CN113518049B - Modulation identification method based on fractional low-order polar coordinates and deep learning - Google Patents
Modulation identification method based on fractional low-order polar coordinates and deep learning Download PDFInfo
- Publication number
- CN113518049B CN113518049B CN202110394895.1A CN202110394895A CN113518049B CN 113518049 B CN113518049 B CN 113518049B CN 202110394895 A CN202110394895 A CN 202110394895A CN 113518049 B CN113518049 B CN 113518049B
- Authority
- CN
- China
- Prior art keywords
- order
- signal
- fractional
- deep learning
- training
- 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
- 238000013135 deep learning Methods 0.000 title claims abstract description 46
- 238000000034 method Methods 0.000 title claims abstract description 32
- 238000012549 training Methods 0.000 claims abstract description 34
- 238000012360 testing method Methods 0.000 claims abstract description 20
- 238000012545 processing Methods 0.000 claims abstract description 16
- 238000004519 manufacturing process Methods 0.000 claims abstract description 4
- 230000006870 function Effects 0.000 claims description 13
- 238000012795 verification Methods 0.000 claims description 9
- 108010076504 Protein Sorting Signals Proteins 0.000 claims description 6
- 238000013507 mapping Methods 0.000 claims description 6
- 238000005070 sampling Methods 0.000 claims description 6
- 238000005094 computer simulation Methods 0.000 claims description 3
- 238000004364 calculation method Methods 0.000 abstract description 7
- 238000010586 diagram Methods 0.000 description 5
- 238000004891 communication Methods 0.000 description 3
- 238000013528 artificial neural network Methods 0.000 description 2
- 230000005540 biological transmission Effects 0.000 description 2
- 238000004422 calculation algorithm Methods 0.000 description 2
- 230000002159 abnormal effect Effects 0.000 description 1
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000013527 convolutional neural network Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
- 238000011161 development Methods 0.000 description 1
- 238000012850 discrimination method Methods 0.000 description 1
- 230000000694 effects Effects 0.000 description 1
- 238000005516 engineering process Methods 0.000 description 1
- 238000002474 experimental method Methods 0.000 description 1
- 238000003909 pattern recognition Methods 0.000 description 1
- 238000011160 research Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
- 238000010200 validation analysis Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
- H04L27/00—Modulated-carrier systems
- H04L27/0012—Modulated-carrier systems arrangements for identifying the type of modulation
-
- 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
Landscapes
- Engineering & Computer Science (AREA)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Health & Medical Sciences (AREA)
- Life Sciences & Earth Sciences (AREA)
- Artificial Intelligence (AREA)
- Biomedical Technology (AREA)
- Biophysics (AREA)
- Computational Linguistics (AREA)
- Signal Processing (AREA)
- Evolutionary Computation (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- Computing Systems (AREA)
- General Engineering & Computer Science (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- Software Systems (AREA)
- Computer Networks & Wireless Communication (AREA)
- Image Analysis (AREA)
- Digital Transmission Methods That Use Modulated Carrier Waves (AREA)
Abstract
A modulation identification method based on fractional lower-order polar coordinates and deep learning comprises the following steps: collecting signals and performing fractional low-order processing on the signals; calculating fractional low-order polar coordinate characteristics and manufacturing a training set and a testing set; constructing and training a light deep learning network; and testing the deep learning network and carrying out signal modulation recognition. The method provided by the invention has wide signal coverage variety, can obviously improve the recognition accuracy of the signal modulation mode under the condition of impulse noise interference, and simultaneously can obviously reduce the calculation cost in the training and using process by the light deep learning network.
Description
Technical Field
The invention relates to the technical field of wireless communication, in particular to a modulation identification method based on fractional low-order polar coordinates and deep learning. The invention is characterized by taking the fractional low-order polar coordinates and taking the deep learning network as a classifier, and efficiently and accurately realizes the modulation mode identification of various signals under the condition of impulse noise interference.
Background
As an intermediate link between signal reception and signal demodulation, signal modulation scheme identification is an indispensable important step. With the development of wireless communication technology, factors such as signal types, transmission modes, transmission environments, etc. are becoming increasingly diverse and complex. As an important research topic in the military and civil fields, signal modulation scheme recognition is increasingly showing its importance in tasks such as communication reconnaissance, electronic interference, electronic countermeasure, abnormal signal recognition, and the like. This requires researchers to develop more efficient, accurate, and widely applicable signal modulation recognition methods.
Whereas in the past conventional feature-based signal modulation recognition methods, different kinds of features are typically manually selected. A large number of technicians observe information such as the frequency spectrum of a signal, and determine the type of signal modulation scheme based on their own experience. Such subjective factors are overly significant and labor and time costs are high, and new methods are being urgently needed to address these drawbacks.
Since the early generation of the century, hinton et al proposed a deep neural network algorithm, the field of deep learning has emerged as CNN, alexNet, googLeNet, VGGNet, resNet a number of excellent networks and algorithms, which are widely used in the fields of computer vision, image processing, and the like. Until now, the accuracy and timeliness of performing classification and identification tasks by deep learning networks represented by ResNet (Residual Network) have been greatly superior to that of the traditional manual-based discrimination method. In addition, by means of the networks, the defect that the traditional method excessively depends on manual experience and subjective judgment in the signal modulation recognition task can be greatly reduced. However, although many methods for identifying signal modulations based on deep learning networks have emerged, the current methods and techniques have not addressed two major problems involved in this type of method at the same time, namely: non-gaussian noise conditions and lightweight deep learning networks.
Disclosure of Invention
The invention provides a modulation recognition method based on fractional low-order polar coordinates and deep learning, which aims to solve the problems of low accuracy, few recognition signal types, high calculation cost, poor timeliness and the like of the existing signal modulation recognition method under the condition that signals are interfered by impulse noise. The present invention first proposes the concept of fractional lower order polar coordinates of typical non-gaussian noise that can suppress impulse noise. Then a lightweight deep learning network with low calculation cost is constructed, and finally a modulation identification method based on the fractional low-order polar coordinates and the deep learning is provided. From the viewpoint of pattern recognition, the fractional lower-order polar coordinates are features and serve as inputs of a network, and the lightweight deep learning network is a classifier and is used for classifying and recognizing various signals. Experiments prove that the method provided by the invention can obviously improve the signal identification accuracy rate interfered by impulse noise, and simultaneously, the lightweight deep learning network can obviously reduce the calculation cost in the training and using processes.
The scheme of the invention is as follows:
a modulation identification method based on fractional lower-order polar coordinates and deep learning comprises the following steps:
A: collecting signals and performing fractional low-order processing on the signals;
b: calculating fractional low-order polar coordinate characteristics and manufacturing a training set and a testing set;
c: constructing and training a light deep learning network;
d: and testing the deep learning network and identifying the signal modulation mode.
Further, the step a includes:
A1: collecting signals of different modulation modes interfered by impulse noise, collecting enough signals interfered by impulse noise from the nature through computer simulation or by utilizing receiving equipment, and describing the non-Gaussian noise by Alpha stable distribution; the generalized signal-to-noise ratio is adopted to measure the intensity of impulse noise, and the definition formula is as follows:
GSNR=10log10(Ps/Pn)
in the above formula, P s represents the power of the signal, P n represents the generalized power of noise, P n =γ, and γ represents the scale parameter of Alpha stable distribution;
A2: performing fractional low-order processing on the acquired signals; and carrying out fractional lower-order processing on the acquired signals by using a fractional lower-order mapping function, wherein the specific formula is as follows:
yFLO(n)=(y(n))<p-1>
=ρFLO(n)exp(jθFLO(n))
Where n represents a discrete time variable corresponding to a sampling time, y (n) represents a sampled signal sequence, y FLO (n) represents a signal sequence after fractional lower order processing, ρ FLO represents a polar path of a signal, θ FLO represents a polar angle of the signal, < · > represents a fractional lower order operator, and the fractional lower order function satisfies the following relation:
Wherein z represents a complex domain The superscript x indicates the conjugate operator.
Further, the step B includes:
B1: calculating the fractional low-order polar coordinate characteristics of the signal subjected to fractional low-order processing; taking a polar angle theta FLO of the signal as an abscissa, taking a polar diameter rho FLO of the signal as an ordinate, and extracting fractional low-order polar coordinate features corresponding to the signal one by one according to a sampling time n;
B2: fully mixing the characteristics, and forming a training set, a verification set and a test set by the mixed characteristics according to a certain proportion; and B1, disturbing and fully mixing the fractional lower-order polar coordinate characteristics obtained in the step, and then respectively forming a training set, a verification set and a test set according to the proportion of m 1:m2:m3. The m 1:m2:m3 =6:2:2.
Further, the step C includes:
c1: adopting fewer convolution layers to construct a lightweight deep learning network;
C2: taking the training set obtained in the step B2 as input, and training the training set; in the training process, cross entropy is adopted as a loss function, RMSprop is adopted as a network optimizer, and the learning rate is set to be 0.01, so that the updating of the modulus parameters is realized;
and C3: the features in the verification set are used as the input of the deep learning network in the step C1, and are verified.
Further, the step D includes:
D1: taking the characteristics in the test set as the input of the deep learning network in C1, and testing the characteristics;
D2: taking the fractional low-order polar coordinate characteristic of the signal of the unknown modulation mode as the input of the deep learning network in the step C1, and obtaining the recognition results of different signal modulation modes after model training
Compared with the prior art, the invention has the beneficial technical effects that:
The method of the invention solves the problem that the accuracy of the existing modulation mode identification method is reduced by improving the typical non-Gaussian noise interference environment of the signal subjected to impulse noise. Meanwhile, the invention relates to various signal modulation modes, solves the problems of high calculation cost, poor timeliness and the like commonly existing in the deep learning network, and realizes the efficient and accurate identification of various signal modulation modes.
Drawings
Fig. 1 is a modulation recognition method based on fractional lower-order polar coordinates and deep learning according to the present invention.
Fig. 2 is a graph of a fractional lower order mapping function in accordance with the present invention. The 5 curves correspond to the fractional lower order parameters p=1.1, 1.3, 1.5, 1.7, 1.9, respectively.
Fig. 3 is a fractional lower order polar characteristic diagram of a signal in accordance with the present invention. Taking the fractional lower order function parameters p=1.2, 1.5 and 1.8 as examples, fractional lower order polar coordinate feature maps of 6 signal modulation modes of 2PSK, 4PSK, 8PSK, 16QAM, 32QAM and 64QAM are listed.
FIG. 4 is a block diagram of a deep learning network according to the present invention
FIG. 5 is a graph of recognition accuracy for different methods under different noise conditions in accordance with the present invention. The values of the parameter p related to the fractional lower-order polar coordinates are 1.1, 1.3 and 1.5 respectively, and the comparison group adopts the original polar coordinates as characteristics.
Detailed Description
In order to facilitate understanding, the implementation process of the present invention will be clearly and specifically described below with reference to the technical scheme and the accompanying drawings.
A modulation identification method based on fractional lower-order polar coordinates and deep learning comprises the following steps:
A: the signal is collected and subjected to fractional lower order processing.
The step A specifically comprises the following steps:
a1: a sufficient amount of signals of different modulation modes interfered by impulse noise are acquired.
A sufficient amount of signals interfered by impulse noise are collected from nature through computer simulation or by using receiving equipment, and the classical non-gaussian noise is characterized by Alpha stable distribution. When the characteristic index α of the Alpha stable distribution is less than 2, the second moment of the noise is not converged, so that the signal-to-noise ratio (SNR) cannot be used for measuring the intensity of the noise, and the generalized signal-to-noise ratio (GSNR) is used for measuring the intensity of the impulse noise, which is defined as:
GSNR=10log10(Ps/Pn)
In the above expression, P s denotes the power of the signal, P n denotes the generalized power of noise, P n =γ, and γ denotes the scale parameter of the Alpha stable distribution. The types of the acquired signals comprise AM, FM, MSK, 2ASK, 2PSK, 4PSK, 8PSK, 16QAM, 32QAM and 64QAM. The generalized signal to noise coverage ranges from-5 dB to +15dB.
A2: and carrying out fractional lower-order processing on the acquired signals.
And carrying out fractional lower-order processing on the acquired signals by using a fractional lower-order mapping function, wherein the specific formula is as follows:
yFLO(n)=(y(n))<p-1>
=ρFLO(n)exp(jθFLO(n))
Where n represents a discrete time variable corresponding to a sampling time, y (n) represents a sampled signal sequence, y FLO (n) represents a signal sequence processed by fractional lower-order (FLO), ρ FLO represents a polar path of a signal, θ FLO represents a polar angle of the signal, and < <· > represents a fractional lower-order operator, where the fractional lower-order function satisfies the following relation:
Wherein z represents a complex domain The superscript x indicates the conjugate operator.
A graph of the fractional lower order mapping function according to the present invention is shown in fig. 2. The 5 curves correspond to parameters p=1.1, 1.3, 1.5, 1.7, 1.9, respectively.
B: and calculating fractional low-order polar coordinate characteristics and manufacturing a training set and a testing set.
The step B specifically comprises the following steps:
B1: and calculating the fractional lower-order polar coordinate characteristic of the signal after fractional lower-order processing.
And extracting fractional low-order polar coordinate features corresponding to the signals one by one according to the sampling time n by taking the polar angle theta FLO of the signals as an abscissa and the polar diameter rho FLO of the signals as an ordinate. The signal type and the generalized signal-to-noise ratio range of the features are matched with the signals acquired in the step A1.
B2: and fully mixing the characteristics, and forming a training set, a verification set and a test set by the mixed characteristics according to a certain proportion.
And B1, disturbing and fully mixing the fractional lower-order polar coordinate characteristics obtained in the step, and then respectively forming a training set, a verification set and a test set according to the proportion of m 1:m2:m3. This ratio is typically set to m 1:m2:m3 =6:2:2.
The fractional lower-order polar coordinate feature diagram related to the step B1 in the present invention is shown in fig. 3. Taking the values of the parameter p related to the fractional lower-order mapping function being equal to 1.2, 1.5 and 1.8 as examples, the fractional lower-order polar coordinate characteristic diagrams of 6 signal modulation modes, namely 2PSK, 4PSK, 8PSK, 16QAM, 32QAM and 64QAM, are listed.
C: and constructing and training a light deep learning network.
The step C specifically comprises the following steps:
C1: and constructing a lightweight deep learning network.
In order to reduce the calculation and time cost of training, a light deep learning network is constructed by adopting fewer convolution layers, so that the problem of overhigh calculation cost in the conventional deep learning network training process is solved.
C2: the deep learning network is trained using a training set.
And B2, training the training set obtained in the step B2 by taking the training set as input. In the training process, cross entropy (cross entropy) is adopted as a loss function, RMSprop is adopted as a network optimizer, the learning rate is set to be 0.01, and the updating of the modulus parameters is realized.
And C3: the deep learning network is validated using a validation set.
The features in the verification set are used as the input of the deep learning network in the step C1, and are verified.
The deep neural network structure diagram according to the present invention is shown in fig. 4, and specific network structure parameter settings are labeled in fig. 4.
D: and testing the deep learning network and carrying out signal modulation recognition.
The step D specifically comprises the following steps:
d1: the deep learning network is tested using a test set.
The features in the test set are tested as input to the deep learning network in C1.
D2: and carrying out signal modulation identification.
And C1, taking the fractional low-order polar coordinate characteristic of the signal of the unknown modulation mode as the input of the deep learning network in the step C1, and obtaining the recognition results of different signal modulation modes after model training.
Table 1 shows the classification and identification results of 10 signals interfered by impulse noise under the conditions of α=1.5 and gsnr=15 dB, taking the fractional lower order parameter p=1.1 as an example.
TABLE 1 Signal modulation recognition results
The average graph of recognition accuracy of different signal modulation modes under different noise conditions is shown in fig. 5. The values of the parameter p related to the fractional lower-order polar coordinates are 1.1, 1.3 and 1.5 respectively, and the comparison method adopts the traditional original polar coordinates as characteristics.
Claims (1)
1. The modulation identification method based on the fractional lower-order polar coordinates and the deep learning is characterized by comprising the following steps of:
A: collecting signals and performing fractional low-order processing on the signals;
b: calculating fractional low-order polar coordinate characteristics and manufacturing a training set and a testing set;
c: constructing and training a light deep learning network;
D: testing the deep learning network and identifying the signal modulation mode;
the step A comprises the following steps:
A1: collecting signals of different modulation modes interfered by impulse noise, collecting enough signals interfered by impulse noise from the nature through computer simulation or by utilizing receiving equipment, and describing the non-Gaussian noise by Alpha stable distribution; the generalized signal-to-noise ratio is adopted to measure the intensity of impulse noise, and the definition formula is as follows:
GSNR=10log10(Ps/Pn)
in the above formula, P s represents the power of the signal, P n represents the generalized power of noise, P n =γ, and γ represents the scale parameter of Alpha stable distribution;
A2: performing fractional low-order processing on the acquired signals; and carrying out fractional lower-order processing on the acquired signals by using a fractional lower-order mapping function, wherein the specific formula is as follows:
yFLO(n)=(y(n))<p-1>
=ρFLO(n)exp(jθFLO(n))
Where n represents a discrete time variable corresponding to a sampling time, y (n) represents a sampled signal sequence, p represents an order of y (n) in a fractional lower order operation, y FLO (n) represents a signal sequence after fractional lower order processing, ρ FLO represents a polar path of a signal, θ FLO represents a polar angle of the signal, < > represents a fractional lower order operator, and the fractional lower order function satisfies the following relation:
Wherein z represents a complex domain The superscript x represents the conjugate operator;
the step B comprises the following steps:
B1: calculating the fractional low-order polar coordinate characteristics of the signal subjected to fractional low-order processing; taking a polar angle theta FLO of the signal as an abscissa, taking a polar diameter rho FLO of the signal as an ordinate, and extracting fractional low-order polar coordinate features corresponding to the signal one by one according to a sampling time n;
B2: fully mixing the characteristics, and forming a training set, a verification set and a test set by the mixed characteristics according to a certain proportion; b1, disturbing and fully mixing the fractional lower-order polar coordinate characteristics obtained in the step, and then respectively forming a training set, a verification set and a test set according to the proportion of m 1:m2:m3;
The m 1:m2:m3 = 6:2:2;
The step C comprises the following steps:
c1: adopting a convolution layer to construct a lightweight deep learning network;
C2: taking the training set obtained in the step B2 as input, and training the training set; in the training process, cross entropy is adopted as a loss function, RMSprop is adopted as a network optimizer, and the learning rate is set to be 0.01, so that the updating of the modulus parameters is realized;
and C3: taking the characteristics in the verification set as the input of the deep learning network in the step C1, and verifying the characteristics;
the step D comprises the following steps:
D1: taking the characteristics in the test set as the input of the deep learning network in C1, and testing the characteristics;
D2: and C1, taking the fractional low-order polar coordinate characteristic of the signal of the unknown modulation mode as the input of the deep learning network in the step C1, and obtaining the recognition results of different signal modulation modes after model training.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110394895.1A CN113518049B (en) | 2021-04-13 | 2021-04-13 | Modulation identification method based on fractional low-order polar coordinates and deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN202110394895.1A CN113518049B (en) | 2021-04-13 | 2021-04-13 | Modulation identification method based on fractional low-order polar coordinates and deep learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN113518049A CN113518049A (en) | 2021-10-19 |
CN113518049B true CN113518049B (en) | 2024-04-26 |
Family
ID=78061491
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN202110394895.1A Active CN113518049B (en) | 2021-04-13 | 2021-04-13 | Modulation identification method based on fractional low-order polar coordinates and deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN113518049B (en) |
Families Citing this family (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN113205000A (en) * | 2021-04-07 | 2021-08-03 | 江苏师范大学 | Modulation mode identification method based on Cauchy Score polar coordinate diagram |
Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE19814770C1 (en) * | 1998-04-02 | 1999-09-23 | Rohde & Schwarz | Process for the automatic, clock-free detection of signal modulation types |
DE102010013637A1 (en) * | 2010-04-01 | 2011-10-06 | Rohde & Schwarz Gmbh & Co. Kg | Method for classifying signals according to the type of modulation used |
CN103326975A (en) * | 2013-07-15 | 2013-09-25 | 西安电子科技大学 | Digital modulation signal identification method under Alpha stable distribution noise |
CN104052702A (en) * | 2014-06-20 | 2014-09-17 | 西安电子科技大学 | Method for identifying digital modulation signals in presence of complicated noise |
CN105785324A (en) * | 2016-03-11 | 2016-07-20 | 西安电子科技大学 | MGCSTFT-based chirp signal parameter estimation method |
CN107135176A (en) * | 2017-07-06 | 2017-09-05 | 电子科技大学 | Figure field communication signal modulate method based on fractional lower-order Cyclic Spectrum |
CN108282426A (en) * | 2017-12-08 | 2018-07-13 | 西安电子科技大学 | Radio signal recognition recognition methods based on lightweight depth network |
CN109309640A (en) * | 2018-09-08 | 2019-02-05 | 苏州大学 | Fanaticism format identification method based on machine learning |
CN110163282A (en) * | 2019-05-22 | 2019-08-23 | 西安电子科技大学 | Modulation Mode Recognition method based on deep learning |
CN110417694A (en) * | 2019-06-07 | 2019-11-05 | 吉林大学 | A kind of modulation mode of communication signal recognition methods |
KR102073935B1 (en) * | 2018-12-24 | 2020-02-05 | 충남대학교산학협력단 | Modulation recognition for radil signal |
CN111259798A (en) * | 2020-01-16 | 2020-06-09 | 西安电子科技大学 | Modulation signal identification method based on deep learning |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
US11443167B2 (en) * | 2019-04-12 | 2022-09-13 | Motorola Solutions, Inc. | Systems and methods for modulation classification of baseband signals using multiple data representations of signal samples |
-
2021
- 2021-04-13 CN CN202110394895.1A patent/CN113518049B/en active Active
Patent Citations (12)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
DE19814770C1 (en) * | 1998-04-02 | 1999-09-23 | Rohde & Schwarz | Process for the automatic, clock-free detection of signal modulation types |
DE102010013637A1 (en) * | 2010-04-01 | 2011-10-06 | Rohde & Schwarz Gmbh & Co. Kg | Method for classifying signals according to the type of modulation used |
CN103326975A (en) * | 2013-07-15 | 2013-09-25 | 西安电子科技大学 | Digital modulation signal identification method under Alpha stable distribution noise |
CN104052702A (en) * | 2014-06-20 | 2014-09-17 | 西安电子科技大学 | Method for identifying digital modulation signals in presence of complicated noise |
CN105785324A (en) * | 2016-03-11 | 2016-07-20 | 西安电子科技大学 | MGCSTFT-based chirp signal parameter estimation method |
CN107135176A (en) * | 2017-07-06 | 2017-09-05 | 电子科技大学 | Figure field communication signal modulate method based on fractional lower-order Cyclic Spectrum |
CN108282426A (en) * | 2017-12-08 | 2018-07-13 | 西安电子科技大学 | Radio signal recognition recognition methods based on lightweight depth network |
CN109309640A (en) * | 2018-09-08 | 2019-02-05 | 苏州大学 | Fanaticism format identification method based on machine learning |
KR102073935B1 (en) * | 2018-12-24 | 2020-02-05 | 충남대학교산학협력단 | Modulation recognition for radil signal |
CN110163282A (en) * | 2019-05-22 | 2019-08-23 | 西安电子科技大学 | Modulation Mode Recognition method based on deep learning |
CN110417694A (en) * | 2019-06-07 | 2019-11-05 | 吉林大学 | A kind of modulation mode of communication signal recognition methods |
CN111259798A (en) * | 2020-01-16 | 2020-06-09 | 西安电子科技大学 | Modulation signal identification method based on deep learning |
Non-Patent Citations (12)
Title |
---|
A Novel Deep Learning and Polar Transformation Framework for an Adaptive Automatic Modulation Classification;Pejman Ghasemzadeh et al.;《IEEE Transactions on Vehicular Technology》;第13245-13255页 * |
Alpha稳定分布噪声下数字调制信号识别;孙汝峰;刘顺兰;;无线通信技术(第01期);全文 * |
Alpha稳定分布噪声下数字调制识别新方法;刘明骞;李兵兵;石亚云;;西安电子科技大学学报(第06期);全文 * |
CNN-based automatic modulation recognition of wireless signal;Kai Ma et al.;《2020 IEEE 3rd international conference on information systems and computer aided education (ICISCAE)》;全文 * |
modulation format identification based on constellation diagrams in adaptive optical OFDM systems;yuanyuan ma et al.;《optics communications》;全文 * |
modulation pattern recognition of communication signals based on fractional Low-Order Choi-Williams distribution and convolutional neural network in impulsive noise environment;xiaodi tian et al.;《2019 IEEE 19th international conference on communication technology (ICCT)》;全文 * |
Robust Modulation Classification Over α-Stable Noise Using Graph-Based Fractional Lower-Order Cyclic Spectrum Analysis;Xiao Yan et al.;《IEEE Transactions on Vehicular Technology》;第2837-2844页 * |
基于卷积神经网络的调制信号识别;李书恋;郑美芳;杨林森;;科技经济导刊(第27期);全文 * |
常用数字通信信号调制方式分类的一种新方法;钟骥等;《电子技术应用》;20081231;全文 * |
袁兰杰等.分数阶傅里叶变换(FRFT)用于调制方式识别.《 电子信息对抗技术》.2011,全文. * |
认知无线电中非高斯噪声下数字调制信号识别方法;刘明骞等;《通信学报》;全文 * |
采用分数低阶循环谱相干系数的调制识别;赵春晖;杨伟超;杜宇;;应用科学学报(第06期);全文 * |
Also Published As
Publication number | Publication date |
---|---|
CN113518049A (en) | 2021-10-19 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110855591B (en) | QAM and PSK signal intra-class modulation classification method based on convolutional neural network structure | |
CN110163282B (en) | Modulation mode identification method based on deep learning | |
CN110086737B (en) | Communication signal modulation mode identification method based on graph neural network | |
CN107124381B (en) | Automatic identification method for digital communication signal modulation mode | |
CN111585671B (en) | Electric power LTE wireless private network electromagnetic interference monitoring and identifying method | |
CN112702294B (en) | Modulation recognition method for multi-level feature extraction based on deep learning | |
CN108764077B (en) | Digital signal modulation classification method based on convolutional neural network | |
CN109450834A (en) | Signal of communication classifying identification method based on Multiple feature association and Bayesian network | |
CN112733811B (en) | Method for identifying underwater sound signal modulation modes based on improved dense neural network | |
CN111428817A (en) | Defense method for resisting attack by radio signal identification | |
CN111585922A (en) | Modulation mode identification method based on convolutional neural network | |
CN110071885A (en) | A kind of deep learning method of discrimination of PSK digital signal subclass Modulation Identification | |
CN101764786A (en) | MQAM signal recognition method based on clustering algorithm | |
CN108052956A (en) | Wireless light communication subcarrier modulation constellation recognition methods under a kind of atmospheric turbulance | |
CN104869096B (en) | Bootstrap-based BPSK signal blind processing result credibility test method | |
CN113518049B (en) | Modulation identification method based on fractional low-order polar coordinates and deep learning | |
CN113037663B (en) | Improved code element rate estimation method suitable for non-constant envelope signal | |
CN110826425B (en) | VHF/UHF frequency band radio signal modulation mode identification method based on deep neural network | |
CN113259289B (en) | Single-channel aliasing signal modulation mode identification method based on residual error neural network | |
CN113239788A (en) | Mask R-CNN-based wireless communication modulation mode identification method | |
CN115409056A (en) | Automatic modulation identification method for large dynamic signal-to-noise ratio | |
CN114422310B (en) | Digital quadrature modulation signal identification method based on joint distribution matrix and multi-input neural network | |
CN115913849A (en) | Electromagnetic signal identification method based on one-dimensional complex value residual error network | |
CN113361569A (en) | Electromagnetic signal modulation type identification model construction method based on time-frequency fusion information | |
CN113269023A (en) | Modulation identification method based on fraction low-order constellation diagram and deep learning |
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 |