CN112953871A - New signal modulation format identification method based on neural network - Google Patents

New signal modulation format identification method based on neural network Download PDF

Info

Publication number
CN112953871A
CN112953871A CN202110203711.9A CN202110203711A CN112953871A CN 112953871 A CN112953871 A CN 112953871A CN 202110203711 A CN202110203711 A CN 202110203711A CN 112953871 A CN112953871 A CN 112953871A
Authority
CN
China
Prior art keywords
network
modulation format
identification
signal modulation
normalized density
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
Application number
CN202110203711.9A
Other languages
Chinese (zh)
Inventor
田清华
忻向军
姚海鹏
王阔
王瑞春
高然
张琦
胡鹏
王光全
付松年
杨雷静
常欢
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Beijing University of Posts and Telecommunications
Original Assignee
Beijing University of Posts and Telecommunications
Priority date (The priority date 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 date listed.)
Filing date
Publication date
Application filed by Beijing University of Posts and Telecommunications filed Critical Beijing University of Posts and Telecommunications
Priority to CN202110203711.9A priority Critical patent/CN112953871A/en
Publication of CN112953871A publication Critical patent/CN112953871A/en
Pending legal-status Critical Current

Links

Images

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
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • 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)
  • Theoretical Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • General Health & Medical Sciences (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • Biophysics (AREA)
  • Computational Linguistics (AREA)
  • Data Mining & Analysis (AREA)
  • Evolutionary Computation (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Molecular Biology (AREA)
  • Artificial Intelligence (AREA)
  • General Engineering & Computer Science (AREA)
  • General Physics & Mathematics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Health & Medical Sciences (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Optical Communication System (AREA)

Abstract

The invention provides a new signal modulation format identification method based on a neural network, which comprises the following steps: performing feature enhancement on the signal constellation diagram based on a normalized density algorithm; performing data enhancement on the signal constellation based on training to generate a countermeasure network; modulation format identification is performed by fine tuning the discrimination network. Meanwhile, a signal modulation format recognition system is also provided, which comprises: an image feature enhancing unit; the device comprises a data enhancement unit and an identification processing unit. The method and the system improve the identification accuracy of the modulation format of the optical signal under the condition of a low signal-to-noise ratio channel, can improve the reliability of a communication system, and have important application prospects in the fields of communication, information transmission processing and the like.

Description

New signal modulation format identification method based on neural network
Technical Field
The invention relates to the technical field of optical fiber communication, in particular to a new method and a system for identifying a signal modulation format based on a neural network.
Background
Under the background of the current information technology, the demand of people for communication equipment is higher and higher due to the high-speed development of information technology, and the amount of various information to be processed by a communication system is increased explosively. In this process, the problem faced by the communication system is the tight communication bandwidth and how to reduce the error rate of signal demodulation. In a conventional communication system, a corresponding hardware circuit is theoretically built for generating and correctly receiving a signal, and a signal receiver generally needs to be capable of receiving signals of dozens of different modulation formats, so that the hardware circuit required to be designed in the receiver is very complicated, and the volume of a communication device is greatly increased due to the limitation of communication materials.
However, nowadays, people have increasingly large requirements on the amount of information which can be transmitted in communication signals and the efficiency of a transmitter and a receiver for transmitting and receiving signals, adaptive modulation and coding solutions are widely proposed, the problems of the existing communication systems cannot be solved by adopting the traditional signal processing mode, and the key for solving the problems lies in exploring a special signal demodulation mode, so that the lower demodulation error rate can be ensured, the original carrier signals do not need to be recovered, and various modulation signals are processed and accurately identified by adopting an intelligent technology.
Disclosure of Invention
In view of this, the present invention provides a new method and system for identifying a signal modulation format based on a neural network, which on one hand improves the accuracy rate of identifying the signal modulation format and on the other hand also improves the reliability of a communication system.
In order to achieve the purpose, the invention provides the following technical scheme: the new method for identifying the signal modulation format based on the neural network is provided, and comprises the following steps: the method comprises the steps of feature enhancement processing of a constellation image based on normalized density, data enhancement processing based on generation of a countermeasure network and signal modulation format identification processing based on generation of the countermeasure network-identification network.
Preferably, the feature enhancement processing of the constellation image based on the normalized density includes:
calculating the normalized density of each signal sampling point; and coloring the constellation diagram according to the normalized density to generate a three-channel constellation diagram, namely the normalized density constellation diagram.
Preferably, the method for calculating the normalized density of each sampling point comprises the following steps:
the normalized density is expressed in a probabilistic form, and a parameter is determined: the number N of sampling points, a square with the side length r is selected by taking the current signal point as the center, the ratio of the number of other signal points to the number N of the sampling points of the received signal in the square is the normalized density value of the point, and the normalized density beta (i) of the ith point is shown as the formula (1):
Figure BDA0002948972400000021
wherein N is the number of sampling points, the side length r of the normalized density drawing square area, x (k) represents the abscissa value of the kth point, y (k) represents the ordinate value of the kth point, and epsilon (x) represents a step function.
Preferably, the data enhancement processing based on the generation countermeasure network is a countermeasure optimization process for the generation network (g (x)) and the authentication network (d (x)).
Preferably, the optimization process of the generation network (g (x)) is:
let the real picture sample be x ∈ RnThe noise z ∈ RmThe sample obtained after the network generation is
Figure BDA0002948972400000022
Is shown as
Figure BDA0002948972400000023
Authentication network output y ∈ [0, 1 ]]The result of the generated sample and the real sample obtained by the authentication network is
Figure BDA0002948972400000024
And X ═ Df(x，θf) The construction loss function is shown in equation (2):
Figure BDA0002948972400000031
wherein P (x) and
Figure BDA0002948972400000032
respectively representing the distribution of the real sample and the generated sample, and optimizing the generated network by combining a back propagation method so as to generate a forged sample which can deceive the authentication network, wherein the specific optimization method is to solve a maximum value, and is as follows:
Figure BDA0002948972400000033
the larger the value of the above equation (3), the better the effect of generating network-generated data.
Preferably, the optimization process of the authentication network (d (x)) is:
let pdata(x) And pg(x) The true constellation sample and the generated constellation sample are respectively input into the identification network, and the loss function can be expressed as formula (4):
Figure BDA0002948972400000034
the patent finally realizes the identification of the modulation format, so that the condition information C for controlling the generation of the network generated picture type is added, and the final form of the loss function is as shown in the formula (5):
Figure BDA0002948972400000035
the optimization direction is the maximum loss function, as in equation (6):
maxD{V(D，G)} (6)
the optimal solution can be obtained as formula (7):
Figure BDA0002948972400000041
when p isdata(x)＝pg(x) Then, a global optimal solution-log 4 can be obtained, and the network achieves Nash balance when the global optimal solution is obtained.
Preferably, the signal modulation format recognition processing based on the generation countermeasure network-identification network is to perform modulation format recognition by finely adjusting the identification network and classifying the extracted deep features with the aid of the full connection layer.
In addition, a signal modulation format recognition system is provided, which is characterized by comprising: the image characteristic enhancement unit is used for converting the common constellation map into a normalized density constellation image; the data enhancement unit is used for training and generating a confrontation network and a discrimination network; and the identification processing unit is used for carrying out signal modulation format identification processing on the identification network fine tuning and training.
Preferably, the feature enhancement unit is generated by Matlab calculation.
Preferably, the generation of the countermeasure network and the training discrimination network are implemented by invoking a computer GPU operation using Python.
Compared with the prior art, the invention has the following beneficial effects:
the novel method and the system for identifying the signal modulation format based on the neural network combine a normalized density algorithm and generate an antagonistic network, extract and highlight deep features of an original single-channel constellation image through feature enhancement and data enhancement operations, train an identification network with higher identification efficiency, improve the identification performance of the whole system, improve the reliability of a communication system, and have important application prospects in the fields of communication, information transmission processing and the like.
Drawings
In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the drawings needed to be used in the embodiments will be briefly described below, and it is obvious that the drawings in the following description are only some embodiments of the present invention, and it is obvious for those skilled in the art that other drawings can be obtained according to these drawings without creative efforts.
FIG. 1 is a flow chart of a new method for identifying a signal modulation format based on a neural network;
FIG. 2 is a diagram of a generating network (left) and an authenticating network (right) structure;
FIG. 3 is a diagram of a trimmed authentication network structure;
fig. 4 a single channel constellation;
FIG. 5 is a normalized density constellation;
fig. 6 generates a false constellation against network generation.
The specific implementation mode is as follows:
in order to make the technical solutions and advantages of the present invention clearer, the technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the drawings in the embodiments of the present application, and it is obvious that the described embodiments are some, but not all embodiments of the present application. All other embodiments, which can be derived by a person skilled in the art from the embodiments given herein without making any creative effort, shall fall within the protection scope of the present invention.
The embodiment provides a new signal modulation format recognition method based on a neural network, which adopts the technical scheme that firstly, three-dimensional characteristics of a constellation image are introduced through a normalized density algorithm, an original single-channel constellation image (figure 4) is converted into a three-channel constellation image, namely, a normalized density constellation image (figure 5), so that high-dimensional data information of the single-channel constellation image is increased, and extra information gain is introduced, thereby achieving the purpose of characteristic enhancement; then, performing data enhancement processing on the normalized density constellation image through a generated countermeasure network (GAN) to achieve the purpose of expanding deep feature information of the image; the characteristics of the constellation diagram data are more prominent through the characteristic enhancement and the data enhancement steps, the capability of extracting complex characteristics of the constellation diagram by the identification network is stronger, so that the identification accuracy of the signal modulation format is improved, the identification network for generating the countermeasure network is directly fine-tuned and trained, part of network structures are changed to enable the identification network to adapt to the current classification problem, and the modulation format is accurately identified. See the flow chart of a new method for identifying signal modulation formats based on neural networks (fig. 1).
The following describes the technical scheme in detail by taking a 16QAM constellation as an example:
first, a parameter is determined: the number of sampling points is N, and the sampling signals are converted into the form of a constellation diagram, namely an original single-channel constellation diagram. Preferably, N65536, that is, 65536 signal points are sampled, and the sampled signal is converted into a constellation diagram, that is, an original single-channel constellation diagram (fig. 4), and 65536 constellation points are mapped to a constellation diagram with a pixel size of 656 x 656.
Then, the normalized density is calculated, the three-dimensional characteristics of the constellation image are introduced through a normalized density algorithm, and the original single-channel constellation image is converted into a three-channel constellation image, namely the normalized density constellation image. The normalized density is expressed in a probability form, a square with the side length r is selected by taking the current signal point as the center, and the ratio of the number of other signal points to the number N of the sampling points of the received signal in the square is the normalized density value of the point. According to the normalized density algorithm, the square side length parameter r is preferably 2 in this embodiment, and then the normalized density β of each constellation point is calculated. Equation (1) below is the normalized density β (i) of the ith point, where N is the number of sampling points, x (k) represents the abscissa value of the kth point, y (k) represents the ordinate value of the kth point, and ∈ (x) represents a step function:
Figure BDA0002948972400000061
and generating a normalized density constellation map (figure 5) according to the normalized density values corresponding to colors with different brightness, wherein the high-density points use a high bright color system, and the low-density points use a low bright color system. Each pixel point in the colored normalized density constellation diagram is not independent and the like, but is integrated with high-dimensional image characteristics, so that more deep information can be condensed, and the primary characteristic enhancement of the constellation diagram is realized.
The normalized density constellation image pair is then used for data-enhanced training of the generation of the countermeasure network (fig. 2). The generation of the countermeasure network is an unsupervised learning model aiming at complex distribution, and the training process is the process of countervailing and optimizing the generation network and the identification network.
Preferably, the optimization process of the generation network (g (x)) is:
let the real picture sample be x ∈ RnThe noise z ∈ RmThe sample obtained after the network generation is
Figure BDA0002948972400000071
Is shown as
Figure BDA0002948972400000072
Authentication network output y ∈ [0, 1 ]]The real sample and the generated sample are obtained as a result through the authentication network
Figure BDA0002948972400000073
And X ═ Df(x，θf) Construction ofThe loss function is shown in equation (2):
Figure BDA0002948972400000074
wherein P (x) and
Figure BDA0002948972400000075
respectively representing the distribution of the real sample and the generated sample, and optimizing the generated network by combining a back propagation method so as to generate a forged sample which can deceive the authentication network, wherein the specific optimization method is to solve a maximum value, and is as follows:
Figure BDA0002948972400000076
the larger the value of the above equation (3), the better the effect of generating network-generated data.
Preferably, the optimization process of the authentication network (d (x)) is:
let pdata(x) And pg(x) The true constellation sample and the generated constellation sample are respectively input into the identification network, and the loss function can be expressed as formula (4):
Figure BDA0002948972400000077
the patent finally realizes the identification of the modulation format, so that the condition information C for controlling the generation of the network generated picture type is added, and the final form of the loss function is as shown in the formula (5):
Figure BDA0002948972400000078
Figure BDA0002948972400000082
the optimization direction is the maximum loss function, as in equation (6):
maxD{V(D，G)} (6)
the optimal solution can be obtained as formula (7):
Figure BDA0002948972400000081
when p isdata(x)＝pg(x) Then, a global optimal solution-log 4 can be obtained, and the network achieves Nash balance when the global optimal solution is obtained.
Preferably, in this embodiment, the data enhancement and training discrimination network is implemented by using Python to call a computer GPU for operation. The generation network generates a false target image according to the random input z, the false target image and the real image are mixed and input into the identification network to carry out true and false identification on the false target image, the false target image gradually becomes real in the countermeasure process, and the identification network training data set is invisibly expanded, so that the identification accuracy is improved. In order to make the network more easily converge, the input constellation image is compressed into 64 × 64 size, the number p of generated network deconvolution layers in the generation countermeasure network is 4, the number q of network convolution layers is identified to be 4, and the generation countermeasure network is trained according to the principle of maximum-minimum loss function to make the network reach nash balance. At this time, the generation network can generate a high-quality false constellation map (fig. 6) distributed in the same way as the original training data set, and in fact, the visual false constellation map is not needed in the training process, and the visual false constellation map provided herein is used for showing the training effect of the generation countermeasure network; the trained discrimination network has strong recognition and classification capability.
And finally, fine-tuning the identification network (fig. 3) and continuing training, changing a part of network structures to enable the identification network to adapt to the current classification problem, adjusting the dimension of the output layer of the full connection layer to be 8, corresponding to 8 modulation formats, and finishing training after the identification network converges. Finally, the system completes the task of identifying the modulation format, and the identification network can realize BPSK, QPSK, 8PSK, 8QAM, 16QAM, 32QAM, 64QAM and 128QAM under the condition of 0db-20db optical signal-to-noise ratio channel by combining a normalized density algorithm and generating an antagonistic network, and the identification accuracy of 8 modulation formats is integrated by 100%.
The present embodiment further provides a system for identifying a signal modulation format, including: the image characteristic enhancement unit is used for converting the common constellation map into a normalized density constellation image; the data enhancement unit is used for generating a training of the countermeasure network and a training identification network; and the identification processing unit is used for carrying out signal modulation format identification processing on the fine adjustment of the identification network.
Preferably, the feature enhancement unit is generated by Matlab calculation.
Preferably, the training of the generation countermeasure network and the identification network is implemented using Python calls to computer GPU operations.
Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and alterations can be made in these embodiments without departing from the principles and spirit of the invention, the scope of which is defined in the appended claims and their equivalents.

Claims (10)

1. A new signal modulation format identification method based on a neural network is characterized by comprising the following steps: the method comprises the steps of feature enhancement processing of a constellation image based on normalized density, data enhancement processing based on generation of a countermeasure network and signal modulation format identification processing based on generation of the countermeasure network-identification network.
2. The new neural network-based signal modulation format recognition method of claim 1, wherein the feature enhancement processing of the normalized density-based constellation image comprises:
calculating the normalized density of each signal sampling point; and coloring the constellation diagram according to the normalized density of the constellation diagram to generate the constellation diagram with the normalized density.
3. The new neural network-based signal modulation format recognition method of claim 1 or 2, wherein the normalized density algorithm is:
the normalized density is expressed in a probabilistic form, and a parameter is determined: the number N of sampling points, a square with the side length r is selected by taking the current signal point as the center, the ratio of the number of other signal points to the number N of the sampling points of the received signal in the square is the normalized density value of the point, and the normalized density beta (i) of the ith point is shown as the formula (1):
Figure FDA0002948972390000011
wherein N is the number of sampling points, the side length r of the normalized density drawing square area, x (k) represents the abscissa value of the kth point, y (k) represents the ordinate value of the kth point, and epsilon (x) represents a step function.
4. The new neural network-based signal modulation format recognition method of claim 1, wherein the generation countermeasure network-based data enhancement processing is a countermeasure optimization process for the generation network (g (x)) and the discrimination network (d (x)).
5. The new neural network-based signal modulation format recognition method of claim 4, wherein the optimization process of the generation network (G (x)) is:
let the real picture sample be x ∈ RnThe noise z ∈ RmThe sample obtained after the network generation is
Figure FDA0002948972390000021
Is shown as
Figure FDA0002948972390000022
Authentication network output y ∈ [0, 1 ]]The results of the generated sample and the real sample obtained by the identification network are respectively
Figure FDA0002948972390000023
And X ═ Df(x，θf) The constructed loss function is shown in formula (2):
Figure FDA0002948972390000024
wherein P (x) and
Figure FDA0002948972390000025
respectively representing the distribution of the real sample and the generated sample, and optimizing the generated network by combining a back propagation method so as to generate a forged sample which can deceive the authentication network, wherein the specific optimization method is to solve a maximum value, and is as follows:
Figure FDA0002948972390000026
the larger the value of the above equation (3), the better the effect of generating network-generated data.
6. The new neural network-based signal modulation format recognition method of claim 4, wherein the optimization process of the discrimination network (D (x)) is:
let pdata(x) And pg(x) The true constellation sample and the generated constellation sample are respectively input into the identification network, and the loss function can be expressed as formula (4):
Figure FDA0002948972390000027
the patent finally realizes the identification of the modulation format, so that the condition information C for controlling the generation of the network generated picture type is added, and the final form of the loss function is as shown in the formula (5):
Figure FDA0002948972390000031
the optimization direction is the maximum loss function, as in equation (6):
maxD{V(D，G)} (6)
the optimal solution can be obtained as formula (7):
Figure FDA0002948972390000032
when p isdata(x)＝pg(x) Then, a global optimal solution-log 4 can be obtained, and the network achieves Nash balance when the global optimal solution is obtained.
7. The method as claimed in claim 1, wherein the signal modulation format recognition process based on the generation countermeasure network-identification network is a modulation format recognition process by fine-tuning the identification network and classifying the extracted deep features with full connectivity.
8. A signal modulation format identification system, comprising:
the characteristic enhancement unit is used for converting the common constellation diagram into a normalized density constellation image;
the data enhancement unit is used for training and generating a confrontation network and a discrimination network;
and the identification processing unit is used for carrying out signal modulation format identification processing on the identification network fine tuning and training.
9. The signal modulation format recognition system of claim 8, wherein the feature enhancement unit is generated by Matlab calculation.
10. The system according to claim 8, wherein the generation countermeasure network and the training discrimination network are implemented using Python calls to computer GPU operations.
CN202110203711.9A 2021-02-23 2021-02-23 New signal modulation format identification method based on neural network Pending CN112953871A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202110203711.9A CN112953871A (en) 2021-02-23 2021-02-23 New signal modulation format identification method based on neural network

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202110203711.9A CN112953871A (en) 2021-02-23 2021-02-23 New signal modulation format identification method based on neural network

Publications (1)

Publication Number Publication Date
CN112953871A true CN112953871A (en) 2021-06-11

Family

ID=76245744

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202110203711.9A Pending CN112953871A (en) 2021-02-23 2021-02-23 New signal modulation format identification method based on neural network

Country Status (1)

Country Link
CN (1) CN112953871A (en)

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113626960A (en) * 2021-07-29 2021-11-09 南京信息工程大学 Anti-mode coupling few-mode signal complex format analysis method and device
CN113723556A (en) * 2021-09-08 2021-11-30 中国人民解放军国防科技大学 Modulation mode identification method based on entropy weighting-multi-mode domain antagonistic neural network
CN113905383A (en) * 2021-08-26 2022-01-07 湖南艾科诺维科技有限公司 IFF signal identification method, device and medium based on radio frequency fingerprint
CN114050953A (en) * 2021-11-22 2022-02-15 常州北邮新一代信息技术研究院有限公司 Signal modulation format identification method and system for coherent optical communication system
CN114065823A (en) * 2021-12-02 2022-02-18 中国人民解放军国防科技大学 Modulation signal identification method and system based on sparse deep neural network
CN114338316A (en) * 2021-12-29 2022-04-12 北京邮电大学 Method for identifying standard modulation format and probability shaping modulation format
CN114978827A (en) * 2022-04-22 2022-08-30 深圳市人工智能与机器人研究院 Modulation identification method for correcting frequency offset based on constellation diagram phase anomaly ratio
CN115622852A (en) * 2022-10-21 2023-01-17 扬州大学 Modulation identification method based on constellation KD tree enhancement and neural network GSENet
CN118300942A (en) * 2024-06-05 2024-07-05 四川轻化工大学 Modulation format identification method based on polar constellation diagram characteristics

Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108092714A (en) * 2017-12-20 2018-05-29 中山大学 Support the ofdm system of color adjustment and the detection of CSK planispheres
CN111614398A (en) * 2020-05-12 2020-09-01 北京邮电大学 Method and device for identifying modulation format and signal-to-noise ratio based on XOR neural network

Patent Citations (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108092714A (en) * 2017-12-20 2018-05-29 中山大学 Support the ofdm system of color adjustment and the detection of CSK planispheres
CN111614398A (en) * 2020-05-12 2020-09-01 北京邮电大学 Method and device for identifying modulation format and signal-to-noise ratio based on XOR neural network

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
BIN TANG等: "Digital Signal Modulation Classification With Data Augmentation Using Generative Adversarial Nets in Cognitive Radio Networks", 《IEEE》 *
BIN TANG等: "Digital Signal Modulation Classification With Data Augmentation Using Generative Adversarial Nets in Cognitive Radio Networks", 《IEEE》, vol. 6, 18 April 2018 (2018-04-18), pages 1 - 11 *
张思成等: "基于轻量级深度神经网络的电磁信号调制识别技术", 《CNKI》 *
张思成等: "基于轻量级深度神经网络的电磁信号调制识别技术", 《CNKI》, vol. 41, no. 11, 30 November 2020 (2020-11-30), pages 1 - 2 *

Cited By (14)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113626960B (en) * 2021-07-29 2022-06-14 南京信息工程大学 Anti-mode coupling few-mode signal complex format analysis method and device
CN113626960A (en) * 2021-07-29 2021-11-09 南京信息工程大学 Anti-mode coupling few-mode signal complex format analysis method and device
CN113905383A (en) * 2021-08-26 2022-01-07 湖南艾科诺维科技有限公司 IFF signal identification method, device and medium based on radio frequency fingerprint
CN113905383B (en) * 2021-08-26 2024-02-06 湖南艾科诺维科技有限公司 IFF signal identification method, device and medium based on radio frequency fingerprint
CN113723556A (en) * 2021-09-08 2021-11-30 中国人民解放军国防科技大学 Modulation mode identification method based on entropy weighting-multi-mode domain antagonistic neural network
CN114050953A (en) * 2021-11-22 2022-02-15 常州北邮新一代信息技术研究院有限公司 Signal modulation format identification method and system for coherent optical communication system
CN114050953B (en) * 2021-11-22 2023-05-23 常州京信新一代信息技术研究院有限公司 Method and system for identifying signal modulation format of coherent optical communication system
CN114065823A (en) * 2021-12-02 2022-02-18 中国人民解放军国防科技大学 Modulation signal identification method and system based on sparse deep neural network
CN114338316A (en) * 2021-12-29 2022-04-12 北京邮电大学 Method for identifying standard modulation format and probability shaping modulation format
CN114338316B (en) * 2021-12-29 2024-01-05 北京邮电大学 Identification method for standard modulation format and probability shaping modulation format
CN114978827A (en) * 2022-04-22 2022-08-30 深圳市人工智能与机器人研究院 Modulation identification method for correcting frequency offset based on constellation diagram phase anomaly ratio
CN114978827B (en) * 2022-04-22 2024-05-10 深圳市人工智能与机器人研究院 Modulation identification method based on constellation diagram phase abnormal ratio correction frequency offset
CN115622852A (en) * 2022-10-21 2023-01-17 扬州大学 Modulation identification method based on constellation KD tree enhancement and neural network GSENet
CN118300942A (en) * 2024-06-05 2024-07-05 四川轻化工大学 Modulation format identification method based on polar constellation diagram characteristics

Similar Documents

Publication Publication Date Title
CN112953871A (en) New signal modulation format identification method based on neural network
Choqueuse et al. Blind modulation recognition for MIMO systems
WO2018176889A1 (en) Method for automatically identifying modulation mode for digital communication signal
CN108173599B (en) Stokes spatial coherent light modulation format identification method based on DENC L UE clustering
CN109802905A (en) Digital signal Automatic Modulation Recognition method based on CNN convolutional neural networks
CN112702294A (en) Modulation recognition method for multi-level feature extraction based on deep learning
CN110113288B (en) Design and demodulation method of OFDM demodulator based on machine learning
CN108965178B (en) Intelligent self-adaptive equalizer based on machine learning and equalization demodulation method
CN114567528B (en) Communication signal modulation mode open set recognition method and system based on deep learning
Huang et al. Hierarchical digital modulation classification using cascaded convolutional neural network
Subbarao et al. Automatic modulation recognition in cognitive radio receivers using multi-order cumulants and decision trees
CN110659684A (en) Convolutional neural network-based STBC signal identification method
CN114398931A (en) Modulation recognition method and system based on numerical characteristic and image characteristic fusion
CN112737992A (en) Underwater sound signal modulation mode self-adaptive in-class identification method
Sun et al. Automatic signal modulation recognition based on deep convolutional neural network
CN114943246A (en) Semi-supervised specific radiation source identification method based on bispectrum feature extraction
Zhang et al. Automatic space-time block code recognition using convolutional neural network with multi-delay features fusion
Ma et al. Deep learning based cognitive radio modulation parameter estimation
CN113239788A (en) Mask R-CNN-based wireless communication modulation mode identification method
CN116016071A (en) Modulation signal identification method based on double-flow fusion CNN-BiLSTM network
CN115913850A (en) Open set modulation identification method based on residual error network
Guo et al. Custom convolutional layer designs for CNN based automatic modulation classification solution
Huang et al. Modulation classification of MQAM signals based on gradient color constellation and deep learning
Zhang et al. Modulation format recognition based on CNN in satellite communication system
Choqueuse et al. Modulation recognition for MIMO communications

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
WD01 Invention patent application deemed withdrawn after publication

Application publication date: 20210611

WD01 Invention patent application deemed withdrawn after publication