGIES FOR ADAPTATION TO CLIMATE CHANGE
- Y02A90/00—Technologies having an indirect contribution to adaptation to climate change
- Y02A90/10—Information and communication technologies [ICT] supporting adaptation to climate change, e.g. for weather forecasting or climate simulation
Abstract
The invention discloses a method, equipment and a storage medium for extracting features between pulses of radar radiation source signals, wherein the method comprises the following steps: s1, forming a two-dimensional sequence by the intercepted pulse parameters and the pulse arrival time respectively; s2, determining optimal finite traversal vision, respectively mapping the two-dimensional sequences into complex networks, obtaining a matrix X based on adjacent matrixes of the complex networks, and constructing a total complex network model according to the matrix X; s3, dividing and sampling the nodes of the total complex network model to obtain a plurality of pulse subsequences, determining the decomposition number, and decomposing the pulse subsequences into a plurality of sub-networks; s4, calculating the average degree vector of each sub-network, reducing the dimension of the vector, and obtaining the two-dimensional feature with the maximum contribution degree; the method has low complexity and high data processing efficiency, and the obtained characteristics can reflect the interaction relation among pulses and the pulse time sequence change and can accurately sort and identify radar pulse radiation source signals based on the interaction relation and the pulse time sequence change.
Description
Technical Field
The invention belongs to the technical field of signal identification, and particularly relates to a method and equipment for extracting features between pulses of a radar radiation source signal and a storage medium.
Background
In the case of dense, complex, staggered and diverse modern electromagnetic environments, sorting and identification of radar radiation source signals are always challenges facing electronic reconnaissance systems; generally, the object of the radar radiation source sorting and identifying modeling is a pulse sequence of full pulses and characteristic parameter information constructed by fine features in the pulses, which is an important component and a basis of a sorting and identifying algorithm and influences the performance of sorting and identifying in both processes and results.
For a long time, the traditional sorting and identifying method based on the radiation source signal directly depends on one or more parameter characteristics in the conventional five-parameter pulse (carrier frequency (RF), Pulse Width (PW), Pulse Amplitude (PA), angle of arrival (DOA) and time of arrival (TOA)) to perform sorting and identifying, and the method has good sorting and identifying effects only when the conventional radar and signal density degree is low in the early period; some recent research methods adopt a neural network and deep learning in artificial intelligence to carry out sorting and identification on radar radiation source signals, but still stay on sorting and identification research based on a traditional Pulse Description Word (PDW), and a new characteristic analysis and characterization method aiming at a full pulse sequence does not appear, so that pulse measurement parameters cannot well reflect the time sequence change characteristics among pulses or among pulses, and meanwhile, the topological structure of the interaction relationship among pulses cannot be reflected; therefore, under the environment of coexistence and high-density signal of the radar with the complex system, the characteristics of waveform change among pulses of the radar full-pulse sequence are researched, and the characteristics of mining and constructing the pulse sequence from different visual angles and the rule of system evolution development are the key of improving the sorting recognition algorithm.
The complex network theory is an important branch of statistical physics developed in recent years, from the perspective of complex networks, a set of method for mapping a time sequence to the complex network has been developed, the structure and the dynamic mechanism of the time sequence can be deeply known, and the method has the advantages of simplicity, intuition, good universality, obvious topological property, strong network robustness and the like.
Disclosure of Invention
The embodiment of the invention aims to provide a method, equipment and a storage medium for extracting features between pulses of a radar radiation source signal, so that the acquired features between the pulses can well reflect the time sequence change characteristics of the pulses, the interaction relation between the pulses is reflected, and the radar pulse radiation source signal can be accurately sorted and identified on the basis.
In order to solve the technical problem, the technical scheme adopted by the invention is that the method for extracting the inter-pulse characteristics of the radar radiation source signal comprises the following steps:
s1, intercepting the pulse data parameter RF i 、PW i 、PA i Respectively with the pulse arrival time t i Form a two-dimensional sequence t i ,RF i }、{t i ,PW i }、{t i ,PA i }；
Wherein i represents a number variable of the number of pulses, RF i Carrier frequency, PW, representing the ith pulse i Indicating the pulse width, PA, of the ith pulse i Represents the pulse amplitude of the ith pulse;
s2, determining the optimal finite traversal sight distance, and respectively connecting the two-dimensional sequences t i ,RF i }、{t i ,PW i }、{t i ,PA i Mapping to complex network, obtaining matrix X by using adjacent matrix of each complex network, and further obtaining matrix XObtaining a total complex network of radiation sources;
s3, respectively adopting an equidistant division and sliding window method to carry out division sampling on pulse nodes of the total complex network to obtain two groups of pulse sequences, wherein each pulse sequence comprises Q pulse subsequences with fixed lengths, determining the decomposition number M of the subsequences, and respectively decomposing the two groups of pulse sequences into QxM sub-networks;
and S4, respectively calculating the average degree vectors of the two groups of pulse sequences corresponding to the sub-networks to obtain two average degree vector matrixes P, and reducing the dimension of the two average degree vector matrixes P by adopting principal component analysis to obtain two-dimensional characteristics with the maximum contribution degree, namely the extracted radar radiation source signals.
Further, the process of determining the optimal limited traversal viewing distance in S2 is as follows:
setting a crossing-limited visual range initial value N as 1, repeating S2-S3 to obtain Q multiplied by M sub-networks corresponding to two groups of pulse sequences, and calculating an average vector of each sub-network;
clustering the two groups of sub-networks respectively, and calculating the class separability measure of each sub-network based on the average degree vector of each sub-network in the clustering result;
and increasing the limited crossing visual distances N one by one, repeatedly calculating the class separability measures of the sub-networks under different limited crossing visual distances, and taking the limited crossing visual distance corresponding to the minimum value of the class separability measures as the optimal limited crossing visual distance.
Further, the matrix X is obtained as follows:
adding adjacent matrixes of each complex network to obtain a matrix M, wherein the element M in the matrix M i,j ＝0,1,2,3，M i,j Representing the total connection number of the pulse i and the pulse j in the three complex networks;
using majority voting to pair M i,j Processing to obtain matrix X, wherein elements in matrix X
X i,j 1 indicates that there is an edge between two pulse nodes, X i,j 0 means that there is no edge between the two pulse nodes.
Further, the category separability measure J is calculated as follows:
wherein S w Representing the mean vector divergence, S, of the intra-class subnetwork b Representing the divergence of the vectors of the average degree of the sub-networks among the classes, C representing the total number of the classes obtained by clustering, C representing the variable of the number of the classes, N c Indicates the total number of subnetworks included in the class c, a indicates the number variable of subnetworks in each class,
represents the mean vector, μ, of the a-th sub-network in class c c Denotes the mean of all sub-network mean vectors in class c, μ denotes the mean of all sub-network mean vectors, and T denotes the transpose.
Further, the vector matrix of the average degree
Where q denotes the number variable of the pulse subsequences, m denotes the number variable of the sub-networks,
respectively representing the average degree vectors of the 1 st, the M th and the M th sub-networks obtained by decomposing the 1 st pulse subsequence,
respectively representing the average degree vectors of the 1 st, the M th and the Mth sub-networks obtained by decomposing the q-th pulse subsequence,
respectively representing the average degree vectors of the 1 st, the M th and the M th sub-networks obtained by decomposing the Q-th pulse subsequence,
N m representing the total number of pulses in the mth subnet,
representing the node degree of the pulse i in the mth subnet, i.e., the number of pulses connected to the pulse i.
An electronic device comprises a processor, a memory and a communication bus, wherein the processor and the memory are communicated with each other through the communication bus;
a memory for storing a computer program;
and the processor is used for realizing the steps of the method when executing the program stored in the memory.
A computer-readable storage medium, in which a computer program is stored which, when being executed by a processor, carries out the above-mentioned method steps.
The invention has the beneficial effects that: according to the embodiment of the invention, the time sequence is used as a bridge, and the radar radiation source inter-pulse signals are converted into network nodes and edges for modeling and analysis, so that the local difference, the relevance and the effectiveness of a radar signal full-pulse sequence are further excavated under the conditions that the pulse is seriously lost and a large number of noise pulses exist.
In the embodiment, intercepted parameters RF, PA and PW among unknown radar signal pulses are used as a joint feature vector, an original pulse sequence is divided and sampled by adopting an equidistant division and sliding window method to obtain a fixed-length pulse subsequence, the dimensionality reduction of a full pulse sequence is realized, the complexity of data processing is reduced, the processing efficiency is improved, local information and a system evolution rule of pulse data are kept as far as possible, and a basis is provided for further analyzing the interaction relation and the topological structure characteristic among radar pulses and pulse groups.
The method comprises the steps of determining the separability measure of pulse sequence parameters according to the intra-class divergence and the inter-class divergence, selecting the optimal limited traversable line of sight to perform network domain transformation on radar radiation source signal pulse subsequences subjected to dimensionality reduction, constructing two groups of complex network models without weights and directions, performing visualization processing and important node analysis on the network, and investigating the quantity of pulse signals which change along with time and are mutually related under the dense environment.
The method is used for uniformly decomposing the complex network, taking the average degree of each sub-network as the extracted statistical characteristics, adopting principal component analysis to reduce the dimension again to construct the principal characteristics of the radar radiation source signals, and improving the aggregation degree of the network statistical characteristic distribution under the condition of losing pulses through comparative analysis.
Drawings
In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings used in the embodiments or the prior art descriptions will be briefly described below, it is obvious that the drawings in the following description are only some embodiments of the present invention, and other drawings can be obtained by those skilled in the art without creative efforts.
Fig. 1 is a flowchart of the present embodiment.
FIG. 2 is a three-dimensional distribution diagram of pulse parameters in the example.
Fig. 3 is a flow chart of an implementation of the present embodiment.
FIG. 4 is a visual range of a pulse sequence complex network modeling after dimension reduction by the method (1) and the method (2) division.
In FIG. 5: a is the node visual relationship diagram of Emitter1 obtained by the method (1), and b is the node visual relationship diagram of Emitter1 obtained by the method (2).
In FIG. 6: a is the node visual relationship diagram of Emitter2 obtained by the method (1), and b is the node visual relationship diagram of Emitter2 obtained by the method (2).
In FIG. 7: a is the node visual relationship diagram of Emitter3 obtained by the method (1), and b is the node visual relationship diagram of Emitter3 obtained by the method (2).
In fig. 8: a is the node visual relationship diagram of Emitter4 obtained by the method (1), and b is the node visual relationship diagram of Emitter4 obtained by the method (2).
FIG. 9 is a graph of the number of important nodes after modeling of the complex network of the radar radiation source pulse sequence in FIG. 4.
Fig. 10 is a distribution diagram of the feature vectors obtained by the method (1) after the dimensionality reduction.
Fig. 11 is a distribution diagram of the combined features after the feature vector dimensionality reduction obtained by the methods (1) and (2).
Detailed Description
The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the drawings in the embodiments of the present invention, and it is obvious that the described embodiments are only a part of the embodiments of the present invention, and not all of the embodiments. All other embodiments, which can be obtained by a person skilled in the art without making any creative effort based on the embodiments in the present invention, belong to the protection scope of the present invention.
As shown in fig. 1, the method for extracting features between pulses of a radar radiation source signal based on complex network modeling includes the following steps:
step S1, selecting three parameters of the radar pulse signal: carrier frequency RF, pulse width PW, pulse amplitude PA, pulse data parameter RF to be intercepted i 、PW i 、PA i Respectively associated with the pulse arrival time t i Form a two-dimensional sequence t i ,RF i }、{t i ,PW i }、{t i ,PA i And i represents a number variable of the number of pulses.
Because of the limitations of direction finding precision and direction finding resolution, the conditions and the technology for pulse de-interlacing or identification through high-precision DOA are not available at present, and the characteristics of a radar radiation source cannot be reflected, so that intercepted radar radiation source signals in the same direction are adopted for analysis in practical application; the PRI modulation of TOA parameters is more complex and important, which determines a working model of a radar, the parameters are generally analyzed individually in the radar radiation source signal processing, PA, PW and RF parameters can also randomly change and agile, but different radiation sources always change in fixed frequency bands, pulse amplitudes and pulse widths in a short time, a formed mode vector is an important combination of multi-parameter sorting and identification, although the space has larger overlap, the sorting and identification rate can be effectively improved through a complex network.
Step S2, setting the initial value N of the finite traversal range to 1, and applying the finite traversal visual map algorithm to map the two-dimensional sequence { t } to the two-dimensional sequence i ,RF i }、{t i ,PW i }、{t i ,PA i Mapping the acquired pulses to a complex network respectively, wherein the nodes of the complex network are the pulses acquired within the interception time; computing adjacency matrix M of each complex network RF 、M PW 、M PA Adding three adjacent matrixes to obtain a matrix M, wherein M is M RF +M PW +M PA 。
Each value M in the matrix M i,j There are four possibilities, M i,j 0,1,2,3 represents the total number of connections of pulse i and pulse j in the three complex networks.
Processing the matrix M by a multi-number table-solving method to obtain an adjacent matrix X, wherein elements in the matrix X are expressed as
X i,j Representing the degree of correlation between three pulse parameters, X i,j 1 denotes the interconnection between two pulse nodes, with an edge, X i,j And 0 means that two pulse nodes are not connected and have no edge, and the total complex network based on RF, PW and PA is obtained according to the adjacency matrix X.
Step S3, dividing and sampling the pulse nodes of the total complex network by using an equidistant division and a sliding window method to obtain two groups of pulse sequences, each pulse sequence including Q pulse subnetworks of fixed length, and dividing and sampling the pulse nodes of the total complex network according to M-2 l And M is more than or equal to 2 and less than or equal to n, the decomposition number M of the pulse subsequences is determined, each pulse subsequence is decomposed, and the two groups of pulse sequences are respectively decomposed into QxMAnd the sub-networks calculate the average degree vector of each sub-network to obtain two matrixes P.
Step S4, the two groups of QxM sub-networks obtained by the decomposition in the step S3 are respectively clustered, and the category separability measure of each clustering result is calculated
The category separability measure is used as an important index for evaluating the quality of feature extraction, a feature set with strong identification performance can be selected, and the smaller the value of the feature set is, the more effective the extracted feature set is in the aspect of sample classification.
Wherein S w Representing the mean vector divergence of the intra-class subnetworks byIn measuring the degree of hashing of samples with their mean, S b Representing the divergence of the vectors of the mean of the inter-class subnetwork, for measuring the degree of hashing of the class mean and the overall mean, S w 、S b Is calculated as follows:
wherein C represents the total number of the categories obtained by clustering, C represents the variable of the number of the categories, and N c Indicates the total number of subnetworks included in the class c, a indicates the number variable of subnetworks in each class,
represents the average degree vector of the a-th sub-network in the c-th class,
from the vector matrix P of the average degree, mu c Denotes the mean of all sub-network mean vectors in class c, μ denotes the mean of all sub-network mean vectors, and T denotes the transpose.
And step S5, increasing the limited crossing visual distances N one by one, repeating the steps S2-S4, constructing a corresponding complex network model, calculating the class separability measure J of the sub-network average degree vector, stopping iteration when the class separability measure J has a minimum value, and taking the limited crossing visual distance corresponding to the minimum value as the optimal limited crossing visual distance.
If the category separability measure is monotonous and no minimum value point exists, the steps S3-S5 are repeatedly executed, the division sampling process and the decomposition number of the pulse sequence are readjusted to calculate until the minimum value point appears.
And step S6, selecting the total complex network corresponding to the optimal limited crossing visual range to obtain the average degree vectors of the decomposed two sub-networks, and reducing the dimensions of the two average degree vector matrixes by adopting Principal Component Analysis (PCA) to obtain the one-dimensional characteristic with the maximum contribution degree, namely the finally obtained inter-pulse characteristic of the radar radiation source signal.
According to the embodiment of the invention, the pulse arrival time is taken as a bridge, and the radar radiation source inter-pulse signals are converted into network nodes and edges for modeling and analysis, so as to further excavate the local difference, relevance and effectiveness shown by a radar signal full-pulse sequence; the characteristic data of the full pulse sequence is acquired through division sampling and decomposition, the complexity of data processing is reduced, the processing efficiency is improved, and the local information and the system evolution rule of the pulse data are kept as far as possible; and the principal component analysis is used for reducing the dimension of the feature data extracted from the sub-network again, so that the aggregation degree of the network statistical feature distribution is increased.
The present invention also encompasses an electronic device comprising a memory for storing various computer program instructions and a processor for executing the computer program instructions to perform all or a portion of the steps recited above; the electronic device may communicate with one or more external devices, may also communicate with one or more devices that enable user interaction with the electronic device, and/or any device that enables the electronic device to communicate with one or more other computing devices, may also communicate with one or more networks (e.g., local area networks, wide area networks, and/or public networks) through a network adapter; the present invention also includes a computer-readable medium that stores a computer program, which can be executed by a processor, which can include, but is not limited to, magnetic storage devices, optical disks, digital versatile disks, smart cards, and flash memory devices, which in addition can represent one or more devices for storing information and/or other machine-readable media, which term "machine-readable media" includes, but is not limited to, wireless channels and various other media (and/or storage media) that can store, contain, and/or carry code and/or instructions and/or data.
Examples
The effectiveness of the method is verified by taking four intercepted unknown radar radiation source signals (marked as Emitter1, Emitter2, Emitter3 and Emitter4) as simulated pulse sequences, wherein 4 complex system radars all come from the same direction, and single pulse or pulse group agility and multi-pulse jumping exist in RF parameters among pulses; both PW and PA parameters are agile within a certain range; the PRI parameters also have slip, spread, and jitter.
Assuming that the total intercept time Tint is 58.4ms, the number of the included intercepted pulses is 2850, 8379, 3664 and 6141, respectively, the pulse loss rate is 10.5%, the typical measured signal-to-noise ratio is 15dB, and the numerical ranges of the three parameters of the 4 radar radiation sources are shown in table 1:
TABLE 1 Radar radiation Source pulse sequence types and parameters thereof
In order to analyze four groups of radar radiation source data more intuitively, three parameters of RF, PW and PA data among pulses are selected as basic data of complex network modeling, and as can be known from a three-dimensional distribution diagram of pulse parameters given by figure 2, the distribution ranges of the RF and PA of pulse signals from a radiation source Emitter1 are relatively dispersed, and other radar pulse signal data are completely overlapped on a two-dimensional plane formed by the projections of the RF and PA; the signal sequences of the radiation sources Emitter2 and Emitter4 are dispersed into a plurality of groups of clusters, projections on different dimensions are partially overlapped, Emititer 3 belongs to high repetition frequency dense signals, and as can be known from three-dimensional diagrams of various radiation sources, the distribution shapes of the radiation sources have larger differences, the difficulty of sorting and identification is increased, and more complex classifier design is needed.
1. Radar radiation source pulse sequence complex network domain transformation
Since the selection of the limited traversal line-of-sight N directly affects the sub-network averaging effectiveness, the present embodiment is based on the pulse parameter RF i 、PW i 、PA i Respectively constructing complex networks by two-dimensional sequences formed by the pulse arrival time, and determining an adjacency matrix of the total complex network based on the adjacency matrices of the complex networks; then, the pulse nodes of the total complex network are divided and sampled through the method (1) and the method (2), and then the average degree of each network is obtained through decompositionAnd (3) clustering the vectors according to the vector to obtain a class separability measure, wherein the class separability measure J changes along with the change of the limited crossing visual range N as shown in fig. 4, the limited crossing visual range is changed, the process is repeated, and the limited crossing visual range corresponding to the minimum value of J is selected as the optimal limited crossing visual range.
Method (1): the method comprises the steps of (1) carrying out dimension reduction on each type of radar radiation source pulse sequences according to arrival time sequence sampling, selecting 10 pulse subsequences (pulse sequences in each subsequence are allowed to have overlapped fragments) with the number of nodes being N-800, determining a limited crossing sight distance N based on the pulse subsequences, and further carrying out complex network construction;
method (2): and (3) selecting a sliding window and a step length by adopting a similarity technology to perform dimension reduction processing on the amplitude characteristics of the pulse sequence parameter values, obtaining 10 pulse subsequences with the number of nodes n being 800, and then performing complex network modeling.
In the method (1), the optimal limited traversal visual range N is 3, and in the method (2), the optimal limited traversal visual range N is 2.
By the method, local information and a system evolution rule of pulse data can be kept as far as possible on the basis of 4 pieces of radar inter-pulse sequence data, and a guarantee is provided for further analyzing interaction relations and topological structure characteristics among traditional radar inter-pulse and inter-pulse group sequence parameters in a complex system.
The general idea of the radar radiation source signal inter-pulse feature extraction method based on the complex network modeling is shown in fig. 3, the complex network model layer performs dimension reduction on a pulse sequence through a method (1) and a method (2) to obtain a 2 x 10 pulse subsequence complex network model; the sub-network average degree feature vector layer decomposes each complex network into 16 sub-networks, extracts the average degree vectors thereof, and forms 2 groups of 10 x 16 dimensional feature vector matrixes; finally, PCA dimensionality reduction is carried out on the 2 groups of matrixes respectively to obtain important characteristics with better separability, and a characteristic matrix of 10 x 2 is formed.
According to the optimal limited traversal sight distance N, an LPVG algorithm is adopted to model each radiation source pulse subsequence to obtain a complex network, Gephi software is used for carrying out visualization processing on the network, the displayed model is a weightless and directionless network model and is laid out as Fruchterman-Reinforld, the size of an important node in a network graph is determined by the node degree, and the larger the value is, the larger the node size is.
In order to further observe the difference of important nodes of a complex network model in a radiation source pulse sequence, in the experiment, the degree of each node is calculated, the node with the node degree k being more than or equal to 50 is defined as the important node, the number of the extracted important node is marked in a diagram a of fig. 5-8, a in fig. 5-8 respectively depicts a node visual relation diagram obtained after the dimension reduction of the pulse sequence of 4 radar radiation source signals is carried out by the method (1), the important nodes of the network corresponding to Emitter1 are 197, 252 and 28, the important nodes of Emitter2 are 739, 752, 804, 43 and 639, the important nodes of Emitter3 are 192, and the important nodes of Emitter4 are 779, 777, 249, 251, 488 and 375.
B in fig. 5 to 8 respectively depict node visual relationship diagrams obtained after dimension reduction is performed on pulse sequences of 4 radar radiation source signals by the method (2), in 4 complex networks obtained by the method (2), important nodes of a network corresponding to Emitter1 are 54 and 69, and important nodes of Emitter2 are 509, 463, 604, 93, 55, 142, 772, 263 and 270. The significant nodes of Emitter3 are 172, 89, and the significant nodes of Emitter4 are 309, 257, 52, 129, 154; as can be seen from the figure, the node degree distribution in the complex network model shown by different radiation source pulse sequences has different degrees of heterogeneity, and nodes with low degrees belong to compact subgraphs which are connected with each other through important nodes.
Based on this, the number of important nodes after the 4 radar radiation source pulse sequence complex network constructed by the method (1) and the method (2) is modeled is plotted in fig. 9, different symbols and sizes respectively represent 4 different radar radiation sources and the total number of the important nodes in the two methods, and it can be known from fig. 9 that the positions and the numbers of the important nodes selected in the complex network can reflect the difference of signals of the 4 radar radiation sources, thereby laying a good foundation for extracting network topology characteristics.
2. Sub-network averaging feature extraction and analysis thereof
Performing network domain transformation on the radar radiation source pulse sequence to obtain 2 groups of feature data respectively containing 10 complex network models, further uniformly decomposing each network model in each group into 16 sub-networks, extracting node average degree of each sub-network to form 2 groups of 10 x 16-dimensional feature vectors, performing PCA dimension reduction on the feature vectors obtained by the method (1), and taking the one-dimensional feature with the largest contribution degree as a horizontal coordinate and the one-dimensional feature with the second largest contribution degree as a vertical coordinate to obtain a feature distribution diagram shown in FIG. 10.
As can be seen from fig. 10, after the complex network modeling decomposition and the PCA dimension reduction, 4 different radar radiation source signal sample sets contain 40 data points, but from the two-dimensional feature distribution, the second-dimensional features generated by the method (1) are not cross-overlapped, the boundary division difference between the first-dimensional features is small, the introduction of the second-dimensional features does not increase the sample concentration, the separability of the sorting identification is difficult to improve, and the purpose of simplifying the classifier cannot be achieved.
Performing complex network modeling on the coarse-grained sampling sequence after the inter-pulse reconstruction generated by the method (2), and after the dimensionality reduction of the generated sub-network average degree feature is performed by PCA, taking the one-dimensional feature with the maximum contribution degree obtained by the method (1) as an abscissa and the one-dimensional feature with the maximum contribution degree obtained by the method (2) as an ordinate to draw a feature distribution map as shown in FIG. 11; as can be seen from fig. 11, different radar pulse signals do not overlap, and the aggregation is significantly improved, and has significant separability, where the sample aggregation of the radiation source Emitter2 is the highest, the sample aggregation of the radiation source Emitter1 is the lowest, and the sample aggregation of the radiation source Emitter3 is higher and farthest from other types of samples, which is consistent with the distribution of the original data in fig. 2.
3. Analysis of recognition results
In order to verify the effect of the sub-network average degree vector extracted in the experiment in radar radiation source signal sorting, a fuzzy C-means (FCM) clustering algorithm is adopted to classify the two-dimensional feature vectors, the classification number is selected to be 4, 40 sample data are independently and repeatedly subjected to experiments for 50 times to obtain an average correct classification value, and the obtained classification correct rate statistical data are shown in table 2.
TABLE 2 sorting identification accuracy statistical data sheet
Meanwhile, as the characteristic distribution corresponding to fig. 11 shows, because the Emitter1 and the Emitter4 are separated from each other by a relatively short distance, 3% of the Emitter1 signal data are misclassified to the Emitter4, and 2% of the Emitter4 signal data are misclassified to the Emitter1, the separation effect of the Emitter2 and Emitter3 signals is the best, the average accuracy is 98%, the separation effect of the Emitter1 signals is poor, and the average accuracy is 95%; however, in an independent classification experiment, the optimal classification result reaches 100% of accuracy, the overall average accuracy can reach 97%, and the average accuracy is compared with 40.7% of the average accuracy of direct classification of the original full pulse sequence of the radar radiation source, so that the features extracted by the embodiment have obvious advantages in the classification result, and the classification accuracy is improved by nearly 50%.
4. Algorithm complexity analysis
For the time complexity of the algorithm, according to the complex network modeling feature extraction algorithm shown in fig. 3, the implementation of the pulse sequence dividing methods (1) and (2) mainly depends on the calculation of obtaining a sliding window in the method (2), so that the method (2) determines the boundary algorithm complexity of the sliding window scale to be o (dff), d-3 represents the pulse parameter dimension, P represents the number of radiation source pulse sequences, and f represents the window scale; the complexity of the complex network modeling process and PCA dimension reduction is O (d (lNi) 2 +i 2 Where N is the number of finite traversal line-of-sight, l ═ 10 is the number divided by methods (1) and (2), i ═ 800 is the number of complex network pulse nodes, and M ═ 16 is the number of complex network decomposition into sub-networks; the classification complexity of the final FCM is O (PdC) 2 f) Where C ═ 4 represents the number of classifications.
Thus, the algorithm complexity of the present invention is substantially O (i) 2 ) In the text, the number of pulse nodes is also tested, i is 200, 400 and 600 respectively, but the classification effect obtained after feature extraction in the text is poor, and although the algorithm complexity of a small number of nodes is reduced, i is 800 which is obviously superiorFCM result of the original sequence and stable classification result meet the requirement of sorting and identifying accuracy under complex electronic countermeasure environment.
Aiming at radar radiation source signal pulse-to-pulse parameters which are flexible, variable, nonlinear and non-stable, the embodiment is based on a complex network construction algorithm of a full pulse sequence, the difference and the relevance shown by the network characteristic quantity of the pulse-to-pulse parameters are mined, the corresponding relation between the network statistical property and the pulse sequence to be sorted and identified is considered, and the method becomes an effective means for analyzing the radar pulse sequence characteristics under the conditions of serious pulse loss and low noise; in addition, as network science gradually permeates into the research of nonlinear time series, the conversion and mutual representation of a complex network and the time series are used as a support and a bridge, the difficulties of the domain transformation and representation of the complex network theory and a radar radiation source signal sequence are mainly solved, the radar radiation source signal is reviewed again from a complex network view angle, the change rule of a radar pulse sequence is favorably known and understood, and the requirement of modern electronic warfare countermeasure is met.
All the embodiments in the present specification are described in a related manner, and the same and similar parts among the embodiments may be referred to each other, and each embodiment focuses on the differences from the other embodiments. In particular, for the system embodiment, since it is substantially similar to the method embodiment, the description is simple, and for the relevant points, reference may be made to the partial description of the method embodiment.
The above description is only for the preferred embodiment of the present invention, and is not intended to limit the scope of the present invention. Any modification, equivalent replacement, or improvement made within the spirit and principle of the present invention shall fall within the protection scope of the present invention.
Claims (7)
1. The method for extracting the inter-pulse characteristics of the radar radiation source signal is characterized by comprising the following steps of:
s1, intercepting the pulse data parameter RF i 、PW i 、PA i Respectively with the pulse arrival time t i Form a two-dimensional sequence t i ,RF i }、{t i ,PW i }、{t i ,PA i }；
Where i represents a number variable of the number of pulses, RF i Indicating the carrier frequency, PW, of the ith pulse i Indicating the pulse width, PA, of the ith pulse i Represents the pulse amplitude of the ith pulse;
s2, determining the optimal finite traversal sight distance, and respectively connecting the two-dimensional sequences t i ,RF i }、{t i ,PW i }、{t i ,PA i Mapping the complex networks, and acquiring a matrix X by using an adjacent matrix of each complex network so as to acquire a total complex network of the radiation source;
s3, respectively adopting an equidistant division and sliding window method to carry out division sampling on pulse nodes of the total complex network to obtain two groups of pulse sequences, wherein each pulse sequence comprises Q pulse subsequences with fixed lengths, determining the decomposition number M of the subsequences, and respectively decomposing the two groups of pulse sequences into QxM sub-networks;
and S4, respectively calculating the average degree vectors of the two groups of pulse sequences corresponding to the sub-networks to obtain two average degree vector matrixes P, and reducing the dimension of the two average degree vector matrixes P by adopting principal component analysis to obtain two-dimensional characteristics with the maximum contribution degree, namely the extracted radar radiation source signals.
2. The method for extracting features between pulses of a radar radiation source signal according to claim 1, wherein the process of determining the optimal limited traversal line-of-sight in S2 is as follows:
setting the initial value N of the limited crossing sight distance as 1, repeating S2-S3 to obtain Q multiplied by M sub-networks corresponding to two groups of pulse sequences, and calculating the average degree vector of each sub-network;
clustering the two groups of sub-networks respectively, and calculating the class separability measure of each sub-network based on the average degree vector of each sub-network in the clustering result;
and increasing the limited crossing visual distances N one by one, repeatedly calculating the class separability measures of the sub-networks under different limited crossing visual distances, and taking the limited crossing visual distance corresponding to the minimum value of the class separability measures as the optimal limited crossing visual distance.
3. The method of claim 1, wherein the matrix X is obtained by the following steps:
adding adjacent matrixes of each complex network to obtain a matrix M, wherein the element M in the matrix M i,j ＝0,1,2,3，M i,j Representing the total connection number of the pulse i and the pulse j in the three complex networks;
using majority voting to pair M i,j Processing to obtain matrix X, wherein elements in matrix X
X i,j 1 indicates that there is an edge between two pulse nodes, X i,j 0 means that there is no edge between the two pulse nodes.
4. The method of claim 2, wherein the category separability measure J is calculated as follows:
wherein S w Representing the mean vector divergence, S, of the intra-class subnetwork b Representing the divergence of the vectors of the average degree of the sub-networks among the classes, C representing the total number of the classes obtained by clustering, C representing the variable of the number of the classes, N c Indicates the total number of subnetworks included in the class c, a indicates the number variable of subnetworks in each class,
represents the mean vector, μ, of the a-th sub-network in class c c Denotes the mean of all sub-network mean vectors in class c, μ denotes the mean of all sub-network mean vectors, and T denotes the transpose.
5. The method of claim 1, wherein the mean vector matrix is a matrix of the radar radiation source signal inter-pulse features
Where q denotes the number variable of the pulse subsequences, m denotes the number variable of the sub-networks,
respectively representing the average degree vectors of the 1 st, the M th and the M th sub-networks obtained by decomposing the 1 st pulse subsequence,
respectively representing the average degree vectors of the 1 st, the M th and the M th sub-networks obtained by decomposing the q-th pulse subsequence,
respectively representing the average degree vectors of the 1 st, the M th and the Mth sub-networks obtained by decomposing the Q-th pulse subsequence,
N m representing the total number of pulses in the mth subnet,
representing the node degree of the pulse i in the mth subnet, i.e., the number of pulses connected to the pulse i.
6. An electronic device is characterized by comprising a processor, a memory and a communication bus, wherein the processor and the memory are communicated with each other through the communication bus;
a memory for storing a computer program;
a processor for implementing the method steps of any one of claims 1 to 5 when executing a program stored in the memory.
7. A computer-readable storage medium, characterized in that a computer program is stored in the computer-readable storage medium, which computer program, when being executed by a processor, carries out the method steps of any one of the claims 1-5.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210587752.7A CN115034261B (en) | 2022-05-26 | 2022-05-26 | Method, equipment and storage medium for extracting inter-pulse characteristics of radar radiation source signals |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210587752.7A CN115034261B (en) | 2022-05-26 | 2022-05-26 | Method, equipment and storage medium for extracting inter-pulse characteristics of radar radiation source signals |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN115034261A true CN115034261A (en) | 2022-09-09 |
| CN115034261B CN115034261B (en) | 2023-08-22 |
Family
ID=83120381
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210587752.7A Active CN115034261B (en) | 2022-05-26 | 2022-05-26 | Method, equipment and storage medium for extracting inter-pulse characteristics of radar radiation source signals |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN115034261B (en) |
Cited By (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115659162A (en) * | 2022-09-15 | 2023-01-31 | 云南财经大学 | Method, system and equipment for extracting features in radar radiation source signal pulse |
| CN117129947A (en) * | 2023-10-26 | 2023-11-28 | 成都金支点科技有限公司 | Planar transformation method radar signal identification method based on mininet |
| CN117272086A (en) * | 2023-11-22 | 2023-12-22 | 中国电子科技集团公司第二十九研究所 | Radar signal scanning envelope segmentation method based on DBSCAN |
Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107727749A (en) * | 2017-08-30 | 2018-02-23 | 南京航空航天大学 | A kind of ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm |
| CN108875205A (en) * | 2018-06-15 | 2018-11-23 | 北京航空航天大学 | System availability efficient simulation method based on reachable matrix and discrete event driving |
| CN111707995A (en) * | 2020-06-19 | 2020-09-25 | 雷震烁 | Radar antenna scanning mode identification method based on limited-penetration visual effect |
| CN112183659A (en) * | 2020-10-15 | 2021-01-05 | 电子科技大学 | Unknown signal radiation source identification method based on convolutional neural network |
| CN112949383A (en) * | 2021-01-22 | 2021-06-11 | 中国人民解放军63892部队 | Waveform agility radar radiation source identification method based on Hydeep-Att network |
| CN114219483A (en) * | 2021-12-14 | 2022-03-22 | 云南财经大学 | Method, equipment and storage medium for sharing block chain data based on LWE-CPBE |
-
2022
- 2022-05-26 CN CN202210587752.7A patent/CN115034261B/en active Active
Patent Citations (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN107727749A (en) * | 2017-08-30 | 2018-02-23 | 南京航空航天大学 | A kind of ultrasonic quantitative detection method based on wavelet packet fusion feature extraction algorithm |
| CN108875205A (en) * | 2018-06-15 | 2018-11-23 | 北京航空航天大学 | System availability efficient simulation method based on reachable matrix and discrete event driving |
| CN111707995A (en) * | 2020-06-19 | 2020-09-25 | 雷震烁 | Radar antenna scanning mode identification method based on limited-penetration visual effect |
| CN112183659A (en) * | 2020-10-15 | 2021-01-05 | 电子科技大学 | Unknown signal radiation source identification method based on convolutional neural network |
| CN112949383A (en) * | 2021-01-22 | 2021-06-11 | 中国人民解放军63892部队 | Waveform agility radar radiation source identification method based on Hydeep-Att network |
| CN114219483A (en) * | 2021-12-14 | 2022-03-22 | 云南财经大学 | Method, equipment and storage medium for sharing block chain data based on LWE-CPBE |
Non-Patent Citations (2)
| Title |
|---|
| 田甜;温广瑞;张志芬;徐斌;: "一种新的复杂网络建模和特征提取方法及应用", 振动.测试与诊断, no. 06 * |
| 陈彬;童创明;李西敏;: "基于动态RCS的典型飞机目标识别", 现代雷达, no. 01 * |
Cited By (6)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN115659162A (en) * | 2022-09-15 | 2023-01-31 | 云南财经大学 | Method, system and equipment for extracting features in radar radiation source signal pulse |
| CN115659162B (en) * | 2022-09-15 | 2023-10-03 | 云南财经大学 | Method, system and equipment for extracting intra-pulse characteristics of radar radiation source signals |
| CN117129947A (en) * | 2023-10-26 | 2023-11-28 | 成都金支点科技有限公司 | Planar transformation method radar signal identification method based on mininet |
| CN117129947B (en) * | 2023-10-26 | 2023-12-26 | 成都金支点科技有限公司 | Planar transformation method radar signal identification method based on mininet |
| CN117272086A (en) * | 2023-11-22 | 2023-12-22 | 中国电子科技集团公司第二十九研究所 | Radar signal scanning envelope segmentation method based on DBSCAN |
| CN117272086B (en) * | 2023-11-22 | 2024-02-13 | 中国电子科技集团公司第二十九研究所 | Radar signal scanning envelope segmentation method based on DBSCAN |
Also Published As
| Publication number | Publication date |
|---|---|
| CN115034261B (en) | 2023-08-22 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN115034261A (en) | Method and equipment for extracting features between pulses of radar radiation source signal and storage medium | |
| CN111551992B (en) | Rock reservoir structure characterization method and device, computer-readable storage medium and electronic equipment | |
| Dudczyk | Radar emission sources identification based on hierarchical agglomerative clustering for large data sets | |
| Gupta et al. | Electronic warfare: Issues and challenges for emitter classification | |
| CN112052755A (en) | Semantic convolution hyperspectral image classification method based on multi-path attention mechanism | |
| CN114429156A (en) | Radar interference multi-domain feature countermeasure learning and detection identification method | |
| CN114595732B (en) | Radar radiation source sorting method based on depth clustering | |
| Liu et al. | Topological properties and fractal analysis of a recurrence network constructed from fractional Brownian motions | |
| CN108171119B (en) | SAR image change detection method based on residual error network | |
| CN106097290A (en) | SAR image change detection based on NMF image co-registration | |
| CN107577742A (en) | A kind of more relation social network patterns method for digging based on bayes method | |
| CN113920255B (en) | High-efficient mapping system based on point cloud data | |
| CN114021458A (en) | Small sample radar radiation source signal identification method based on parallel prototype network | |
| CN116797846A (en) | Method and device for identifying small sample radar radiation source based on RoAtten-PN network | |
| CN110956221A (en) | Small sample polarization synthetic aperture radar image classification method based on deep recursive network | |
| CN113361204B (en) | Radar radiation source PRI modulation identification method based on SAE | |
| CN113869289B (en) | Multi-channel ship radiation noise feature extraction method based on entropy | |
| CN113419278B (en) | Well-seismic joint multi-target simultaneous inversion method based on state space model and support vector regression | |
| Tsonis et al. | Searching for determinism in observed data: a review of the issues involved | |
| Wang et al. | FCM algorithm and index CS for the signal sorting of radiant points | |
| CN112666528A (en) | Multi-station radar system interference identification method based on convolutional neural network | |
| Das et al. | Non-parametric Bayesian mixture of sparse regressions with application towards feature selection for statistical downscaling | |
| CN111083632A (en) | Ultra-wideband indoor positioning method based on support vector machine | |
| Chen et al. | An integrated graph Laplacian downsample (IGLD)-based method for DEM generalization | |
| Sunori et al. | Classification of rainfall levels using various machine learning techniques |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |