CN108509911A - Interference signal recognition methods based on convolutional neural networks - Google Patents
Interference signal recognition methods based on convolutional neural networks Download PDFInfo
- Publication number
- CN108509911A CN108509911A CN201810286070.6A CN201810286070A CN108509911A CN 108509911 A CN108509911 A CN 108509911A CN 201810286070 A CN201810286070 A CN 201810286070A CN 108509911 A CN108509911 A CN 108509911A
- Authority
- CN
- China
- Prior art keywords
- convolutional neural
- neural networks
- signal
- interference signal
- sample
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Granted
Links
Classifications
-
- 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
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F18/00—Pattern recognition
- G06F18/20—Analysing
- G06F18/21—Design or setup of recognition systems or techniques; Extraction of features in feature space; Blind source separation
- G06F18/214—Generating training patterns; Bootstrap methods, e.g. bagging or boosting
-
- 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
- G06N3/045—Combinations of networks
Landscapes
- Engineering & Computer Science (AREA)
- Theoretical Computer Science (AREA)
- Physics & Mathematics (AREA)
- General Engineering & Computer Science (AREA)
- Artificial Intelligence (AREA)
- General Physics & Mathematics (AREA)
- Data Mining & Analysis (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Biophysics (AREA)
- Biomedical Technology (AREA)
- Computing Systems (AREA)
- General Health & Medical Sciences (AREA)
- Computational Linguistics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Molecular Biology (AREA)
- Health & Medical Sciences (AREA)
- Signal Processing (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Evolutionary Biology (AREA)
- Noise Elimination (AREA)
- Image Analysis (AREA)
Abstract
The invention belongs to wireless communication technology fields, and in particular to a kind of interference signal recognition methods based on convolutional neural networks.The method of the present invention includes mainly structure convolutional neural networks；The interference signal received is pre-processed, as the input sample of convolutional neural networks；According to the classification of signal to be identified, sample of signal and its corresponding classification are configured to training set, utilize the convolutional neural networks of the training set training structure of structure；According to trained convolutional neural networks, each pretreated sample of signal is identified, the generic of unknown signaling is obtained.Beneficial effects of the present invention are, interference signal identification is carried out using convolutional neural networks, it overcomes traditional needs and artificially extracts interference signal feature the shortcomings that carrying out Classification and Identification, there is universality and flexibility to the identification of interference signal, and improve the accuracy of identification.
Description
Technical field
The invention belongs to wireless communication technology fields, and in particular to a kind of interference signal identification based on convolutional neural networks
Method.
Background technology
With the development of wireless communication technique, the electromagnetic environment that wireless communication system faces is increasingly complicated severe, both may
It can be by being communicated from one's own side, it is also possible to the interference sample deliberately discharged by enemy.In order to ensure the normal communication under environment,
For a variety of interference means occurred in communication countermeasure, corresponding disturbance restraining method also comes into being.However, each is interfered
Suppression technology is typically only capable to that for a certain certain types of jamming signal type best anti-jamming effectiveness could be obtained.Therefore, having must
Classification and Identification is carried out to the interference signal classification in complex electromagnetic environment, targetedly to take best interference protection measure
Technical foundation is provided.
Currently, existing interference signal recognition methods can be divided into two important stages.First stage carries out signal characteristic abstraction,
The feature of interference signal is generally extracted from the dimensions such as time domain, frequency domain, Higher Order Cumulants domain and time-frequency domain.Second stage is logical
It crosses grader to classify to interference signal, certain grader is trained according to the interference signal feature of extraction.Grader is usually adopted
With decision tree, support vector machines, neural network etc..It can refer to：Wu Hao, Hangzhoupro are dry based on Higher Order Cumulants and neural network
Disturb recognizer [J] military communication technologies, 2008 (1)；It is dry based on SVM in the directly-enlarging systems such as Yu Bo, Shao Gaoping, Sun Hongsheng
Disturb automatic classifying identification method [J], signal processing, 2010,26 (10):1539-1543.However, there are following two for this method
The problem of aspect：(1) feature extracted determines the upper limit of Classification and Identification performance, but more difficult in practice extracts to various interference
Signal has the feature of distinction；(2) feature extracted does not have universality and flexibility, can only identify that some are specific dry
It disturbs, if there is new jamming signal type, needs to re-start extraction feature.
Invention content
A kind of purpose of the present invention, aiming at the above problem, it is proposed that convolutional Neural net for interference signal identification
Network structure and the method for carrying out disturbance ecology using the convolutional neural networks, know interference signal to automatically extract feature
Not.
The technical solution adopted in the present invention is：
Interference signal recognition methods based on convolutional neural networks, which is characterized in that include the following steps：
A, convolutional neural networks are built：
Input layer is inputted using the sample format of N × 1 × 3, and N is positive integer；Then by two layers by convolution, activation letter
The network of number, Chi Hua, batch normalization composition；Then the network being made of Inception, Chi Hua, batch normalization by four layers,
Middle Inception is the operation by being concatenated together after a variety of nonidentity operations；Finally by a full articulamentum with
Softmax classifies；
B, the interference signal received is pre-processed, as the input sample of convolutional neural networks；
C, according to the classification of signal to be identified, sample of signal and its corresponding classification is configured to training set, utilize structure
Training set training step a in the convolutional neural networks that build；Specially：
By the pretreated sample of each known class and class label composing training collection, using the sample in training set as
The input of convolutional neural networks, output of the label vector as convolutional neural networks；
Using the training data of composition to train invention convolutional neural networks；
D, according to trained convolutional neural networks, each pretreated sample of signal is identified, unknown letter is obtained
Number generic, specially：
Each unknown sample is inputed into trained convolutional neural networks, obtains corresponding output vector；
According to greatest member manipulative indexing in the output categorization vector, the classification of signal is obtained.
Further, pretreated specific method is in the step b：
B1, down coversion is carried out to the interference signal received, complex base band is then obtained by low-pass filter and down-sampling
Signal；
B2, to complex baseband signal power normalized, FFT then is done to normalized complex baseband signal and obtains its width
The real part of complex baseband signal after normalization, imaginary part and amplitude spectrum are formed a sample of signal by degree spectrum.
Beneficial effects of the present invention are to carry out interference signal identification using convolutional neural networks, overcome traditional needs
Artificial extraction interference signal feature the shortcomings that carrying out Classification and Identification, has universality and flexibility to the identification of interference signal,
And improve the accuracy of identification.
Description of the drawings
Fig. 1 is the overall structure figure of the convolutional neural networks of design；
Fig. 2 is the Inception structure charts in convolutional neural networks structure；
Fig. 3 is the flow chart that sample is configured to after handling interference signal in the embodiment of the present invention；
Fig. 4 is the accuracy rate curve under different signal-to-noise ratio environment to interference backward in present example.
Specific implementation mode
With reference to embodiment and attached drawing, detailed description of the present invention technical solution：
Embodiment：
The convolutional neural networks structure for disturbance ecology of this example is as shown in Figure 1：Input layer uses the sample of N × 1 × 3
Format inputs, and N is positive integer；Then the network being made of convolution, activation primitive, Chi Hua, batch normalization by two layers；Then lead to
Four layers of network being made of Inception, Chi Hua, batch normalization are crossed, wherein Inception structures are as shown in Fig. 2, it is by passing through
Cross the operation being concatenated together after a variety of nonidentity operations (convolution or pond)；Finally by a full articulamentum and Softmax into
Row classification.
The convolution that is wherein arrived involved in structure, Chi Hua, batch normalization, Softmax operations are as follows：
Convolution algorithm is by taking three-dimensional input data as an example, if three-dimensional input data is I, corresponding convolution kernel K is also three
Dimension, the third dimension dimension of wherein K is always consistent with input I, before the size of the convolution kernel introduced below all only illustrates
Two dimensions.After convolution operation, exports and be
There are maximum pond, mean value pond in pond, and operation is by taking three-dimensional input data as an example, if three-dimensional input data is
X, calculation are as follows：
Mean value pond：
Maximum value pond：
Normalization is criticized, operation is as follows
Wherein, m is the number of samples in a batch, xiFor i-th of sample in batch, ε is set as one and levels off to 0
Positive number, yiFor i-th of output in output batch, the parameter learnt is needed when γ, β are network training.
Softmax is the operation that input data is switched to probability, and neuron i exports oiIt is defined as follows：
Wherein xkIt is inputted for k-th of neuron.
Specifically, in the present embodiment, each layer of structure is specific as follows shown：
First layer is input layer, and input size is identical as sample tensor size, and 1024 are taken for N × 1 × 3, such as N.
The second layer is made of convolution, Relu activation primitives, pondization and batch normalization.Specifically, in the present embodiment, first
It is to carry out convolution, wherein convolution kernel size is 5 × 1, step-length 2, and convolution kernel number is 8；Then pass through Relu activation primitives；It connects
Carry out pond, pondization is using maximum pond, and pond size is 3 × 1, step-length 2；Finally by batch normalization.Wherein Relu
Activation primitive is to carry out nonlinear transformation to the number of input, is exported to inputting number x, after operation as follows
Y=max { 0, x } (6)
Third layer is equally made of convolution, Relu activation primitives, pondization and batch normalization.It is progress convolution first, wherein
Convolution kernel size is 3 × 1, step-length 2, and convolution kernel number is 8；Then pass through Relu activation primitives；Then pond, Chi Hua are carried out
Using maximum pond, pond size is 3 × 1, step-length 2；Finally by batch normalization.
It 4th layer, is made of Inception, pondization and batch normalization.Pass through Inception, Inception structures first
As shown in Fig. 2, in structure, the convolution kernel size of convolution 1 and convolution 4 is 1 × 1, and convolution kernel number is 4, convolution 2 and convolution 3
Convolution kernel size is 1 × 1, and convolution kernel number is 3；Pond in Inception is using maximum pond；Convolution kernel size is
3 × 1,5 × 1 convolution kernel number is all 4.Then pond, using maximum pond, size is 3 × 1, step-length 2.It is most laggard
Row batch normalization.
Layer 5 is made of Inception, pondization and batch normalization.Pass through Inception, Inception structures first
As shown in Fig. 2, in structure, the convolution kernel size of convolution 1 and convolution 4 is 1 × 1, and convolution kernel number is 8, convolution 2 and convolution 3
Convolution kernel size is 1 × 1, and convolution kernel number is 6；Pond in Inception is using maximum pond；Convolution kernel size is
3 × 1,5 × 1 convolution kernel number is all 8.Then pond, using maximum pond, size is 3 × 1, step-length 2.It is most laggard
Row batch normalization.
Layer 6 is made of Inception, pondization and batch normalization.Pass through Inception, Inception structures first
As shown in Fig. 2, in structure, the convolution kernel number that the convolution kernel size of convolution 1 and convolution 4 is 1 × 1 is 8, convolution 2 and convolution 3
Convolution kernel size is 1 × 1, and convolution kernel number is 6；Pond in Inception is using maximum pond；Convolution kernel size is
3 × 1,5 × 1 convolution kernel number is all 8.Then pond, using maximum pond, size is 3 × 1, step-length 2.It is most laggard
Row batch normalization.
Layer 7 is made of Inception, pondization and batch normalization.Pass through Inception, Inception structures first
As shown in Fig. 2, in structure, the convolution kernel number that the convolution kernel size of convolution 1 and convolution 4 is 1 × 1 is 16, convolution 2 and convolution 3
Convolution kernel size be 1 × 1, convolution kernel number be 12；Pond in Inception is using maximum pond；Convolution kernel ruler
The very little convolution kernel number for being 3 × 1,5 × 1 is all 16.Then pond, using mean value pond, size is 8 × 1, step-length 8.Most
After carry out batch normalization.
It 8th layer, is formed with softmax by connecting entirely.Wherein full connection weight coefficient matrix is the size of 64 × M, behind
Connect the probability of a softmax output M kind sorting signal.
Fig. 3 be the present invention case study on implementation in the preprocessing part of signal, explain each step in detail with reference to Fig. 3
Rapid and its principle.
Step S1.1 in step S1 after receiving terminal receives signal first, down-converts the signals to base band, then uses low pass
Filter filters out of band signal, is sampled using A/D, sampling number N, such as N=1024, obtains digital complex baseband signal y
(n), n=1,2,3...., 1024.
Step S1.2 normalizes to obtain to digital signal powerI.e.
WhereinThen rightIt is the FFT that length is N and modulus is worth to amplitude spectrum, by amplitude spectrum
WithReal part, imaginary part form one 1024 × 1 × 3 sample tensor.
Step S2.1 in step S2 will by the pretreated sample of each known class and class label composing training collection
Input of the sample as convolutional neural networks in training set, output of the label vector as convolutional neural networks.
Specifically, need the signal kinds identified for five kinds of interference signals and as noiseless mark in the implementation case
Gaussian noise signal, five kinds of interference signals are respectively：Single tone jamming, linear frequency sweep interference, Partial band noise jammer, is made an uproar at Multi-tone jamming
Tone frequency interferes.This six kinds of signals are respectively labeled as 1,2 ..., 6.For arbitrary signal, it is assumed that label is k ∈ belonging to it
Its classification is then mapped as the vector of one 6 dimension by [1,6], wherein the kth position of vector is 1, remaining is all 0, this vector is made
For the label of signal.In training set, five kinds of interference signals of present example 51 different signal-to-noise ratio (- 5dB arrives 20dB,
Using 0.5dB as interval) under respectively generate 100 interference signal sample tensors, Gaussian noise signal randomly generates 5100 samples
Amount, finally by each sample tensor and its label vector to should be used as a training set.Sample tensor in training set is as volume
The input of product neural network, output of the label vector as convolutional neural networks.
Step S2.2 upsets the data in training set at random, and then (number of samples can be set in a batch in batches
Set suitable value, such as 128,256 etc., the present embodiment 128) in the convolutional neural networks of In-put design, obtain the defeated of batch
Go out.Convolutional neural networks structure therein and its design parameter are using the example in embodiment one.
Weight parameter is learnt using the method for error back propagation, wherein error function is to intersect entropy loss letter
Number.The loss function of each of which sample is defined as：
Wherein ymFor m-th of element in sample label vector, M is the dimension of label vector,To pass through Softmax layers
M-th of element of the vector exported afterwards.Each layer needs the parameter learnt update rule as follows：
Wherein η is learning rate, and W is the parameter that the needs in network learn, and 0.001 is taken in this reality embodiment.
Each unknown sample is inputed to trained convolutional neural networks by step S3.1 in step S3, and it is right with it to obtain
The output vector answered.In the implementation case above-mentioned five kinds of interference signals and gaussian signal are generated by emulating.To every a kind of signal,
In 16 different signal-to-noise ratio, (- 10dB arrives 20dB, and 1000 sample of signal collection are respectively generated under being divided into 2).Then to these signals
It does pretreatment and obtains sample tensor.These samples are input to trained convolutional neural networks, obtain corresponding output vector.
Step S3.2 obtains the recognition result of signal to the index where maximum value in obtained output vector amount of orientation,
Then by the recognition result compared with actual signal classification, discrimination of the disturbance signal under different signal-to-noise ratio is obtained.
Further, Fig. 4 is illustrated in present example and is produced to step S3 using the neural network that step S2 learns
The recognition accuracy of unlike signal is as a result, wherein the discrimination of gaussian signal is also drawn in as a comparison under raw different signal-to-noise ratio
In figure.It can be seen from the figure that the discrimination of gaussian signal is almost 100%, when signal-to-noise ratio is higher than -2dB, each signal
Correct recognition rata is also almost 100%.
Claims (2)
1. the interference signal recognition methods based on convolutional neural networks, which is characterized in that include the following steps：
A, convolutional neural networks are built：
Input layer is inputted using the sample format of N × 1 × 3, and N is positive integer；Then by two layers by convolution, activation primitive, pond
Change, the network of batch normalization composition；Then the network being made of Inception, Chi Hua, batch normalization by four layers, wherein
Inception is the operation by being concatenated together after a variety of nonidentity operations；Finally by a full articulamentum and Softmax
Classify；
B, the interference signal received is pre-processed, as the input sample of convolutional neural networks；
C, according to the classification of signal to be identified, sample of signal and its corresponding classification is configured to training set, utilize the instruction of structure
Practice the convolutional neural networks built in collection training step a；
D, according to trained convolutional neural networks, each pretreated sample of signal is identified, unknown signaling is obtained
Generic.
2. the interference signal recognition methods according to claim 1 based on convolutional neural networks, which is characterized in that the step
Pretreated specific method is in rapid b：
B1, down coversion is carried out to the interference signal received, complex baseband signal is then obtained by low-pass filter and down-sampling；
B2, to complex baseband signal power normalized, FFT then is done to normalized complex baseband signal and obtains its amplitude
The real part of complex baseband signal after normalization, imaginary part and amplitude spectrum are formed a sample of signal by spectrum.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810286070.6A CN108509911B (en) | 2018-04-03 | 2018-04-03 | Interference signal identification method based on convolutional neural network |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201810286070.6A CN108509911B (en) | 2018-04-03 | 2018-04-03 | Interference signal identification method based on convolutional neural network |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN108509911A true CN108509911A (en) | 2018-09-07 |
| CN108509911B CN108509911B (en) | 2020-06-12 |
Family
ID=63379710
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201810286070.6A Active CN108509911B (en) | 2018-04-03 | 2018-04-03 | Interference signal identification method based on convolutional neural network |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN108509911B (en) |
Cited By (16)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109274624A (en) * | 2018-11-07 | 2019-01-25 | 中国电子科技集团公司第三十六研究所 | A kind of carrier frequency bias estimation based on convolutional neural networks |
| CN109299697A (en) * | 2018-09-30 | 2019-02-01 | 泰山学院 | Deep neural network system and method based on underwater sound communication Modulation Mode Recognition |
| CN109709553A (en) * | 2019-02-01 | 2019-05-03 | 北京航空航天大学 | A kind of radio frequency compacting interference classification method based on convolutional neural networks |
| CN110515096A (en) * | 2019-08-20 | 2019-11-29 | 东南大学 | Satellite navigation interference signal identification device and its method based on convolutional neural networks |
| CN110557209A (en) * | 2019-07-19 | 2019-12-10 | 中国科学院微电子研究所 | Broadband signal interference monitoring method |
| CN110659684A (en) * | 2019-09-23 | 2020-01-07 | 中国人民解放军海军航空大学 | Convolutional neural network-based STBC signal identification method |
| CN110996343A (en) * | 2019-12-18 | 2020-04-10 | 中国人民解放军陆军工程大学 | Interference recognition model based on deep convolutional neural network and intelligent recognition algorithm |
| CN111562597A (en) * | 2020-06-02 | 2020-08-21 | 南京敏智达科技有限公司 | Beidou satellite navigation interference source identification method based on BP neural network |
| CN111585671A (en) * | 2020-04-15 | 2020-08-25 | 国网河南省电力公司郑州供电公司 | Electric power LTE wireless private network electromagnetic interference monitoring and identifying method |
| CN111783558A (en) * | 2020-06-11 | 2020-10-16 | 上海交通大学 | Satellite navigation interference signal type intelligent identification method and system |
| CN112285666A (en) * | 2020-10-21 | 2021-01-29 | 电子科技大学 | Radar interference suppression method based on deep U-inclusion network |
| CN112332866A (en) * | 2020-10-28 | 2021-02-05 | 成都海擎科技有限公司 | Method for identifying cascade code parameters based on DVB-S and DVB-S2 signals |
| CN112491442A (en) * | 2020-11-17 | 2021-03-12 | 中山大学 | Self-interference elimination method and device |
| CN113011556A (en) * | 2021-02-20 | 2021-06-22 | 安徽大学 | Method for establishing network identification model based on INC-DenseUnet |
| CN113569742A (en) * | 2021-07-29 | 2021-10-29 | 西南交通大学 | Broadband electromagnetic interference source identification method based on convolutional neural network |
| CN115563485A (en) * | 2022-09-14 | 2023-01-03 | 电子科技大学 | Low-complexity interference identification method based on deep learning |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103226028A (en) * | 2013-05-02 | 2013-07-31 | 电子科技大学 | Method for identifying and detecting disturbance signals of phase-sensitive optical time domain reflectometer |
| CN106357575A (en) * | 2016-10-17 | 2017-01-25 | 中国电子科技集团公司第五十四研究所 | Multi-parameter jointly-estimated interference type identification method |
| CN107203782A (en) * | 2017-05-23 | 2017-09-26 | 哈尔滨工业大学 | Communication interference signals recognition methods under Larger Dynamic signal to noise ratio based on convolutional neural networks |
| CN107220606A (en) * | 2017-05-22 | 2017-09-29 | 西安电子科技大学 | The recognition methods of radar emitter signal based on one-dimensional convolutional neural networks |
-
2018
- 2018-04-03 CN CN201810286070.6A patent/CN108509911B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN103226028A (en) * | 2013-05-02 | 2013-07-31 | 电子科技大学 | Method for identifying and detecting disturbance signals of phase-sensitive optical time domain reflectometer |
| CN106357575A (en) * | 2016-10-17 | 2017-01-25 | 中国电子科技集团公司第五十四研究所 | Multi-parameter jointly-estimated interference type identification method |
| CN107220606A (en) * | 2017-05-22 | 2017-09-29 | 西安电子科技大学 | The recognition methods of radar emitter signal based on one-dimensional convolutional neural networks |
| CN107203782A (en) * | 2017-05-23 | 2017-09-26 | 哈尔滨工业大学 | Communication interference signals recognition methods under Larger Dynamic signal to noise ratio based on convolutional neural networks |
Cited By (21)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN109299697A (en) * | 2018-09-30 | 2019-02-01 | 泰山学院 | Deep neural network system and method based on underwater sound communication Modulation Mode Recognition |
| CN109274624B (en) * | 2018-11-07 | 2021-04-27 | 中国电子科技集团公司第三十六研究所 | Carrier frequency offset estimation method based on convolutional neural network |
| CN109274624A (en) * | 2018-11-07 | 2019-01-25 | 中国电子科技集团公司第三十六研究所 | A kind of carrier frequency bias estimation based on convolutional neural networks |
| CN109709553A (en) * | 2019-02-01 | 2019-05-03 | 北京航空航天大学 | A kind of radio frequency compacting interference classification method based on convolutional neural networks |
| CN110557209A (en) * | 2019-07-19 | 2019-12-10 | 中国科学院微电子研究所 | Broadband signal interference monitoring method |
| CN110557209B (en) * | 2019-07-19 | 2021-08-31 | 中国科学院微电子研究所 | Broadband signal interference monitoring method |
| CN110515096A (en) * | 2019-08-20 | 2019-11-29 | 东南大学 | Satellite navigation interference signal identification device and its method based on convolutional neural networks |
| CN110659684A (en) * | 2019-09-23 | 2020-01-07 | 中国人民解放军海军航空大学 | Convolutional neural network-based STBC signal identification method |
| CN110996343A (en) * | 2019-12-18 | 2020-04-10 | 中国人民解放军陆军工程大学 | Interference recognition model based on deep convolutional neural network and intelligent recognition algorithm |
| CN111585671A (en) * | 2020-04-15 | 2020-08-25 | 国网河南省电力公司郑州供电公司 | Electric power LTE wireless private network electromagnetic interference monitoring and identifying method |
| CN111585671B (en) * | 2020-04-15 | 2022-06-10 | 国网河南省电力公司郑州供电公司 | Electric power LTE wireless private network electromagnetic interference monitoring and identifying method |
| CN111562597A (en) * | 2020-06-02 | 2020-08-21 | 南京敏智达科技有限公司 | Beidou satellite navigation interference source identification method based on BP neural network |
| CN111783558A (en) * | 2020-06-11 | 2020-10-16 | 上海交通大学 | Satellite navigation interference signal type intelligent identification method and system |
| CN112285666A (en) * | 2020-10-21 | 2021-01-29 | 电子科技大学 | Radar interference suppression method based on deep U-inclusion network |
| CN112332866A (en) * | 2020-10-28 | 2021-02-05 | 成都海擎科技有限公司 | Method for identifying cascade code parameters based on DVB-S and DVB-S2 signals |
| CN112332866B (en) * | 2020-10-28 | 2024-04-30 | 成都海擎科技有限公司 | Cascade code parameter identification method based on DVB-S and DVB-S2 signals |
| CN112491442A (en) * | 2020-11-17 | 2021-03-12 | 中山大学 | Self-interference elimination method and device |
| CN113011556A (en) * | 2021-02-20 | 2021-06-22 | 安徽大学 | Method for establishing network identification model based on INC-DenseUnet |
| CN113011556B (en) * | 2021-02-20 | 2022-10-11 | 安徽大学 | Method for establishing network identification model based on INC-DenseUnet |
| CN113569742A (en) * | 2021-07-29 | 2021-10-29 | 西南交通大学 | Broadband electromagnetic interference source identification method based on convolutional neural network |
| CN115563485A (en) * | 2022-09-14 | 2023-01-03 | 电子科技大学 | Low-complexity interference identification method based on deep learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN108509911B (en) | 2020-06-12 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN108509911A (en) | Interference signal recognition methods based on convolutional neural networks | |
| CN111913156B (en) | Radar radiation source individual identification method based on deep learning model and feature combination | |
| CN107682109B (en) | A kind of interference signal classifying identification method suitable for UAV Communication system | |
| CN107203782A (en) | Communication interference signals recognition methods under Larger Dynamic signal to noise ratio based on convolutional neural networks | |
| CN108647614A (en) | The recognition methods of electrocardiogram beat classification and system | |
| CN109495214B (en) | Channel coding type identification method based on one-dimensional inclusion structure | |
| Errington et al. | Application of artificial neural networks to chromosome classification | |
| CN110120926A (en) | Modulation mode of communication signal recognition methods based on evolution BP neural network | |
| Cruz et al. | Handwritten digit recognition using multiple feature extraction techniques and classifier ensemble | |
| CN110751121B (en) | Unsupervised radar signal sorting method based on clustering and SOFM | |
| CN106628097A (en) | Ship equipment fault diagnosis method based on improved radial basis function neutral network | |
| CN110417694A (en) | A kind of modulation mode of communication signal recognition methods | |
| CN110175615A (en) | The adaptive visual position recognition methods in model training method, domain and device | |
| CN111680642B (en) | Terrain classification method and device | |
| CN110515096A (en) | Satellite navigation interference signal identification device and its method based on convolutional neural networks | |
| CN113759323B (en) | Signal sorting method and device based on improved K-Means joint convolution self-encoder | |
| CN116866129A (en) | Wireless communication signal detection method | |
| CN113098862A (en) | Intrusion detection method based on combination of hybrid sampling and expansion convolution | |
| CN114943253A (en) | Radio frequency fingerprint small sample identification method based on meta-learning model | |
| CN114492522B (en) | Automatic modulation classification method based on improved stacked hourglass neural network | |
| CN117233706B (en) | Radar active interference identification method based on multilayer channel attention mechanism | |
| CN112231775A (en) | Hardware Trojan horse detection method based on Adaboost algorithm | |
| CN116482618B (en) | Radar active interference identification method based on multi-loss characteristic self-calibration network | |
| CN113109782A (en) | Novel classification method directly applied to radar radiation source amplitude sequence | |
| Ruan et al. | Automatic recognition of radar signal types based on CNN-LSTM |
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 |