CN116626631A - Automatic radar model identification method and system combining intra-pulse and inter-pulse characteristics - Google Patents
Automatic radar model identification method and system combining intra-pulse and inter-pulse characteristics Download PDFInfo
- Publication number
- CN116626631A CN116626631A CN202310127615.XA CN202310127615A CN116626631A CN 116626631 A CN116626631 A CN 116626631A CN 202310127615 A CN202310127615 A CN 202310127615A CN 116626631 A CN116626631 A CN 116626631A
- Authority
- CN
- China
- Prior art keywords
- pulse
- radar
- modulation
- intra
- signals
- 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.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 61
- 238000013527 convolutional neural network Methods 0.000 claims abstract description 31
- 238000004458 analytical method Methods 0.000 claims abstract description 13
- 238000013145 classification model Methods 0.000 claims description 21
- 238000012549 training Methods 0.000 claims description 19
- 238000006243 chemical reaction Methods 0.000 claims description 8
- 238000011176 pooling Methods 0.000 claims description 7
- 230000004913 activation Effects 0.000 claims description 6
- 230000006870 function Effects 0.000 claims description 6
- 230000005855 radiation Effects 0.000 claims description 6
- 238000012360 testing method Methods 0.000 claims description 6
- 239000002131 composite material Substances 0.000 claims description 4
- 238000010276 construction Methods 0.000 claims description 4
- 238000012216 screening Methods 0.000 claims description 3
- 238000004590 computer program Methods 0.000 claims description 2
- 238000003062 neural network model Methods 0.000 claims description 2
- 108010076504 Protein Sorting Signals Proteins 0.000 claims 1
- 238000013135 deep learning Methods 0.000 abstract description 5
- 238000004364 calculation method Methods 0.000 description 11
- 238000003491 array Methods 0.000 description 2
- 150000001875 compounds Chemical class 0.000 description 2
- 238000013461 design Methods 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000009471 action Effects 0.000 description 1
- 238000013528 artificial neural network Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000001514 detection method Methods 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 230000004927 fusion Effects 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 210000002569 neuron Anatomy 0.000 description 1
- 230000003287 optical effect Effects 0.000 description 1
- 230000008092 positive effect Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 239000004065 semiconductor Substances 0.000 description 1
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
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
- G01S7/417—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section involving the use of neural networks
-
- 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
- G01S7/021—Auxiliary means for detecting or identifying radar signals or the like, e.g. radar jamming signals
-
- 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
- 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
- G01S7/40—Means for monitoring or calibrating
-
- 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
- G01S7/40—Means for monitoring or calibrating
- G01S7/4052—Means for monitoring or calibrating by simulation of echoes
-
- 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
- G01S7/40—Means for monitoring or calibrating
- G01S7/4052—Means for monitoring or calibrating by simulation of echoes
- G01S7/406—Means for monitoring or calibrating by simulation of echoes using internally generated reference signals, e.g. via delay line, via RF or IF signal injection or via integrated reference reflector or transponder
-
- 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
- G01S7/40—Means for monitoring or calibrating
- G01S7/4052—Means for monitoring or calibrating by simulation of echoes
- G01S7/4082—Means for monitoring or calibrating by simulation of echoes using externally generated reference signals, e.g. via remote reflector or transponder
-
- 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
- G01S7/40—Means for monitoring or calibrating
- G01S7/4052—Means for monitoring or calibrating by simulation of echoes
- G01S7/4082—Means for monitoring or calibrating by simulation of echoes using externally generated reference signals, e.g. via remote reflector or transponder
- G01S7/4091—Means for monitoring or calibrating by simulation of echoes using externally generated reference signals, e.g. via remote reflector or transponder during normal radar operation
-
- 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
- G01S7/41—Details of systems according to groups G01S13/00, G01S15/00, G01S17/00 of systems according to group G01S13/00 using analysis of echo signal for target characterisation; Target signature; Target cross-section
Landscapes
- Engineering & Computer Science (AREA)
- Radar, Positioning & Navigation (AREA)
- Remote Sensing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Physics & Mathematics (AREA)
- General Physics & Mathematics (AREA)
- Artificial Intelligence (AREA)
- Evolutionary Computation (AREA)
- Radar Systems Or Details Thereof (AREA)
Abstract
The invention belongs to the technical field of radar signal recognition, and discloses a radar model automatic recognition method and system combining intra-pulse and inter-pulse characteristics, wherein a characteristic parameter matching method, a deep learning convolutional neural network and a three-level classification method of intra-pulse parameter analysis are adopted; the first stage is based on tolerance range classification of characteristic parameter matching method, the second stage is image classification of convolutional neural network, and the convolutional neural network is utilized to realize secondary classification of radar signal modulation mode based on image recognition; the third stage is the classification of the pulse modulation parameters of the conventional pulse (no modulation in pulse), frequency coding, complex modulation, nonlinear frequency modulation, linear frequency modulation signals and phase coding signals, and the third stage of the radar signals is realized by using a time-frequency domain multi-parameter joint classification method. According to the invention, the intra-pulse modulation characteristic of the radar signal is obtained by analyzing the intra-pulse parameters of the radar; by combining the radar signal parameters between the pulses and in the pulses, the radar signal identification accuracy is improved.
Description
Technical Field
The invention belongs to the technical field of radar signal identification, and particularly relates to a radar model automatic identification method and system combining intra-pulse and inter-pulse characteristics.
Background
Radar radiation source identification (Emitter Identification, EID) is a key core of electronic intelligence reconnaissance, and battlefield situation information and tactical decision action information are provided for combat commanders by analyzing and identifying intercepted enemy radar signals. With the rapid development of electronic science and technology, the traditional radar signal identification means based on three-dimensional characteristics of pulse descriptors (Pulse Description Word, PDW) including Pulse Width (PW), carrier frequency (RF) and repetition Period (PRI) have been difficult to cope with the complex battlefield electromagnetic environment nowadays. Meanwhile, the problems that radar signal parameters of different types are aliased, the limit is fuzzy, pulse splitting is difficult to avoid and the like in a complex electromagnetic environment exist in the prior art. Therefore, there is a need to design an automatic radar signal recognition method combining intra-pulse and inter-pulse features.
Through the above analysis, the problems and defects existing in the prior art are as follows:
the traditional radar signal reconnaissance means based on pulse descriptors including pulse width, carrier frequency and repetition period three-dimensional characteristics is difficult to deal with complex battlefield electromagnetic environments nowadays, and the existing radar signal reconnaissance equipment can only perform template matching recognition of signals according to the traditional three-dimensional characteristics, so that when two or more radars are subjected to parameter aliasing and boundary blurring in the three-dimensional characteristics, the target judgment accuracy is influenced.
Disclosure of Invention
Aiming at the problems existing in the prior art, the invention provides a radar model automatic identification method and system combining intra-pulse and inter-pulse characteristics, and aims to solve the problems that the identification method for traditional template matching of radar signals in a complex electromagnetic environment is low in accuracy, pulse splitting is difficult to avoid and the like in the prior art.
The invention is realized in such a way that the radar model automatic identification method combining intra-pulse and inter-pulse features comprises the following steps:
the method comprises the steps of adopting a traditional characteristic parameter template matching method, a deep convolutional neural network and a gradual recognition method of intra-pulse parameter analysis, wherein the first stage is a radar platform which is based on EDW (Emitter Discreption Word) parameters and has different intra-pulse modulation modes, the recognition result is directly pushed when the recognition result is only a 1-type radar signal, the second stage is carried out when the recognition result is a radar signal with more than 2 types, the second stage utilizes the convolutional neural network to realize the intra-pulse parameter analysis of radar signal modulation modes (comprising linear frequency modulation, phase coding, conventional pulse (no modulation in pulse), frequency coding, complex modulation, nonlinear frequency modulation and the like) based on a time-frequency image, aims to distinguish radar platforms which are similar in three characteristic parameters of carrier frequency, pulse width and repetition period in the first stage, enter a third stage recognition when the classification result is still a radar signal with more than 2 types, the third stage carries out intra-pulse parameter analysis of radar intermediate frequency signals for the linear signal and the phase coding signal classified in the second stage recognition when the recognition result is still a radar with multiple types, carries out the intra-pulse parameter analysis of the radar intermediate frequency signal, carries out the calculation of the radar signal with tolerance calculation, and carries out the calculation of the bandwidth and carries out the radar parameter library, and the final classification method is carried out the method by carrying out the radar parameter matching.
Further, the radar model automatic identification method combining intra-pulse and inter-pulse features comprises the following steps:
step one, constructing a radar signal classification model;
step two, classifying and training radar signals and identifying the radar signals;
and thirdly, updating the radar signal classification model.
Further, the constructing of the radar signal classification model in the first step includes;
the classification method is divided into 3 stages, the core idea is to gradually narrow the classification recognition range, the 1 st stage is classification based on characteristic parameter matching, the 2 nd stage is classification based on the intra-pulse modulation mode of the convolutional neural network radar signal based on the time-frequency image under the condition of the 1 st stage classification result, (the classification based on the intra-pulse modulation mode of the radar signal is carried out under the classification model), the classification specifically comprises a linear frequency modulation signal, a phase coding signal, a conventional pulse (no-pulse modulation signal, a frequency coding signal, a nonlinear frequency modulation signal and a composite modulation signal), and the 3 rd stage is the recognition based on the intra-pulse parameter analysis under the condition of the 1 st stage classification and the 2 nd stage classification result.
Further, the constructing the radar signal classification model in the step one includes:
the constructing of the part 1 radar parameter model database comprises the following steps:
(1) Storing clarified radar intermediate frequency signals and sorted parameters, wherein the parameters comprise radar time domain waveforms, intra-pulse parameters and inter-pulse parameters, the inter-pulse parameters comprise pulse repetition periods PRI, carrier frequencies RF and pulse widths PW after sorting, and the intra-pulse parameters comprise signal bandwidths (B), frequency modulation slopes (y) of chirp signals, symbol widths (mu) of phase coded signals and coding modes (z);
(2) Establishing a corresponding radar radiation source model database S (hereinafter referred to as database S) according to the detected and clarified radar parameter data, wherein fields in the database S comprise: in the first step, the radar signal carrier frequency (RF), pulse Width (PW), pulse repetition Period (PRI), time-frequency image (PIC) after single pulse time-frequency conversion, intra-pulse modulation type (Modu), signal bandwidth (B), chirp rate (y) of the chirp signal, symbol width (μ) of the phase-coded signal, coding mode (z), and radar model (Target) are obtained in step (1).
Part 2 is a convolutional neural network model construction based on time-frequency images. And forming a convolutional neural network model of the radar signal intra-pulse modulation mode.
(1) After passing through time-frequency conversion, the radar time domain signal obtained in step (1) of part 1 is screened out signals with the same modulation mode and different signal to noise ratios, and a plurality of time-frequency image sets lambda are obtained 1 ；
(2) The time-frequency image set lambda obtained in the step (1) of the 2 nd part is processed 1 Respectively establishing radar signal intra-pulse modulation modesClassification data set lambda 2 And test set lambda 3 ；
(3) Constructing convolutional neural network, and integrating time-frequency image data set lambda 2 Inputting convolutional neural network for training to form one-time classified neural network model, and using test set lambda 3 Verifying the accuracy of the model;
in the step (1), a Short Time Fourier Transform (STFT) isochronous frequency transform method is used to perform time-frequency transform on the radar intermediate frequency signal, so as to obtain a plurality of time-frequency images.
The convolutional neural network structure in the step (3) of the 2 nd part comprises a network structure adopting N convolutional layers, M pooling layers and K full-connection layers, wherein N, M and K are integers which are more than or equal to 2; the activation function of N convolution layers is a Relu activation function, the softmax is adopted to carry out multi-objective classification, the convolution kernel size is m x m, wherein m is an odd integer which is more than 3 and less than 31, and the output is the corresponding radar signal intra-pulse modulation type or the direct classification identification of a special intra-pulse modulation mode.
The 3 rd part is to analyze the pulse modulation parameters of signals such as linear frequency modulation, phase coding, normal pulse (no modulation in pulse), complex modulation and the like and record the parameters into a database S, the linear frequency modulation signals obtain a frequency modulation slope y through calculation, the phase coding signals obtain a code element width mu and a coding mode z through calculation, and the other modulation modes such as normal pulse (no modulation in pulse), complex modulation and the like obtain a signal bandwidth B through calculation.
Further, the radar signal classification training and recognition in the second step includes:
(1) When a radar signal is newly detected, the sorted radar signal comprises three parameters including Carrier Frequency (CF), repetition Period (PRI) and pulse width, the three parameters of the signal and three template parameter data in a database S are subjected to comparison inquiry with tolerance, and the primary identification of a target is performed within a set tolerance range;
(2) Under the primary identification condition, when the template data in the database S is matched with and only one target result exists, the identification result is directly pushed to the user;
(3) Under the first-stage recognition condition, when two or more target results are matched with the template data in the database S, a second-stage signal recognition flow is carried out.
(4) In the signals with multiple radar signal boundary ambiguity in the last step, the pulse signal r to be identified x Performing time-frequency conversion, forming a time-frequency image, inputting a2 nd part model for secondary identification, and outputting a classification result of a modulation mode;
(5) The output modulation mode identification result is a signal of a conventional pulse (no modulation in pulse), frequency coding, nonlinear frequency modulation, compound modulation and other modulation modes, a3 rd part model is input, a signal bandwidth calculation result under the same modulation mode is output, template data is established by a radar radiation source signal model database S, database statement comparison query with tolerance is carried out, and the identification result is pushed;
(6) Analyzing the pulse modulation parameters of the output modulation mode identification result as a linear frequency modulation signal and a phase coding signal, calculating the frequency modulation slope of the signal to form a data parameter y, calculating the phase coding signal to obtain the coding mode z and the code element width mu of the signal, establishing template data by using a radar radiation source signal model database S, comparing and inquiring database sentences with tolerance, and pushing the identification result;
further, the updating of the radar signal classification model in the third step includes:
(1) In the second step (1), when the matching of the sorted signals and the template data in the database S has no result, the template matching of the radar-free signals is prompted to the user, the user can update parameters of the database S, and unknown signals can be further screened in the characteristic parameter matching.
(2) Building a label for the radar signal with the new unknown modulation mode, importing a training data set for retraining, and updating a convolutional neural network model;
(3) And (5) carrying out intra-pulse parameter and inter-pulse parameter analysis on the newly detected radar signals, and updating the database S.
Another object of the present invention is to provide a radar model automatic recognition system for combining intra-pulse and inter-pulse features using the method for combining intra-pulse and inter-pulse features, the radar model automatic recognition system for combining intra-pulse and inter-pulse features comprising:
the model construction module is used for constructing a radar signal classification model;
the training and identifying module is used for carrying out classification training and identification on radar signals;
and the model updating module is used for updating the radar signal classification model.
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:
adopting a characteristic parameter matching method, a deep learning convolutional neural network and an intra-pulse parameter estimation three-level classification method; the first stage is three inter-pulse classifications based on carrier frequency, repetition period and pulse width of a traditional characteristic parameter matching method, the second stage is image classification of a convolutional neural network, and the convolutional neural network is utilized to realize secondary classification of a radar signal modulation mode based on image recognition; and the third stage is classification of radar intra-pulse parameters after estimation, and three-time classification identification of radar signals is realized by using a time-frequency domain multi-parameter joint classification method.
Another object of the present invention is to provide an information data processing terminal for implementing the radar model automatic recognition system combining intra-pulse and inter-pulse features.
In combination with the technical scheme and the technical problems to be solved, the technical scheme to be protected has the following advantages and positive effects:
the radar signal automatic identification method combining the intra-pulse and inter-pulse features combines the multi-stage classification modes, utilizes multi-parameter combined judgment, introduces a deep learning framework, and realizes rapid identification of radar signals while improving accuracy.
According to the invention, the intra-pulse modulation characteristic of the radar signal is obtained by analyzing the intra-pulse parameters of the radar, and the accuracy of radar signal identification is improved by combining the inter-pulse and intra-pulse radar signal parameters.
According to the method, the radar model primary identification based on the characteristic parameters is realized by utilizing a traditional characteristic parameter matching method, and a convolutional neural network is used under the condition of fuzzy identification, so that the secondary classification of the radar signal modulation mode based on the time-frequency image is realized; under the condition of fuzzy secondary classification and identification, the intra-pulse parameter analysis method is utilized to realize the three-time identification of signals such as intra-pulse linear frequency modulation, phase coding, regular pulse (no modulation in the pulse) and the like of the radar, and the accuracy rate of radar signal identification can be effectively improved.
In the engineering application field of radar signal identification, only the traditional characteristic parameter matching method is used, and the characteristic parameter matching, convolutional neural network time-frequency image classification and intra-pulse parameter feature combination method used in the invention are used for identification, so that the dimensional information of signal identification is increased, and the reliability and accuracy of the signal identification method are improved.
The invention is very different from the patent number CN 113156391A, the application of the patent is fundamentally different from the application of the invention, the patent is applied to radar signal sorting, and the invention is applied to accurate identification of radar types.
Drawings
In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings that are needed in the embodiments of the present invention will be briefly described below, and it is obvious that the drawings described below are only some embodiments of the present invention, and other drawings may be obtained according to these drawings without inventive effort for a person skilled in the art.
FIG. 1 is a flowchart of a method for automatically identifying radar model of combining intra-pulse and inter-pulse features according to an embodiment of the present invention;
FIG. 2 is a block diagram of a radar model automatic identification system combining intra-pulse and inter-pulse features according to an embodiment of the present invention;
FIG. 3 is a specific network architecture diagram provided by an embodiment of the present invention;
in the figure: 1. a model building module; 2. training an identification module; 3. and a model updating module.
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 model automatic identification method and a radar model automatic identification system combining intra-pulse and inter-pulse characteristics, and the invention is described in detail below with reference to the accompanying drawings.
In order to fully understand how the invention may be embodied by those skilled in the art, this section is an illustrative embodiment in which the claims are presented for purposes of illustration.
The radar model automatic identification method combining intra-pulse and inter-pulse features provided by the embodiment of the invention comprises the following steps: adopting a characteristic parameter matching method, a deep learning convolutional neural network and a three-level classification method of intra-pulse parameter estimation fusion; the method comprises the steps of classifying a tolerance range based on a characteristic parameter matching method at a first stage, classifying images of a convolutional neural network at a second stage, and realizing secondary classification of a radar signal modulation mode based on image identification by using the convolutional neural network; the third level is the bandwidth parameter classification of signals such as normal pulse (no modulation in pulse), complex modulation and the like aiming at the intra-pulse modulation parameter classification of the linear frequency modulation signals and the phase coding signals under different modulation modes, and the third level classification of radar signals is realized by utilizing a time-frequency domain multi-parameter joint classification method. According to the radar model automatic identification method, multistage classification is combined, multi-parameter combined judgment is utilized, a deep learning framework is introduced, and the radar signal is quickly identified while the accuracy is improved. According to the invention, the intra-pulse modulation characteristic of the radar signal is obtained by analyzing the intra-pulse parameters of the radar; by combining the radar signal parameters between the pulses and in the pulses, the radar signal identification accuracy is improved.
As shown in fig. 1, the method for automatically identifying radar model of combined intra-pulse and inter-pulse features provided by the embodiment of the invention comprises the following steps:
s101, constructing a radar signal classification model;
SA1, detecting and storing radar signals and parameters, namely, equipment for collecting radar time domain signals, and storing the sorted parameters, wherein the parameters comprise pulse repetition period PRI, carrier frequency RF and pulse width PW after sorting, and the pulse parameters comprise signal bandwidth (B), frequency modulation slope (y) of a linear frequency modulation signal, code element width (mu) of a phase coding signal, coding mode (z) and the like;
SA2, performing time-frequency conversion on the original radar signal detected in the step SA1, and manually screening signals with similar modulation modes and different signal to noise ratios according to different time-frequency image characteristics to obtain a plurality of time-frequency image sets lambda 1 ；
SA3, forming a radar radiation source parameter model database S, wherein fields in the database comprise radar signal carrier frequency (RF), pulse Width (PW), pulse repetition Period (PRI), pulse time-frequency image (PIC), pulse internal modulation type (Modu), signal bandwidth (B) (regular pulse (pulse internal modulation), frequency coding, composite modulation signal and the like), radar model (Target), frequency modulation slope (y) (chirp signal only), symbol width (mu) and coding mode (z) (phase coding signal only).
SA4, the time-frequency image set lambda obtained in the step SA2 is collected 1 Establishing a primary classification data set lambda of radar signal intra-pulse modulation 2 And test set lambda 3 ；
SA5, constructing a convolutional neural network with modulation mode classification:
SA5-1 specifically comprises: adopting a network structure of N convolution layers, M pooling layers and K full connection layers, wherein N, M, K is an integer greater than or equal to 2; the activation functions of the N convolution layers are the Relu activation functions, the softmax is adopted for classification, the output is the corresponding modulation type, the convolution kernel size is m x m, wherein m is an odd integer which is more than 3 and less than 31, as shown in fig. 3, the specific network structure is provided with n=2 and m=3 according to the invention, wherein N is the number of target types to be classified, and the convolution layer representation method in fig. 3 is "convolution layer (convolution kernel) @ (convolution kernel size)"; the representation method of the pooling layer in fig. 3 is "pooling layer @ (pooling window size)"; the fully connected layer is shown in the figure as "fully connected layer @ (number of neurons)". The convolution kernel sizes corresponding to the convolution layers are respectively as follows: m is 3, 16 m is m, m is an odd integer greater than 3 and less than 31; to reduce the number of training parameters after the convolutional layer outputs, the pooling layer is utilized to reduce the training parameters.
SA5-2, classifying the time-frequency image obtained in the step SA3, and establishing a classified data set lambda of the radar signal modulation mode 2 The radar signal classification data set is divided into six types according to modulation modes in radar signal pulses: the method comprises the steps of inputting a linear frequency modulation signal, a non-linear frequency modulation signal, a phase coding signal, a conventional pulse (no modulation in pulse), a frequency coding signal and a composite modulation signal into a neural network to finally form a classification model of a convolutional neural network modulation mode based on a time-frequency image, and using a test set lambda 3 Verifying the accuracy of the model;
s102, performing radar signal classification training and recognition;
SB1, when the radar signal is newly detected, comparing and inquiring the template data in the database S in a set tolerance range through calculation, and carrying out primary judgment on the target;
SB2, after SB1 calculation, when the template data in the database S is matched with and has only one target result within the tolerance range of the new detection signal, directly pushing the identification result to the user;
SB3, after SB1 calculation, when the new detected signal is matched with the template data in the database S to obtain two or more target results, proving that a plurality of radar signal boundaries are fuzzy in the sorted signal, and carrying out the next signal identification flow;
SB4, pulse signal r of ambiguous signal in signal where aliasing of radar signal occurs in SB3 x Performing time-frequency conversion to form a time-frequency image, inputting the time-frequency image into the model of the 2 nd part, and outputting a classification result of the modulation mode;
SB5-1, the output modulation mode identification result is the signal of the modulation modes such as the conventional pulse (no modulation in the pulse), the frequency coding, the nonlinear frequency modulation, the compound modulation and the like, the 3 rd part model is input to obtain the parameter of the signal bandwidth, and the identification result is searched and pushed by comparing with the database S with the tolerance;
SB5-2, analyzing the pulse modulation parameters of the output modulation mode identification result which is the linear frequency modulation signal and the phase coding signal, inputting the 3 rd part model to obtain the pulse modulation parameters, comparing and inquiring the pulse modulation parameters with a database S with tolerance, and pushing the identification result;
SB5-3, in the step (6), classifying the step (4) to obtain a radar monopulse time domain signal with a modulation mode of a linear frequency modulation signal, calculating a frequency modulation slope of the signal to form a data parameter y, and calculating a coding mode z and a code element width mu of the signal with the modulation mode of a phase coding signal obtained by classification.
S103, updating a radar signal classification model:
and SC1, after calculation by SB1, if no result exists in the tolerance range of the template data matching with the template data in the database S, prompting the template matching without radar signals to a user, updating the template database and the parameter feature library by the user, and further screening unknown signals in the feature parameter matching.
SC2, building a label for the newly detected radar signal with an unknown modulation mode, importing a training data set for retraining, and updating a convolutional neural network model;
and SC3, performing intra-pulse parameter and inter-pulse parameter analysis on the newly detected radar signal, and updating the database S.
As shown in fig. 2, the radar model automatic identification system combining intra-pulse and inter-pulse features provided by the embodiment of the invention includes:
the model construction module 1 is used for constructing a radar signal classification model;
the training and identifying module 2 is used for carrying out classification training and identification on the radar signals;
and the model updating module 3 is used for updating the radar signal classification model.
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 (9)
1. A radar model automatic identification method combining intra-pulse and inter-pulse features is characterized by comprising the following steps of
The radar model automatic identification method combining intra-pulse and inter-pulse features comprises the following steps:
the final recognition result pushing of the radar signal is carried out by adopting a characteristic parameter template matching method, a deep convolutional neural network and a step-by-step recognition method of intra-pulse parameter analysis, wherein the step-by-step recognition method comprises the following steps: the first stage is to identify based on EDW (Emitter Discreption Word) parameter characteristic parameter matching method, and directly push identification result when identifying as 1 type radar radiation source; the second stage uses convolutional neural network to realize the classification of the radar signal modulation mode based on the image, and carries out the secondary classification on the result of the first stage identification; and thirdly, carrying out intra-pulse parameter analysis on the radar intermediate frequency signals on signals such as conventional pulse (no modulation in pulse), complex modulation, linear frequency modulation, phase coding and the like, and carrying out three-time classification on the result of the secondary identification.
2. The method for automatically identifying radar model of combined intra-pulse and inter-pulse features according to claim 1, wherein the method for automatically identifying radar model of combined intra-pulse and inter-pulse features specifically comprises the following steps:
step one, constructing a radar signal classification model;
step two, classifying and training radar signals and identifying the radar signals;
and thirdly, updating the radar signal classification model.
3. The method for automatically identifying radar models that combine intra-and inter-pulse features according to claim 2, wherein in step one, the classification model includes:
a model based on feature parameter matching;
convolutional neural network radar signal intra-pulse modulation mode classification model based on time-frequency image;
the pulse modulation parameters of the signals such as conventional pulse (no modulation in pulse), complex modulation, linear modulation, phase coding and the like are analyzed and identified.
4. The method for automatically identifying radar models combining intra-pulse and inter-pulse features according to claim 3, wherein the radar signal classification model comprises:
(1) Storing radar intermediate frequency signals and sorted parameters, wherein the radar intermediate frequency signals comprise radar time domain waveforms, intra-pulse parameters and inter-pulse parameters, the inter-pulse parameters comprise pulse repetition periods PRI, carrier frequencies RF and pulse widths PW after sorting, and the intra-pulse parameters comprise signal bandwidths (B), frequency modulation slopes (y) of chirp signals, symbol widths (mu) of phase coded signals, coding modes (z) and the like;
(2) Establishing a corresponding radar parameter database S according to the detected and interpreted radar parameter data, wherein fields in the database comprise: the radar signal carrier frequency (RF), pulse Width (PW), pulse repetition Period (PRI), time-frequency image (PIC) after single pulse is subjected to short-time Fourier transform, pulse modulation type (Modu), signal bandwidth (B), chirp signal frequency modulation slope (y), phase coding signal code element width (mu), coding mode (z) and radar platform model (Target) obtained in the step (1);
the identification of the modulation mode, namely under the condition of fuzzy matching of characteristic parameters of the sorting signals, based on the image identification of a convolutional neural network, forming an identification model of the radar signal intra-pulse modulation mode;
1) After passing through time-frequency conversion, the radar time domain signals obtained in the step (1) are screened out signals with the same modulation mode and different signal to noise ratios, and a plurality of time-frequency image sets lambda are obtained 1 The method comprises the steps of carrying out a first treatment on the surface of the Performing time-frequency conversion on the radar intermediate frequency signal by adopting short-time Fourier transform (STFT) and other methods to obtain a plurality of time-frequency images;
2) The time-frequency image set lambda obtained in the step 1) is processed 1 Respectively establishing a radar signal intra-pulse modulation mode classification data set lambda 2 And test set lambda 3 ；
3) Constructing convolutional neural network, and integrating time-frequency image data set lambda 2 Inputting convolutional neural network for training to form one-time classified neural network model, and using test set lambda 3 Verifying the accuracy of the model;
the convolutional neural network structure comprises N convolutional layers, M pooling layers and M full-connection layers, wherein N, M, K is an integer greater than or equal to 2; the activation functions of the N convolution layers are Relu activation functions, the softmax is adopted to carry out multi-objective classification, and the output is the corresponding radar signal intra-pulse modulation type or the direct classification identification of a special intra-pulse modulation mode.
5. The method for automatically identifying radar models by combining intra-pulse and inter-pulse features according to claim 4, wherein the intra-pulse modulation parameters of the signals such as normal pulse (no intra-pulse modulation), complex modulation, frequency coding, chirp, phase coding, etc. are analyzed and recorded in the database S, the signal bandwidth B is calculated for the signals such as normal pulse (no intra-pulse modulation), complex modulation, frequency coding, etc., the chirp rate y is calculated for the chirp signal, the symbol width μ is calculated for the phase-coded signal, and the coding scheme z is calculated.
6. The method for automatically identifying radar models by combining intra-pulse and inter-pulse features according to claim 2, wherein the training and identifying radar signal classification in the second step comprises:
(1) When the radar signal is newly detected, comparing and inquiring the sorted radar signal with template data in a database S with tolerance, and carrying out primary discrimination of the target within a set tolerance range;
(2) When the target result is matched with the template data in the database S within the tolerance range and only one target result exists, the identification result is directly pushed to the user;
(3) When two or more target results are matched with template data in the database S within the tolerance range, proving that a plurality of radar signal boundary blurring or sorting parameters exist in the sorted signals, and carrying out the next signal identification flow;
(4) Pulse signal r for primary identification and sorting x Performing time-frequency conversion to form a time-frequency image, inputting the time-frequency image into the model of the 2 nd part, and outputting a classification result of the modulation mode;
(5) And (3) classifying the radar monopulse time domain signals with modulation modes of the chirp signals obtained in the step (4), calculating frequency modulation slope to form a data parameter y, classifying the radar monopulse time domain signals with modulation modes of the chirp signals to obtain a coding mode z of the signals and a code element width mu of the signals obtained by classifying the signals, performing frequency coding and non-chirp on the conventional pulses obtained by classifying (pulse without modulation), calculating the composite modulation signals to obtain a signal bandwidth B, and comparing and inquiring through a database S based on tolerance to output identification results.
7. The method for automatically identifying radar model combining intra-pulse and inter-pulse features according to claim 2, wherein the updating of radar signal classification model in the third step comprises:
(1) In the second step (1), when no result of matching with the template data in the database S occurs within the tolerance range, prompting the user of template matching without radar signals, updating the template database and the parameter feature library by the user, and further screening unknown signals in feature parameter matching;
(2) Building a label for the radar signal with the new unknown modulation mode, importing a training data set for retraining, and updating a convolutional neural network model;
(3) And (5) carrying out intra-pulse parameter and inter-pulse parameter analysis on the newly detected radar signals, and updating the database S.
8. A radar model automatic identification system combining intra-pulse and inter-pulse features, characterized in that the radar model automatic identification system combining intra-pulse and inter-pulse features comprises:
the model construction module is used for constructing a radar signal classification model;
the training and identifying module is used for carrying out classification training and identification on radar signals;
and the model updating module is used for updating the radar signal classification model.
9. 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 method of automatic identification of radar models combining intra-and inter-pulse features as claimed in any one of claims 1 to 8.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310127615.XA CN116626631A (en) | 2023-02-15 | 2023-02-15 | Automatic radar model identification method and system combining intra-pulse and inter-pulse characteristics |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202310127615.XA CN116626631A (en) | 2023-02-15 | 2023-02-15 | Automatic radar model identification method and system combining intra-pulse and inter-pulse characteristics |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN116626631A true CN116626631A (en) | 2023-08-22 |
Family
ID=87637136
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202310127615.XA Pending CN116626631A (en) | 2023-02-15 | 2023-02-15 | Automatic radar model identification method and system combining intra-pulse and inter-pulse characteristics |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN116626631A (en) |
Cited By (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
-
2023
- 2023-02-15 CN CN202310127615.XA patent/CN116626631A/en active Pending
Cited By (2)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| 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 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN107220606B (en) | Radar radiation source signal identification method based on one-dimensional convolutional neural network | |
| Samadi et al. | Change detection in SAR images using deep belief network: a new training approach based on morphological images | |
| Babak et al. | Methods and models for information data analysis | |
| Wei et al. | Intra‐pulse modulation radar signal recognition based on CLDN network | |
| CN111444790B (en) | Pulse-level intelligent identification method for multifunctional radar working mode sequence | |
| Zhai et al. | Real‐time optical character recognition on field programmable gate array for automatic number plate recognition system | |
| CN112668498B (en) | Method, system, terminal and application for identifying individual intelligent increment of aerial radiation source | |
| Abdelwahab | Fast approach for efficient vehicle counting | |
| Hong-hai et al. | Radar emitter multi-label recognition based on residual network | |
| KR101313051B1 (en) | Method and device for recognizing pulse repetition interval modulation type of a radar signal | |
| CN116626631A (en) | Automatic radar model identification method and system combining intra-pulse and inter-pulse characteristics | |
| Liu | Pulse deinterleaving for multifunction radars with hierarchical deep neural networks | |
| CN114114166A (en) | Radar pulse de-interlacing method based on DTM algorithm | |
| Liu et al. | Radar signal recognition based on triplet convolutional neural network | |
| Nuhoglu et al. | Image segmentation for radar signal deinterleaving using deep learning | |
| Robertson et al. | Efficient approximations of the multi-sensor labelled multi-Bernoulli filter | |
| CN111413682A (en) | Synthetic extremely narrow pulse radar detection threshold calculation method based on sequence statistics | |
| 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 | |
| Zhu et al. | Covariance control joint integrated probabilistic data association filter for multi‐target tracking | |
| Guven et al. | Classifying LPI radar waveforms with time-frequency transformations using multi-stage CNN system | |
| CN116797796A (en) | Signal identification method based on time-frequency analysis and deep learning under DRFM intermittent sampling | |
| CN116758922A (en) | Voiceprint monitoring and diagnosing method for transformer | |
| Ma et al. | Radar working mode recognition based on hierarchical feature representation and clustering | |
| Zhuojun et al. | Stagger period estimation algorithm for multiple sets of radar pulses |
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 |