CN112036074B - Radar signal sorting method and system in high pulse density environment - Google Patents
Radar signal sorting method and system in high pulse density environment Download PDFInfo
- Publication number
- CN112036074B CN112036074B CN202010731268.8A CN202010731268A CN112036074B CN 112036074 B CN112036074 B CN 112036074B CN 202010731268 A CN202010731268 A CN 202010731268A CN 112036074 B CN112036074 B CN 112036074B
- Authority
- CN
- China
- Prior art keywords
- pulse
- sorting
- radar
- radar signal
- toa
- 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.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 38
- 238000001208 nuclear magnetic resonance pulse sequence Methods 0.000 claims abstract description 29
- 230000015654 memory Effects 0.000 claims abstract description 26
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 21
- 238000013528 artificial neural network Methods 0.000 claims abstract description 19
- 238000013507 mapping Methods 0.000 claims abstract description 11
- 230000009466 transformation Effects 0.000 claims description 15
- 210000002569 neuron Anatomy 0.000 claims description 13
- 210000004205 output neuron Anatomy 0.000 claims description 12
- 238000005070 sampling Methods 0.000 claims description 9
- 238000012549 training Methods 0.000 claims description 8
- 238000007689 inspection Methods 0.000 claims description 6
- 238000012545 processing Methods 0.000 claims description 5
- 238000004590 computer program Methods 0.000 claims description 4
- 230000008859 change Effects 0.000 claims description 3
- 230000006870 function Effects 0.000 claims description 3
- 238000012163 sequencing technique Methods 0.000 claims description 3
- 230000000694 effects Effects 0.000 description 8
- 230000005855 radiation Effects 0.000 description 8
- 238000005516 engineering process Methods 0.000 description 5
- 238000012360 testing method Methods 0.000 description 4
- 230000007547 defect Effects 0.000 description 3
- 230000004927 fusion Effects 0.000 description 3
- 230000006872 improvement Effects 0.000 description 3
- 238000001228 spectrum Methods 0.000 description 3
- 238000003491 array Methods 0.000 description 2
- 230000001537 neural effect Effects 0.000 description 2
- 238000011160 research Methods 0.000 description 2
- 239000013598 vector Substances 0.000 description 2
- 238000004458 analytical method Methods 0.000 description 1
- 230000008485 antagonism Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 230000006835 compression Effects 0.000 description 1
- 238000007906 compression Methods 0.000 description 1
- 238000013461 design Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 238000011156 evaluation Methods 0.000 description 1
- 238000000605 extraction Methods 0.000 description 1
- 230000007787 long-term memory Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 238000007500 overflow downdraw method Methods 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 230000008569 process Effects 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
- 230000006403 short-term memory Effects 0.000 description 1
- 238000004088 simulation Methods 0.000 description 1
- 238000012706 support-vector machine Methods 0.000 description 1
Classifications
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F30/00—Computer-aided design [CAD]
- G06F30/20—Design optimisation, verification or simulation
- G06F30/27—Design optimisation, verification or simulation using machine learning, e.g. artificial intelligence, neural networks, support vector machines [SVM] or training a model
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/23—Clustering techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/24—Classification techniques
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/04—Architecture, e.g. interconnection topology
- G06N3/049—Temporal neural networks, e.g. delay elements, oscillating neurons or pulsed inputs
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
- G06N3/00—Computing arrangements based on biological models
- G06N3/02—Neural networks
- G06N3/08—Learning methods
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F2218/00—Aspects of pattern recognition specially adapted for signal processing
- G06F2218/12—Classification; Matching
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Evolutionary Computation (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- General Engineering & Computer Science (AREA)
- Life Sciences & Earth Sciences (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Software Systems (AREA)
- Mathematical Physics (AREA)
- Computing Systems (AREA)
- Evolutionary Biology (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Bioinformatics & Computational Biology (AREA)
- General Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Health & Medical Sciences (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Computer Networks & Wireless Communication (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Medical Informatics (AREA)
- Computer Hardware Design (AREA)
- Geometry (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention belongs to the technical field of radar signal sorting in electronic countermeasure, and discloses a radar signal sorting method and a radar signal sorting system in a high pulse density environment, wherein received radar pulse descriptive words are divided according to pulse arrival angles to obtain a plurality of groups of pulse sequences from different directions; inputting the pulse frequency domain parameters and pulse width of each group of pulse sequences into a cascade structure self-organizing mapping neural network to obtain a primary sorting result; an improved sequence difference histogram algorithm is used for the primary sorting result to obtain a main sorting result; and traversing the combination condition of the main sorting result, inputting the pulse frequency domain parameters, the pulse width and the pulse arrival time into a long-short time memory network for repeated frequency identification, and combining to obtain the repeated frequency group variable radar. The invention can effectively realize the sorting of common radar signals (fixed heavy frequency, staggered heavy frequency, heavy frequency jitter, sliding heavy frequency, group heavy frequency variation and the like) in a high pulse density environment, and can realize the real-time sorting of radar signals in an actual electromagnetic environment.
Description
Technical Field
The invention belongs to the technical field of radar signal sorting in electronic countermeasure, and particularly relates to a radar signal sorting method, a radar signal sorting system and application in a high pulse density environment.
Background
Currently, radar signal sorting is generally performed by evaluating and acquiring the number of radar radiation sources existing in a current electromagnetic environment. With the continuous increase in radar technology demand and the continuous innovation and development of various processing means and techniques for modern signals, many new-system radars are continuously developed and applied in practice. In particular, the application of novel radars such as pulse compression radars and agile radars, makes the specific location information of the radar not effectively detectable, especially in increasingly complex electromagnetic environments. In order to effectively improve the sorting capability of radar signals, a good evaluation effect on a complex electromagnetic combat environment is ensured. At present, radar signal sorting in the radar countermeasure field mainly focuses on sorting algorithms based on intra-pulse information, and sorting research for inter-pulse information remains at an earlier research level. First of the prior art, a minimum L is disclosed 1 The radar signal sorting method under the norm has higher sorting efficiency aiming at the problem of radar signal sorting under the highly dense and complex signal environment, but the method is required to have stronger correlation or matching property with the input signal, and has certain adaptability, can self-evolve, and has low radar signal sorting accuracy and reliability when the correlation between the signal and the overcomplete dictionary is lower; in the second prior art, a multimode radar signal sorting method based on data field hierarchical clustering is disclosed, the method searches the maximum value of a local potential value by calculating the potential value of the data field, selects sample data closest to the maximum value as an initial clustering center, and then clusters by using a traditional clustering algorithm, so that the sorting efficiency is higher for the radar signal sorting problem under the highly dense and complex signal environment, but the operation complexity of the potential value of the data field is higher, the calculation amount is larger, and the sorting instantaneity is lower; in the third prior art, a radar radiation source signal sorting method based on sample entropy is disclosed, and the method is implemented by the sample entropy and the power spectrum entropy characteristics of the radar radiation source signalThe extraction and support vector machine is used for classifying, the overall sorting rate is good, and the distribution of similar vectors in the sequence and the influence of the complexity of the sequence vectors on the time sequence complexity are not considered when the sample entropy calculates the time sequence complexity, so that the single signal recognition rate is not high when the signal-to-noise ratio is low. The fourth prior art discloses a method for applying information fusion to radar signal sorting, which performs data level fusion on pulse description words before radar signal sorting, performs feature level fusion on sorting results after sorting, unifies parameters describing the same radar and sorts the sorting results in credibility, and solves sorting failure possibly existing in the loss of single receiving equipment receiving pulse, but because the D-S data fusion method cannot accurately judge under the special condition that provided evidence is very large in direction conflict, sorting errors are likely to occur when radar parameters overlap, and further improvement is needed; in the fifth prior art, the spread histogram is generated by using the pulse arrival time difference, and then the pulse sequences of different radars are recursively sorted, and the method weakens the influence of noise through the spread operation, so that the sorting effect is improved. The histogram method is suitable for processing the aliasing radar signals with low pulse density and less pulse loss, but the time division effect is drastically reduced when the pulse density and the pulse loss rate are increased; in the sixth prior art, the arrival time difference of the radar sequence is converted into a PRI spectrum, the PRI value is locked through the position of a spectrum peak, and the PRI box with a non-constant width is adopted to improve the precision and time consumption of PRI estimation, but the precision and time consumption can not meet the practical requirements when the method is used for real-time sorting in an actual combat environment; in the seventh prior art, the carrier frequency and the arrival angle of the radar signal are clustered within a certain tolerance range, and the average value of the clustering results of the two parameters of the pulse width and the arrival angle is further utilized to perform secondary clustering, but the technology relies on accurate estimation of the pulse arrival angle, and when the pulse arrival angle estimation is inaccurate or cannot be estimated, the effectiveness of the technology is greatly reduced. The prior art, namely the prior art seven, solves the problem of radar signal sorting to a certain extent, and the more common problem is that the process of extracting the intra-pulse features is higher due to the higher frequency of the radar signalsThis feature is not suitable for real-time sorting of radar signals, which is complex and time consuming. In terms of inter-pulse features, when a plurality of algorithms are used for solving the sorting problem in a high-density pulse environment, the effectiveness of the algorithms is rapidly reduced, and the existing radar with complex modulation in repeated frequency has no good sorting effect. In addition, the sorting technology is complex to implement, and the batch increasing phenomenon is easy to generate.
Through the above analysis, the problems and defects existing in the prior art are as follows: the existing radar signal sorting method has poor sorting real-time performance, cannot sort the radar modulated by complex repetition frequency, easily generates the problem of batch increase, and has lower reliability.
The difficulty of solving the problems and the defects is as follows:
in a pulse dense environment, real-time sorting of radar signals is realized, and a repeated frequency group variable radar cannot be batched into multiple repeated frequency fixed radars, so that the problem of batched is formed.
The meaning of solving the problems and the defects is as follows:
in an actual electronic countermeasure environment, along with the improvement of the technology level, the electromagnetic environment is more and more complex. This results in a particularly important determination of the number of radiation sources in the current environment, and solving the above problems can improve the antagonism activity to a certain extent, which is of great importance for subsequent radiation source identification, threat level assessment, etc.
Disclosure of Invention
Aiming at the problems existing in the prior art, the invention provides a radar signal sorting method, a radar signal sorting system and application in a high pulse density environment.
The invention is realized in such a way, a radar signal sorting method under a high pulse density environment is realized, and the radar signal sorting method under the high pulse density environment divides the received radar pulse description words according to the pulse arrival angles to obtain a plurality of groups of pulse sequences from different directions; inputting the pulse frequency domain parameters and pulse width of each group of pulse sequences into a cascade structure self-organizing mapping neural network to obtain a primary sorting result; an improved sequence difference histogram algorithm is used for the primary sorting result to obtain a main sorting result; and traversing the combination condition of the main sorting result, inputting the pulse frequency domain parameters, the pulse width and the pulse arrival time into a long and short time memory network (LSTM) for repeated frequency identification, and combining to obtain the repeated frequency group variable radar.
Further, the received radar pulse descriptors are divided according to the pulse arrival angles, and the method for obtaining the multiple groups of pulse sequences from different directions specifically comprises the following steps:
1) Detecting an object P which is not checked in a database, if the object P is not processed and is classified into a certain cluster or marked as noise, checking a neighborhood of the object P, if the number of the included objects is not less than the minimum allowable number of objects minPts, establishing a new cluster C, and adding all points in the new cluster C into a candidate set N;
2) Checking the neighborhood of all the objects q which are not processed in the candidate set N, and adding the objects into N if the objects at least contain minPts objects; if q does not fall into any cluster, adding q to C;
3) Repeating the step 2), and continuously checking the unprocessed object in the N, wherein the current candidate set N is empty;
4) Repeating steps 1) to 3) until all objects fall into a certain cluster or are marked as noise.
Further, the cascade self-organizing map neural network primary sorting steps are as follows:
1) Initializing. Determining thresholds for determining neuronal merger or divisionσ i Corresponding to the error of each parameter value obtained by the reconnaissance receiver; setting an initial value m of the number of output neurons of the self-organizing map neural network 0 The maximum number of cycles allowed is K, and neurons undergo one-time merging or splitting into one cycle;
2) Training by using a traditional self-organizing map learning algorithm to enable the self-organizing map learning algorithm to reach an ordered map to obtain an initial clustering center;
3) Calculating the average distance in each classAnd inter-class distance D j ＝||m j -m j+1 | (j=1, 2,., c-1) and comparing with a set threshold R; if d j > R, then neuron j splits; if D j If R is less than R, merging the neurons j, determining whether two types are merged into one type or split into two types, and adjusting the scale of the self-organizing map neural network to obtain the number of new output neurons, namely m i Specific numerical values; if all output neurons are neither merged nor split, go to step 5); otherwise, go to step 4);
4) Judging whether the circulation turns are finished or not, if so, turning to the step 5); otherwise, turning to the step 2);
5) Calculating the value J of the clustering criterion function 1,m Obtaining various clustering center values;
6) Calculating J corresponding to the number of output neurons m+1 and m-1 1,m+1 And J 1,m-1 And sum J 1,m Comparing, taking max (J 1,m+1 ,J 1,m ,J 1,m-1 ) The corresponding neuron number is the final result, and various clustering center values are obtained.
Further, the main sorting steps of the improved sequence difference histogram algorithm are as follows:
1) Inputting a pulse arrival time sequence to be sorted;
2) Performing PRI classification statistics of all possible pulse repetition intervals;
3) Setting a statistical threshold epsilon, and sequencing and de-duplicating PRI larger than the threshold;
4) Traversing PRI passing through threshold;
5) Traversing all pulse arrival times TOAs;
6) Calculating the allowable time range [ TOA+PRI-mu, TOA+PRI+mu ] of the PRI under the current arrival time according to the noise tolerance mu;
7) Judging whether pulses exist in the allowable time range, if so, continuing to execute the step 6) by TOA=TOA', otherwise, increasing the missing pulse number by a misscount++;
8) Judging whether the missing pulse number misscunt reaches a set maximum value misscunt_max, if so, setting the missing pulse number to zero, changing TOA into TOA and executing step 5), otherwise, continuing to execute step 6) by TOA=TOA';
9) And carrying out subharmonic inspection on all the extracted TOA sequences, and combining the extracted TOA sequences to meet inspection rules so as to obtain a final main sorting result.
Further, the steps of the long-short time memory network for carrying out the identification of the repeated frequency group change are as follows:
1) Carrying out combined sorting on a radar Pulse Description Word (PDW) sequence according to the pulse arrival time obtained by main sorting;
2) Determining parameters of a segmented random feature sampling method: segment length k, number of segments d and interval g between segments;
3) Randomly selecting k sequence data from the PDW sequence as first segment characteristic data;
4) Judging whether the data is sampled; if yes, executing the step 4), otherwise, executing the step 2) by adding the interval g between the segments to the sampling start position;
5) Inputting the sampling result into a long-short-time memory network after training is completed, and obtaining a judgment result of the repetition frequency group variation;
6) If the repetition frequency group transformation rule is met, the PDW sequences are combined to form a repetition frequency group transformation radar pulse sequence, otherwise, the PDW sequences are not combined.
It is a further object of the present invention to provide a computer device comprising a memory and a processor, the memory storing a computer program which, when executed by the processor, causes the processor to perform the steps of: dividing the received radar pulse description words according to the pulse arrival angles to obtain a plurality of groups of pulse sequences from different directions; inputting the pulse frequency domain parameters and pulse width of each group of pulse sequences into a cascade structure self-organizing mapping neural network to obtain a primary sorting result; an improved sequence difference histogram algorithm is used for the primary sorting result to obtain a main sorting result; and traversing the combination condition of the main sorting result, inputting the pulse frequency domain parameters, the pulse width and the pulse arrival time into a long and short time memory network (LSTM) for repeated frequency identification, and combining to obtain the repeated frequency group variable radar.
Another object of the present invention is to provide a computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the steps of: dividing the received radar pulse description words according to the pulse arrival angles to obtain a plurality of groups of pulse sequences from different directions; inputting the pulse frequency domain parameters and pulse width of each group of pulse sequences into a cascade structure self-organizing mapping neural network to obtain a primary sorting result; an improved sequence difference histogram algorithm is used for the primary sorting result to obtain a main sorting result; and traversing the combination condition of the main sorting result, inputting the pulse frequency domain parameters, the pulse width and the pulse arrival time into a long and short time memory network (LSTM) for repeated frequency identification, and combining to obtain the repeated frequency group variable radar.
Another object of the present invention is to provide a radar signal sorting system under a high pulse density environment for operating the radar signal sorting method under a high pulse density environment, the radar signal sorting system under a high pulse density environment comprising:
the high-density pulse sparse module is used for carrying out sparse on the high-density pulse environment to obtain a multi-channel pulse sequence for parallel processing;
the cascade self-organizing mapping neural network primary sorting module is used for carrying out primary clustering sorting on radar pulses to obtain a clustering result of primary sorting of current radar signals;
the improved sequence difference histogram main sorting module is used for carrying out main sorting on the sparse radar pulse sequence to obtain a main sorting result of the repeated frequency fixed, repeated frequency sliding and repeated frequency dithering radar signal;
and the long-and-short-term memory network repetition frequency group transformation identification module is used for identifying and combining radar pulse sequences conforming to a repetition frequency group transformation rule in the repetition frequency fixed radar, and finally obtaining a repetition frequency group transformation radar signal sorting result.
Another object of the present invention is to provide a radar signal sorting system, which is equipped with the radar signal sorting system in a high pulse density environment.
It is another object of the present invention to provide a radar equipped with the radar signal sorting system in a high pulse density environment.
By combining all the technical schemes, the invention has the advantages and positive effects that: the invention can effectively realize the sorting of common radar signals (fixed repetition frequency, staggered repetition frequency, repeated frequency jitter, repeated frequency sliding variation, repeated frequency group variation, fixed pulse width, rapid frequency pulse interval variation, rapid frequency pulse group variation, simultaneous frequency diversity and time-sharing frequency diversity) in a high pulse density environment, and can realize the real-time sorting of radar signals in an actual electromagnetic environment.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present application, the following description will briefly explain the drawings needed in the embodiments of the present application, and it is obvious that the drawings described below are only some embodiments of the present application, and that other drawings can be obtained according to these drawings without inventive effort for a person skilled in the art.
Fig. 1 is a flowchart of a radar signal sorting method in a high pulse density environment according to an embodiment of the present invention.
FIG. 2 is a schematic diagram of a radar signal sorting system in a high pulse density environment according to an embodiment of the present invention;
in the figure: 1. a high-density pulse sparse module; 2. the cascade self-organizing mapping neural network primary sorting module; 3. a main sorting module of the improved sequence difference histogram; 4. and a long-short-term memory network repetition frequency group change identification module.
Fig. 3 to fig. 6 are graphs of clustering results after the cascade structure self-organizing map network according to the embodiment of the present invention.
Fig. 7 and fig. 8 are graphs of test results after training of a long-short-term memory network according to an embodiment of the present invention.
Detailed Description
The present invention will be described in further detail with reference to the following examples in order to make the objects, technical solutions and advantages of the present invention more apparent. It should be understood that the specific embodiments described herein are for purposes of illustration only and are not intended to limit the scope of the invention.
Aiming at the problems existing in the prior art, the invention provides a radar signal sorting method, a radar signal sorting system and application in a high pulse density environment, and the invention is described in detail below with reference to the accompanying drawings.
As shown in fig. 1, the radar signal sorting method under the high pulse density environment provided by the invention comprises the following steps:
s101: dividing radar radiation source pulse descriptors according to pulse arrival angles to obtain a plurality of groups of pulse sequences from different directions;
s102: inputting the pulse frequency domain parameters and pulse width of each group of pulse sequences into a cascade self-organizing mapping neural network to obtain a primary sorting result;
s103: an improved sequence difference histogram algorithm is used for the primary sorting result to obtain a main sorting result;
s104: and traversing the combination condition of the main sorting result, inputting the pulse frequency domain parameters, the pulse width and the pulse arrival time into a long-short time memory network for repeated frequency identification, and combining to obtain the repeated frequency group variable radar.
Other steps may be performed by those skilled in the art of the radar signal sorting method in a high pulse density environment provided by the present invention, and the radar signal sorting method in a high pulse density environment provided by the present invention of fig. 1 is merely a specific embodiment.
As shown in fig. 2, the radar signal sorting system under the high pulse density environment provided by the invention comprises:
the high-density pulse sparse module 1 is used for carrying out sparse on a high-density pulse environment to obtain a multichannel pulse sequence for parallel processing;
the cascade self-organizing map neural network primary sorting module 2 is used for carrying out primary clustering sorting on radar pulses to obtain a clustering result of primary sorting of current radar signals;
the improved sequence difference histogram main sorting module 3 is used for carrying out main sorting on the sparse radar pulse sequence to obtain a main sorting result of the heavy frequency fixed, heavy frequency sliding and heavy frequency dithering radar signal;
and the long-short-term memory network repetition frequency group transformation identification module 4 is used for identifying and combining radar pulse sequences conforming to a repetition frequency group transformation rule in the repetition frequency fixed radar, and finally obtaining a repetition frequency group transformation radar signal sorting result.
The technical scheme of the invention is further described below with reference to the accompanying drawings.
The embodiment of the invention provides a radar signal sorting method and a radar signal sorting system under a high pulse density environment, which specifically comprise the following steps:
step one, the received radar pulse description word is divided according to the pulse arrival angle, and the obtaining of a plurality of groups of pulse sequences from different directions specifically comprises the following steps:
1) Detecting an object P which is not checked in a database, checking a neighborhood of the object P if the object P is not processed (classified into a certain cluster or marked as noise), establishing a new cluster C if the number of the included objects is not less than the minimum allowable number (minPts), and adding all points in the new cluster C into a candidate set N;
2) Checking the neighborhood of all the objects q which are not processed in the candidate set N, and adding the objects into N if the objects at least contain minPts objects; if q does not fall into any cluster, adding q to C;
3) Repeating the step 2), and continuously checking the unprocessed object in the N, wherein the current candidate set N is empty;
4) Repeating steps 1) to 3) until all objects fall into a certain cluster or are marked as noise.
Inputting pulse frequency domain parameters and pulse width of each group of pulse sequences into a cascade self-organizing mapping neural network to obtain a primary sorting result:
1) Initializing. Determining thresholds for determining neuronal merger or divisionσ i Corresponding to the error in the values of the parameters obtained by the scout receiver. Setting an initial value m of the number of output neurons of the self-organizing map neural network 0 The most number of cycles allowed is K (here the neurons undergo a merger or split into cycles);
2) Training by using a traditional self-organizing map learning algorithm to enable the self-organizing map learning algorithm to reach an ordered map to obtain an initial clustering center;
3) Calculating the average distance in each classAnd inter-class distance D j ＝||m j -m j+1 | (j=1, 2,., c-1) and compared to a set threshold R. If d j > R, then neuron j splits; if D j If R is less than R, merging the neurons j to determine whether two types are merged into one type or split into two types, and adjusting the scale of the self-organizing map neural network to obtain the number of new output neurons, namely m i Specific numerical values; if all output neurons are neither merged nor split, go to step 5); otherwise, go to step 4);
4) Judging whether the circulation turns are finished or not, if so, turning to the step 5); otherwise, turning to the step 2);
5) Calculating the value J of the clustering criterion function 1,m Obtaining various clustering center values;
6) Calculating J corresponding to the number of output neurons m+1 and m-1 1,m+1 And J 1,m-1 And sum J 1,m Comparing, taking max (J 1,m+1 ,J 1,m ,J 1,m-1 ) The corresponding neuron number is the final result, and various clustering center values are obtained.
Thirdly, an improved sequence difference histogram algorithm is used for the primary sorting result, and a main sorting result is obtained:
1) Inputting a pulse arrival time sequence to be sorted;
2) Performing all possible Pulse Repetition Interval (PRI) classification statistics;
3) Setting a statistical threshold epsilon, and sequencing and de-duplicating PRI larger than the threshold;
4) Traversing PRI passing through threshold;
5) Traversing all pulse arrival Times (TOAs);
6) Calculating the allowable time range [ TOA+PRI-mu, TOA+PRI+mu ] of the PRI under the current arrival time according to the noise tolerance mu;
7) Judging whether pulses exist in the allowable time range, if so, continuing to execute the step 6) by TOA=TOA', otherwise, increasing the missing pulse number by a misscount++;
8) Judging whether the missing pulse number misscunt reaches a set maximum value misscunt_max, if so, setting the missing pulse number to zero, changing TOA into TOA and executing step 5), otherwise, continuing to execute step 6) by TOA=TOA';
9) And carrying out subharmonic inspection on all the extracted TOA sequences, and combining the extracted TOA sequences to meet inspection rules so as to obtain a final main sorting result.
Step four, traversing the combination condition of the main sorting result, inputting the pulse frequency domain parameters, the pulse width and the pulse arrival time into a long-short-time memory network for repeated frequency identification, and combining to obtain a repeated frequency group variable radar:
1) Carrying out combined sorting on a radar Pulse Description Word (PDW) sequence according to the pulse arrival time obtained by main sorting;
2) Determining parameters of a segmented random feature sampling method: segment length k, number of segments d and interval g between segments;
3) Randomly selecting k sequence data from the PDW sequence as first segment characteristic data;
4) And judging whether the data is sampled. If yes, executing the step 4), otherwise, executing the step 2) by adding the interval g between the segments to the sampling start position;
5) Inputting the sampling result into a long-short-time memory network after training is completed, and obtaining a judgment result of the repetition frequency group variation;
6) If the repetition frequency group transformation rule is met, the PDW sequences are combined to form a repetition frequency group transformation radar pulse sequence, otherwise, the PDW sequences are not combined.
The technical effects of the present invention will be described in detail with reference to simulation.
To test the clustering effect of the cascade structure self-organizing map network, 50 groups of real acquisition signals with different time periods are randomly selected, as shown in fig. 3-6.
The subsequent long-short-term memory network algorithm requires that the number of the clustering result categories is not less than the number of the repeated frequency group variable radar radiation sources. And (5) evaluating 50 groups of clustering effects by combining standard sorting reference results. The 48 groups meet the requirement of the subsequent algorithm, namely the number of the clustering result categories is not less than the number of the variable radar radiation sources of the repeated frequency group, and the total accuracy is 96%.
In order to test the judging performance of the long-short time memory network, 136 time periods of the actual acquisition signal are sorted to obtain 63 groups of effective radar repetition frequency group variable pulse description word sequences. Training and testing of long and short term memory networks was performed using the pulse descriptor sequence structured data set described above, as shown in fig. 7 and 8.
It should be noted that the embodiments of the present invention can be realized in hardware, software, or a combination of software and hardware. The hardware portion may be implemented using dedicated logic; the software portions may be stored in a memory and executed by a suitable instruction execution system, such as a microprocessor or special purpose design hardware. Those of ordinary skill in the art will appreciate that the apparatus and methods described above may be implemented using computer executable instructions and/or embodied in processor control code, such as provided on a carrier medium such as a magnetic disk, CD or DVD-ROM, a programmable memory such as read only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The device of the present invention and its modules may be implemented by hardware circuitry, such as very large scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, etc., or programmable hardware devices such as field programmable gate arrays, programmable logic devices, etc., as well as software executed by various types of processors, or by a combination of the above hardware circuitry and software, such as firmware.
The foregoing is merely illustrative of specific embodiments of the present invention, and the scope of the invention is not limited thereto, but any modifications, equivalents, improvements and alternatives falling within the spirit and principles of the present invention will be apparent to those skilled in the art within the scope of the present invention.
Claims (6)
1. The radar signal sorting method under the high pulse density environment is characterized in that the radar signal sorting method under the high pulse density environment divides the received radar pulse description words according to the pulse arrival angles to obtain a plurality of groups of pulse sequences from different directions; inputting the pulse frequency domain parameters and pulse width of each group of pulse sequences into a cascade structure self-organizing mapping neural network to obtain a primary sorting result; an improved sequence difference histogram algorithm is used for the primary sorting result to obtain a main sorting result; traversing the combination condition of the main sorting result, inputting pulse frequency domain parameters, pulse width and pulse arrival time into a long-short time memory network (LSTM) for repeated frequency identification, and combining to obtain a repeated frequency group variable radar;
the method for obtaining the multi-group pulse sequences from different directions specifically comprises the following steps of:
1) Detecting an object P which is not checked in a database, if the object P is not processed and is classified into a certain cluster or marked as noise, checking a neighborhood of the object P, if the number of the included objects is not less than the minimum allowable number of objects minPts, establishing a new cluster C, and adding all points in the new cluster C into a candidate set N;
2) Checking the neighborhood of all the objects q which are not processed in the candidate set N, and adding the objects into N if the objects at least contain minPts objects; if q does not fall into any cluster, adding q to C;
3) Repeating the step 2), and continuously checking the unprocessed object in the N, wherein the current candidate set N is empty;
4) Repeating steps 1) to 3) until all objects fall into a certain cluster or are marked as noise;
the cascade self-organizing mapping neural network primary sorting steps are as follows:
1) Initialization, determining threshold for judging neuron merging or splittingσ i Corresponding to the error of each parameter value obtained by the reconnaissance receiver; setting an initial value m of the number of output neurons of the self-organizing map neural network 0 The maximum number of cycles allowed is K,neurons undergo one-time merging or splitting into one round of cycles;
2) Training by using a traditional self-organizing map learning algorithm to enable the self-organizing map learning algorithm to reach an ordered map to obtain an initial clustering center;
3) Calculating the average distance in each classAnd inter-class distance D j ＝||m j -m j+1 | (j=1, 2,., c-1) and comparing with a set threshold R; if d j > R, then neuron j splits; if D j If R is less than R, merging the neurons j, determining whether two types are merged into one type or split into two types, and adjusting the scale of the self-organizing map neural network to obtain the number of new output neurons, namely m i Specific numerical values; if all output neurons are neither merged nor split, go to step 5); otherwise, go to step 4);
4) Judging whether the circulation turns are finished or not, if so, turning to the step 5); otherwise, turning to the step 2);
5) Calculating the value J of the clustering criterion function 1,m Obtaining various clustering center values;
6) Calculating J corresponding to the number of output neurons m+1 and m-1 1,m+1 And J 1,m-1 And sum J 1,m Comparing, taking max (J 1,m+1 ,J 1,m ,J 1,m-1 ) The corresponding neuron number is the final result, and various clustering center values are obtained;
the main sorting steps of the improved sequence difference histogram algorithm are as follows:
1) Inputting a pulse arrival time sequence to be sorted;
2) Performing PRI classification statistics of all possible pulse repetition intervals;
3) Setting a statistical threshold epsilon, and sequencing and de-duplicating PRI larger than the threshold;
4) Traversing PRI passing through threshold;
5) Traversing all pulse arrival times TOAs;
6) Calculating the allowable time range [ TOA+PRI-mu, TOA+PRI+mu ] of the PRI under the current arrival time according to the noise tolerance mu;
7) Judging whether pulses exist in the allowable time range, if so, continuing to execute the step 6) by TOA=TOA', otherwise, increasing the missing pulse number by a misscount++;
8) Judging whether the missing pulse number misscunt reaches a set maximum value misscunt_max, if so, setting the missing pulse number to zero, changing TOA into TOA and executing step 5), otherwise, continuing to execute step 6) by TOA=TOA';
9) Sub-harmonic inspection is carried out on all the extracted TOA sequences, and the merging of inspection rules is met, so that a final main sorting result is obtained;
the steps of the long and short time memory network for carrying out the identification of the repetition frequency group change are as follows:
1) Carrying out combined sorting on a radar Pulse Description Word (PDW) sequence according to the pulse arrival time obtained by main sorting;
2) Determining parameters of a segmented random feature sampling method: segment length k, number of segments d and interval g between segments;
3) Randomly selecting k sequence data from the PDW sequence as first segment characteristic data;
4) Judging whether the data is sampled; if yes, executing the step 4), otherwise, executing the step 2) by adding the interval g between the segments to the sampling start position;
5) Inputting the sampling result into a long-short-time memory network after training is completed, and obtaining a judgment result of the repetition frequency group variation;
6) If the repetition frequency group transformation rule is met, the PDW sequences are combined to form a repetition frequency group transformation radar pulse sequence, otherwise, the PDW sequences are not combined.
2. A computer device comprising a memory and a processor, the memory storing a computer program that, when executed by the processor, causes the processor to perform the radar signal sorting method in a high pulse density environment of claim 1.
3. A computer readable storage medium storing a computer program which, when executed by a processor, causes the processor to perform the method of radar signal sorting in a high pulse density environment of claim 1.
4. A radar signal sorting system in a high pulse density environment for operating the radar signal sorting method in a high pulse density environment of claim 1, the radar signal sorting system in a high pulse density environment comprising:
the high-density pulse sparse module is used for carrying out sparse on the high-density pulse environment to obtain a multi-channel pulse sequence for parallel processing;
the cascade self-organizing mapping neural network primary sorting module is used for carrying out primary clustering sorting on radar pulses to obtain a clustering result of primary sorting of current radar signals;
the improved sequence difference histogram main sorting module is used for carrying out main sorting on the sparse radar pulse sequence to obtain a main sorting result of the repeated frequency fixed, repeated frequency sliding and repeated frequency dithering radar signal;
and the long-and-short-term memory network repetition frequency group transformation identification module is used for identifying and combining radar pulse sequences conforming to a repetition frequency group transformation rule in the repetition frequency fixed radar, and finally obtaining a repetition frequency group transformation radar signal sorting result.
5. A radar signal sorting system, characterized in that it is equipped with a radar signal sorting system according to claim 4 in a high pulse density environment.
6. A radar, wherein the radar is equipped with a radar signal sorting system in a high pulse density environment according to claim 4.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010731268.8A CN112036074B (en) | 2020-07-27 | 2020-07-27 | Radar signal sorting method and system in high pulse density environment |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202010731268.8A CN112036074B (en) | 2020-07-27 | 2020-07-27 | Radar signal sorting method and system in high pulse density environment |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN112036074A CN112036074A (en) | 2020-12-04 |
| CN112036074B true CN112036074B (en) | 2024-04-02 |
Family
ID=73583230
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202010731268.8A Active CN112036074B (en) | 2020-07-27 | 2020-07-27 | Radar signal sorting method and system in high pulse density environment |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN112036074B (en) |
Families Citing this family (22)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN112698274A (en) * | 2020-12-05 | 2021-04-23 | 扬州宇安电子科技有限公司 | Radar signal sorting and pulse sequence extraction system based on hierarchical PRI transformation |
| CN112799053B (en) * | 2020-12-09 | 2023-08-15 | 中国人民解放军战略支援部队信息工程大学 | Real-time pairing and sorting method for monopulse signals |
| CN112526456B (en) * | 2020-12-28 | 2022-07-15 | 扬州健行电子科技有限公司 | Radar signal identification method |
| CN112904287B (en) * | 2021-01-14 | 2021-10-22 | 无锡国芯微电子***有限公司 | Tracking radiation source parameter estimation method based on passive radar finder and application thereof |
| CN112904288B (en) * | 2021-01-14 | 2023-08-29 | 无锡国芯微电子***有限公司 | Pulse signal combining system and pulse signal combining method |
| CN112949383B (en) * | 2021-01-22 | 2023-04-18 | 中国人民解放军63892部队 | Waveform agility radar radiation source identification method based on HyDeep-Att network |
| CN112986925B (en) * | 2021-02-01 | 2024-03-12 | 东南大学 | Radar pulse sequence sorting method based on image features |
| CN113033629B (en) * | 2021-03-09 | 2022-08-05 | 中南大学 | Radar signal sorting method and device based on improved cuckoo algorithm |
| CN112986928B (en) * | 2021-03-11 | 2022-06-17 | 哈尔滨工程大学 | Signal sorting multi-source fusion processing method in complex electromagnetic environment |
| CN113128584B (en) * | 2021-04-15 | 2022-11-11 | 北京理工大学 | Mode-level unsupervised sorting method of multifunctional radar pulse sequence |
| CN113127807B (en) * | 2021-04-19 | 2023-02-24 | 深圳华创电科技术有限公司 | Mode5 leading pulse jitter value calculation method based on constrained least square algorithm |
| CN113259017B (en) * | 2021-05-13 | 2022-02-08 | 电子科技大学 | Intra-frame signal sorting method based on time hopping relation |
| CN113253217B (en) * | 2021-05-27 | 2021-10-01 | 南京雷电信息技术有限公司 | Method for constructing complex electromagnetic environment based on pulse stream |
| CN113759323B (en) * | 2021-07-19 | 2024-02-06 | 西安电子科技大学 | Signal sorting method and device based on improved K-Means joint convolution self-encoder |
| CN113721219B (en) * | 2021-10-08 | 2023-07-04 | 中国电子科技集团公司第三十八研究所 | Radar signal sorting method and system based on multi-parameter clustering |
| CN113962261B (en) * | 2021-10-21 | 2024-05-14 | 中国人民解放军空军航空大学 | Deep network model construction method for radar signal sorting |
| CN113985356B (en) * | 2021-10-29 | 2022-12-27 | 扬州宇安电子科技有限公司 | Method for realizing radar signal pulse description word sorting function |
| CN114624645B (en) * | 2022-03-10 | 2022-09-30 | 扬州宇安电子科技有限公司 | Miniature rotor unmanned aerial vehicle radar reconnaissance system based on micro antenna array |
| CN114924234B (en) * | 2022-07-21 | 2022-10-21 | 中国人民解放军国防科技大学 | Radar radiation source target signal detection method based on regional contrast |
| CN115128569B (en) * | 2022-08-30 | 2022-11-25 | 中国电子科技集团公司第十五研究所 | Radar reconnaissance full-pulse data visual sorting method, server and storage medium |
| CN117554919B (en) * | 2024-01-11 | 2024-03-29 | 成都金支点科技有限公司 | Radar signal sorting and searching method based on bidirectional LSTM network |
| CN118131168A (en) * | 2024-05-08 | 2024-06-04 | 成都玖锦科技有限公司 | Complex environment signal sorting system and method |
Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110764063A (en) * | 2019-10-15 | 2020-02-07 | 哈尔滨工程大学 | Radar signal sorting method based on combination of SDIF and PRI transformation method |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| KR101007662B1 (en) * | 2009-05-08 | 2011-01-13 | 국방과학연구소 | Radar Signals Clustering Method using Frequency Modulation Characteristics and Combination Characteristics of Signals, and System for Receiving and Processing Radar Signals using the same |
-
2020
- 2020-07-27 CN CN202010731268.8A patent/CN112036074B/en active Active
Patent Citations (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110764063A (en) * | 2019-10-15 | 2020-02-07 | 哈尔滨工程大学 | Radar signal sorting method based on combination of SDIF and PRI transformation method |
Non-Patent Citations (6)
| Title |
|---|
| 一种密集信号环境下雷达脉冲分选方法;魏娟, 杨万麟, 植强, 王松煜;电子科技大学学报;20050625(第03期);全文 * |
| 基于融合算法的多参数雷达信号分选;陈恒;张友益;王玉梅;;舰船电子工程;20180920(第09期);全文 * |
| 岳志勇 ; 肖军鹏 ; 王鹏 ; .重频抖动脉冲信号分选研究.遥测遥控.2007,(S1),全文. * |
| 重频抖动脉冲信号分选研究;岳志勇;肖军鹏;王鹏;;遥测遥控;20071115(第S1期);全文 * |
| 陈恒 ; 张友益 ; 王玉梅 ; .基于融合算法的多参数雷达信号分选.舰船电子工程.2018,(09),全文. * |
| 魏娟,杨万麟,植强,王松煜.一种密集信号环境下雷达脉冲分选方法.电子科技大学学报.2005,(03),全文. * |
Also Published As
| Publication number | Publication date |
|---|---|
| CN112036074A (en) | 2020-12-04 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN112036074B (en) | Radar signal sorting method and system in high pulse density environment | |
| Bortolato et al. | Bump hunting in latent space | |
| Heredia et al. | Density-based clustering methods for unsupervised separation of partial discharge sources | |
| CN112036239B (en) | Radar signal working mode identification method and system based on deep learning network | |
| CN113721219B (en) | Radar signal sorting method and system based on multi-parameter clustering | |
| Cain et al. | Convolutional neural networks for radar emitter classification | |
| CN112986925B (en) | Radar pulse sequence sorting method based on image features | |
| KR101313051B1 (en) | Method and device for recognizing pulse repetition interval modulation type of a radar signal | |
| CN109766926B (en) | Radar radiation source signal intra-pulse characteristic comprehensive evaluation method and system | |
| CN111693945B (en) | Inter-pulse modulation method, inter-pulse modulation system, computer device, and storage medium | |
| CN111487592B (en) | Radar signal sorting method based on GPU and CPU platform drift mean clustering | |
| Chunjie et al. | Synthetic algorithm for deinterleaving radar signals in a complex environment | |
| Nuhoglu et al. | Image segmentation for radar signal deinterleaving using deep learning | |
| Xie et al. | A novel method for deinterleaving radar signals: First‐order difference curve based on sorted TOA difference sequence | |
| Shapero et al. | Identifying agile waveforms with neural networks | |
| CN103076595B (en) | Abnormal type identifying method for multivariate discrete radar emitter signal | |
| CN112904292A (en) | Broadband radar detection threshold determination and target detection method based on importance sampling | |
| Zhang et al. | PRI modulation recognition and sequence search under small sample prerequisite | |
| CN115980689A (en) | Point cloud detection-based radiation source signal sorting method, device, equipment and medium | |
| Dai et al. | Clustering of DOA data in radar pulse based on SOFM and CDbw | |
| CN115629365A (en) | Method and device for sorting few sample signals in high-density background signals | |
| CN108197646A (en) | A kind of target classification identification method for distributed optical fiber sensing system | |
| Perďoch et al. | An improved radar clutter suppression by simple neural network | |
| Han et al. | Radar clutter classification using expectation-maximization method | |
| Jakubowski et al. | Recognition of moving tracked and wheeled vehicles based on sound analysis and machine learning algorithms |
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 |