CN114501451A - Physical layer channel security authentication method based on limited reorganization data enhancement - Google Patents
Physical layer channel security authentication method based on limited reorganization data enhancement Download PDFInfo
- Publication number
- CN114501451A CN114501451A CN202210305200.2A CN202210305200A CN114501451A CN 114501451 A CN114501451 A CN 114501451A CN 202210305200 A CN202210305200 A CN 202210305200A CN 114501451 A CN114501451 A CN 114501451A
- Authority
- CN
- China
- Prior art keywords
- channel
- matrix
- limited
- fingerprint
- initial
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Pending
Links
- 238000000034 method Methods 0.000 title claims abstract description 35
- 230000008521 reorganization Effects 0.000 title claims abstract description 12
- 238000012549 training Methods 0.000 claims abstract description 51
- 238000012545 processing Methods 0.000 claims abstract description 12
- 230000006798 recombination Effects 0.000 claims abstract description 10
- 238000005215 recombination Methods 0.000 claims abstract description 10
- 238000007781 pre-processing Methods 0.000 claims abstract 2
- 239000011159 matrix material Substances 0.000 claims description 63
- 238000004891 communication Methods 0.000 claims description 16
- 238000013528 artificial neural network Methods 0.000 claims description 10
- 238000010801 machine learning Methods 0.000 claims description 7
- 238000004364 calculation method Methods 0.000 claims description 6
- 239000013598 vector Substances 0.000 claims description 6
- 238000013527 convolutional neural network Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 238000010276 construction Methods 0.000 claims description 2
- 125000004122 cyclic group Chemical group 0.000 claims description 2
- 238000007477 logistic regression Methods 0.000 claims description 2
- 238000012706 support-vector machine Methods 0.000 claims description 2
- 238000011176 pooling Methods 0.000 description 5
- 230000004913 activation Effects 0.000 description 2
- 238000010586 diagram Methods 0.000 description 2
- 230000006870 function Effects 0.000 description 2
- 238000012986 modification Methods 0.000 description 2
- 230000004048 modification Effects 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 230000005540 biological transmission Effects 0.000 description 1
- 238000013135 deep learning Methods 0.000 description 1
- 238000013136 deep learning model Methods 0.000 description 1
- 230000007547 defect Effects 0.000 description 1
Images
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04W—WIRELESS COMMUNICATION NETWORKS
- H04W12/00—Security arrangements; Authentication; Protecting privacy or anonymity
- H04W12/06—Authentication
-
- G—PHYSICS
- G06—COMPUTING; CALCULATING OR COUNTING
- G06F—ELECTRIC DIGITAL DATA PROCESSING
- G06F17/00—Digital computing or data processing equipment or methods, specially adapted for specific functions
- G06F17/10—Complex mathematical operations
- G06F17/16—Matrix or vector computation, e.g. matrix-matrix or matrix-vector multiplication, matrix factorization
-
- 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)
- Physics & Mathematics (AREA)
- Theoretical Computer Science (AREA)
- Data Mining & Analysis (AREA)
- General Physics & Mathematics (AREA)
- Mathematical Physics (AREA)
- General Engineering & Computer Science (AREA)
- Software Systems (AREA)
- Life Sciences & Earth Sciences (AREA)
- Evolutionary Computation (AREA)
- Pure & Applied Mathematics (AREA)
- Mathematical Optimization (AREA)
- Computing Systems (AREA)
- Mathematical Analysis (AREA)
- Computational Mathematics (AREA)
- Artificial Intelligence (AREA)
- Computer Vision & Pattern Recognition (AREA)
- Biophysics (AREA)
- Bioinformatics & Cheminformatics (AREA)
- Bioinformatics & Computational Biology (AREA)
- Computational Linguistics (AREA)
- Evolutionary Biology (AREA)
- Biomedical Technology (AREA)
- Health & Medical Sciences (AREA)
- Molecular Biology (AREA)
- General Health & Medical Sciences (AREA)
- Algebra (AREA)
- Databases & Information Systems (AREA)
- Computer Security & Cryptography (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses a physical layer channel security authentication method based on limited reorganization data enhancement, which comprises the following steps: collecting an initial channel fingerprint; preprocessing the initial channel fingerprint to obtain an initial training sample set; performing limited reorganization data enhancement processing on the initial training sample set; training the network model by using the training data set after data enhancement, thereby obtaining a physical layer channel fingerprint authentication model; and carrying out security authentication on the unknown channel fingerprint. Under the condition of acquiring less original channel fingerprint data, the method quickly generates more training data samples by a limited recombination data enhancement method, and improves the training speed of the network model.
Description
Technical Field
The invention belongs to the field of transmission data packet security identification, and particularly relates to a physical layer channel security authentication method based on limited recombinant data enhancement.
Background
The physical layer channel authentication is to authenticate a data packet through unique channel fingerprint information of a wireless channel, has the characteristics of light weight and high reliability, and is very suitable for application scenes of future large-scale machine communication. Compared with the traditional physical layer channel authentication method based on the threshold value, the channel authentication method based on the machine learning can effectively improve the authentication success rate of the channel fingerprint, especially under the support of edge calculation, edge training can be realized, and a terminal or a node can hardly perform any calculation.
However, the channel authentication method based on machine learning requires a large amount of data to train the authentication model, however, in some wireless communication application scenarios with limited resources or sensitive time delay, there is usually not enough time to collect sufficient channel fingerprint training data set samples, thereby affecting the training and authentication performance of the network model.
Disclosure of Invention
The invention aims to overcome the defects of the prior art and provide a physical layer channel security authentication method based on limited reorganization data enhancement, under the condition of obtaining less original channel training samples, more training data samples are rapidly generated through the limited reorganization data enhancement method so as to accelerate the training speed of a network model.
The purpose of the invention is realized by the following technical scheme: a physical layer channel security authentication method based on limited reorganization data enhancement comprises the following steps:
s1, collecting initial channel fingerprint matrixes of K communication nodes
Wherein K is 1, 2, …, K, t is 1, 2, …, psik，ψkInitial channel fingerprint matrix representing a collecting communication node k
The maximum value of the number of the first and second,
is a complex matrix with dimension m × N, m ═ Ns×Nt，NsDenotes the number of subcarriers, NtRepresenting the number of transmit antennas, n representing the number of receive antennas;
s2, fingerprint matrix of initial channel
Carrying out pretreatment:
s201, according to the initial channel fingerprint matrix
Generating an initial real channel fingerprint matrix
Wherein t is 1, 2, …, psik，
Representing initial channel fingerprint matrix
The matrix of the real part of (a),
representing initial channel fingerprint matrix
The imaginary matrix of (a);
s202, when K is equal to each of 1, 2, …, and K, repeatedly executing step S201, taking the result obtained under each value of K as a row, and finally obtaining the following matrix:
s3, giving each initial real number channel fingerprint matrix
A corresponding label IkThen the initial training sample set DtrainComprises the following steps:
Dtrain＝{Xtrain，Ytrain} (2)
wherein, Xtrain，YtrainAll represent intermediate variables of the calculation process, and have no specific meaning;
s4, for the initial training sample set DtrainPerforming limited recombination data enhancement processing;
s401, real channel fingerprint matrix
Dividing into beta blocks according to rows or columns, dividing into blocks according to rows as shown in formula (5), and dividing into blocks according to columns as shown in formula (6):
wherein,
all represent row vectors of dimension 1 × 2n [. ]]' denotes the transpose of the matrix;
each represents a column vector of dimension m × 1; beta represents the number of original adjacent channel matrixes participating in the construction of the new channel matrix, is a positive integer, and has to divide m or 2n by 1<β≤ψk；
Then, adjacent beta real channel fingerprint matrixes
The block elements are recombined to generate a new channel matrix, as shown in equation (7) or equation (8):
wherein,
representing a channel matrix divided by beta blocks
In which 1 block is not selected repeatedly,
representing a channel matrix divided by beta blocks
1 block is selected without repetition in the other same way; upsilon is an index of the newly generated channel samples and 1, 2, …, (ψ)k-β+1)×ββ；
At this time, the finite recombination real number channel fingerprint matrix for the kth communication node of the training network is obtained as follows:
the limited reassembly label matrix for the kth communication node is:
therefore, the new training sample set of the kth communication node obtained after the processing by the limited reassembly data enhancement method is:
s402, when K is equal to each of 1, 2, …, and K, repeating step S401, thereby performing the training sample set DtrainPerforming limited recombination data enhancement processing to obtain a new training data sample set
S5, utilizing a new training data sample set
And training the network model to obtain the trained physical layer channel fingerprint authentication model.
Preferably, the network model in step S5 includes, but is not limited to, a machine learning algorithm or a Neural network algorithm, wherein the machine learning algorithm includes a support vector machine algorithm, a K-Nearest Neighbor (KNN) algorithm, a classification tree algorithm, and the like, and the Neural network algorithm includes a logistic regression algorithm, a shallow Neural network algorithm, a Deep Neural Network (DNN) algorithm, a Global-Connected convolutional Neural network (GC-Net) algorithm, a cyclic Neural network algorithm, and the like.
And S6, inputting the unknown channel fingerprint into the physical layer channel fingerprint authentication model in the step S5, and finishing authentication and identification of the unknown channel fingerprint according to an output result.
The invention has the beneficial effects that: according to the method, a limited recombination data enhancement method is adopted, a small amount of initially acquired channel fingerprint samples are subjected to data enhancement to generate more training data samples, the training data sample set subjected to data enhancement is used for training the network model, the training speed of the network model can be increased, the problem that the channel authentication performance based on machine learning is influenced due to insufficient training data samples is solved, and the method is very suitable for an application scene of improving the security authentication based on deep learning under 5G edge calculation.
Drawings
Fig. 1 is a flowchart illustrating channel fingerprint authentication according to an embodiment of the present invention.
Fig. 2 is a diagram illustrating a success rate of channel fingerprint authentication in an embodiment.
Detailed Description
The technical solutions of the present invention are further described in detail below with reference to the accompanying drawings, but the scope of the present invention is not limited to the following.
As shown in fig. 1, a physical layer channel security authentication method based on limited reassembly data enhancement includes the following steps:
s1, in the embodiment, the number N of receiving antennas is 8, and the number N of transmitting antennastIs 1, the number of subcarriers NsIs 128; 15 communication nodes acquire 100 initial channel fingerprint matrixes
Wherein
t＝1、2、…、100，k＝1、2、…、15。
S2, fingerprint matrix of initial channel
Carrying out pretreatment:
s201, according to the initial channel fingerprint matrix
Generating an initial real channel fingerprint matrix
Wherein,
representing initial channel fingerprint matrix
The matrix of the real part of (a),
representing initial channel fingerprint matrix
The imaginary matrix of (a);
then the initial real channel fingerprint matrix
The expression of (a) is:
wherein, amnMatrix representing real part
M-th row and n-th column of elements, bmnRepresenting the imaginary matrix
Row m and column n.
S202, when K is equal to each of 1, 2, …, and K, repeatedly executing step S201, taking the result obtained under each value of K as a row, and finally obtaining the following matrix:
s3, giving each initial real number channel fingerprint matrix
A corresponding label IkThen the initial training sample set DtrainComprises the following steps:
Dtrain＝{Xtrain，Ytrain} (2)
wherein, the label IkExpressed in one-hot code, Xtrain，YtrainAll represent intermediate variables of the calculation process, and have no specific meaning.
S4, adopting a limited reorganization data enhancement method to carry out initial training sample set DtrainCarrying out data enhancement processing;
s401, firstly, real channel fingerprint matrix
According to line orThe block division is performed by column average into 2 (i.e., β ═ 2) blocks, the block division is performed by row as shown in formula (5), and the block division is performed by column as shown in formula (6):
wherein,
all represent row vectors of dimension 1 × 16 [. ]]' denotes the transpose of the matrix,
and
respectively representing real channel fingerprint matrices
The row of the first block element matrix is divided into a 1 st block element matrix and the row of the second block element matrix is divided into a 2 nd block element matrix;
each representing a column vector of dimension 128 x 1,
and
respectively representing real channel fingerprint matrices
Is divided into a 1 st block element matrix and a 2 nd block element matrix;
then, adjacent beta is 2 real channel fingerprint matrixes
The block elements are recombined to generate a new channel matrix, as shown in equation (7) or equation (8):
wherein τ is 1, 2, …, 99.
The newly generated channel fingerprint matrix is as follows:
at this time, the finite recombination real number channel fingerprint matrix for the kth communication node of the training network is obtained as follows:
the limited reassembly label matrix for the kth communication node is:
therefore, the new training sample set of the kth communication node obtained after the processing by the limited reassembly data enhancement method is:
s402, when K is equal to each of 1, 2, …, and K, repeating step S401, thereby performing the training sample set DtrainPerforming limited recombination data enhancement processing to obtain a new training data sample set
S5, utilizing a new training data sample set
And training the network model to obtain the trained physical layer channel fingerprint authentication model.
And S6, inputting the unknown channel fingerprint into the physical layer channel fingerprint authentication model in the step S5, and finishing authentication and identification of the unknown channel fingerprint according to an output result.
In this embodiment, a fully-connected convolutional neural network (GC-Net) is selected for training and classification, which contains convolutional layers, pooling layers, activation layers, and finally fully-connected layers.
In this example, GC-Net is set to three layers, the convolution kernel size of each layer is set to 3 × 3, the step size is 1, the number of convolution kernels is 64, the pooling layer selects the maximum pooling, the pooling filter is 2 × 2, the pooling step size is 2, and the activation function selects the ELU function.
Fig. 2 is a diagram of success rate of channel fingerprint authentication in this embodiment, as shown in the figure, after multiple iterations, the success rate of channel authentication is converged, and the convergence rate of the channel authentication method subjected to the limited reassembly data enhancement processing is faster, that is, the number of iterations required is less than that of the authentication method without data enhancement. Therefore, the limited reorganization data enhancement method can quickly generate more training samples and accelerate the training speed of the deep learning model.
In summary, the invention performs data enhancement processing on a small amount of initially acquired channel fingerprint samples to generate more training samples by a limited recombination data enhancement method, and trains a network model by using a training sample set after data enhancement, so that not only more training samples can be generated, but also the training speed of the network model can be accelerated.
The foregoing is a preferred embodiment of the present invention, it is to be understood that the invention is not limited to the form disclosed herein, but is not to be construed as excluding other embodiments, and is capable of other combinations, modifications, and environments and is capable of changes within the scope of the inventive concept as expressed herein, commensurate with the above teachings, or the skill or knowledge of the relevant art. And that modifications and variations may be effected by those skilled in the art without departing from the spirit and scope of the invention as defined by the appended claims.
Claims (7)
1. A physical layer channel security authentication method based on limited reorganization data enhancement is characterized in that: the method comprises the following substeps:
s1, collecting an initial channel fingerprint;
s2, preprocessing the initial channel fingerprint to generate an initial real channel fingerprint matrix;
s3, giving a label to obtain an initial training sample set;
s4, performing limited reorganization data enhancement processing on the initial training sample set;
s5, training a network model by using the training data set after data enhancement, thereby obtaining a physical layer channel fingerprint authentication model;
and S6, carrying out security authentication on the unknown channel fingerprint.
2. The physical layer channel security authentication method based on limited reassembly data enhancement as claimed in claim 1, wherein: the step S1 includes:
collecting initial channel fingerprint matrix of K communication nodes
Wherein K is 1, 2, …, K, t is 1, 2, …, psik，ψkInitial channel fingerprint matrix representing a collecting communication node k
The maximum value of the number of the first and second,
is a complex matrix with dimension m × N, m ═ Ns×Nt，NsDenotes the number of subcarriers, NtRepresenting the number of transmit antennas and n representing the number of receive antennas.
3. The physical layer channel security authentication method based on limited reassembly data enhancement as claimed in claim 1, wherein: the step S2 includes:
s201, according to the initial channel fingerprint matrix
Generating an initial real channel fingerprint matrix
Wherein t is 1, 2, …, psik，
Representing initial channel fingerprint matrix
The matrix of the real part of (a),
representing initial channel fingerprint matrix
The imaginary matrix of (a);
s202, when K is equal to each of 1, 2, …, and K, repeatedly executing step S201, taking the result obtained under each value of K as a row, and finally obtaining the following matrix:
4. the physical layer channel security authentication method based on limited reassembly data enhancement as claimed in claim 1, wherein: the step S3 includes:
giving each initial real channel fingerprint matrix
A corresponding label IkThen the initial training sample set DtrainComprises the following steps:
Dtrain＝{Xtrain，Ytrain} (2)
wherein, Xtrain，YtrainBoth represent intermediate variables of the calculation process.
5. The physical layer channel security authentication method based on limited reassembly data enhancement as claimed in claim 1, wherein: the step S4 includes:
s401, real channel fingerprint matrix
Dividing into beta blocks according to rows or columns, dividing into blocks according to rows as shown in formula (5), and dividing into blocks according to columns as shown in formula (6):
wherein,
all represent row vectors of dimension 1 × 2n [. ]]' denotes the transpose of the matrix;
each represents a column vector of dimension m × 1; beta represents the number of original adjacent channel matrixes participating in the construction of the new channel matrix, is a positive integer, and has to divide m or 2n by 1 < beta ≦ ψk；
Then, adjacent beta real channel fingerprint matrixes
The block elements are recombined to generate a new channel matrix, as shown in equation (7) or equation (8):
wherein,
representing a channel matrix from divided beta blocks
In the selection of 1 block is not repeated,
representing a channel matrix divided by beta blocks
1 block is selected without repetition in the other same way; upsilon is an index of the newly generated channel samples and 1, 2, …, (ψ)k-β+1)×ββ；
At this time, the finite recombination real number channel fingerprint matrix for the kth communication node of the training network is obtained as follows:
the limited reassembly label matrix for the kth communication node is:
therefore, the new training sample set of the kth communication node obtained after the processing by the limited reassembly data enhancement method is:
s402, when K is equal to each of 1, 2, …, and K, repeating step S401, thereby performing the training sample set DtrainPerforming limited recombination data enhancement processing to obtain a new training data sample set
6. The physical layer channel security authentication method based on limited reassembly data enhancement as claimed in claim 1, wherein: the step S5 includes:
utilizing a new training data sample set
Training the network model to obtain a trained physical layer channel fingerprint authentication model;
the network model comprises but is not limited to a network model constructed by a machine learning algorithm or a neural network algorithm, wherein the machine learning algorithm comprises a support vector machine algorithm, a K nearest neighbor algorithm and a classification tree algorithm, and the neural network algorithm comprises a logistic regression algorithm, a shallow layer neural network algorithm, a deep neural network algorithm, a full-connection convolutional neural network algorithm and a cyclic neural network algorithm.
7. The physical layer channel security authentication method based on limited reorganization data enhancement of claim 1, wherein: in step S6, the unknown channel fingerprint is input into the physical layer channel fingerprint authentication model in step S5, and authentication and identification of the unknown channel fingerprint are completed according to the output result.
Applications Claiming Priority (2)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN2022101013342 | 2022-01-27 | ||
| CN202210101334 | 2022-01-27 |
Publications (1)
| Publication Number | Publication Date |
|---|---|
| CN114501451A true CN114501451A (en) | 2022-05-13 |
Family
ID=81488536
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210305200.2A Pending CN114501451A (en) | 2022-01-27 | 2022-03-25 | Physical layer channel security authentication method based on limited reorganization data enhancement |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114501451A (en) |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105635125A (en) * | 2015-12-25 | 2016-06-01 | 电子科技大学 | Physical layer combined authentication method based on RF fingerprint and channel information |
| US20190260739A1 (en) * | 2016-10-19 | 2019-08-22 | Politecnico Di Torino | Device and Methods for Authenticating a User Equipment |
| CN111541632A (en) * | 2020-04-20 | 2020-08-14 | 四川农业大学 | Physical layer authentication method based on principal component analysis and residual error network |
| CN113704737A (en) * | 2021-07-26 | 2021-11-26 | 西安交通大学 | Small sample physical layer equipment authentication method, system, terminal and storage medium |
| CN113784349A (en) * | 2021-11-09 | 2021-12-10 | 电子科技大学 | Method for improving safety certification based on deep learning under 5G edge calculation |
-
2022
- 2022-03-25 CN CN202210305200.2A patent/CN114501451A/en active Pending
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN105635125A (en) * | 2015-12-25 | 2016-06-01 | 电子科技大学 | Physical layer combined authentication method based on RF fingerprint and channel information |
| US20190260739A1 (en) * | 2016-10-19 | 2019-08-22 | Politecnico Di Torino | Device and Methods for Authenticating a User Equipment |
| CN111541632A (en) * | 2020-04-20 | 2020-08-14 | 四川农业大学 | Physical layer authentication method based on principal component analysis and residual error network |
| CN113704737A (en) * | 2021-07-26 | 2021-11-26 | 西安交通大学 | Small sample physical layer equipment authentication method, system, terminal and storage medium |
| CN113784349A (en) * | 2021-11-09 | 2021-12-10 | 电子科技大学 | Method for improving safety certification based on deep learning under 5G edge calculation |
Non-Patent Citations (3)
| Title |
|---|
| SONGLIN CHEN 等: "Physical-Layer Channel Authentication for 5G via Machine Learning Algorithm", WIRELESS COMMUNICATIONS AND MOBILE COMPUTING, pages 1 - 5 * |
| YI CHEN 等: "On Physical-Layer Authentication via Online Transfer Learning", IEEE INTERNET OF THINGS JOURNAL, pages 1 - 6 * |
| 李古月;俞佳宝;胡爱群;: "基于设备与信道特征的物理层安全方法", 密码学报, no. 02 * |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Wen et al. | Deep learning for massive MIMO CSI feedback | |
| Li et al. | Robust automated VHF modulation recognition based on deep convolutional neural networks | |
| CN109727246B (en) | Comparative learning image quality evaluation method based on twin network | |
| Cai et al. | Attention model for massive MIMO CSI compression feedback and recovery | |
| CN110113288B (en) | Design and demodulation method of OFDM demodulator based on machine learning | |
| CN113222179A (en) | Federal learning model compression method based on model sparsification and weight quantization | |
| CN113472706A (en) | MIMO-OFDM system channel estimation method based on deep neural network | |
| CN111464465A (en) | Channel estimation method based on integrated neural network model | |
| CN112910811B (en) | Blind modulation identification method and device under unknown noise level condition based on joint learning | |
| CN111224905B (en) | Multi-user detection method based on convolution residual error network in large-scale Internet of things | |
| CN114268388B (en) | Channel estimation method based on improved GAN network in large-scale MIMO | |
| CN113965233A (en) | Multi-user broadband millimeter wave communication resource allocation method and system based on deep learning | |
| CN114117945B (en) | Deep learning cloud service QoS prediction method based on user-service interaction graph | |
| CN113784349A (en) | Method for improving safety certification based on deep learning under 5G edge calculation | |
| CN112686376A (en) | Node representation method based on timing diagram neural network and incremental learning method | |
| Chuan et al. | Uplink NOMA signal transmission with convolutional neural networks approach | |
| CN112153615A (en) | Deep learning-based user association method in multi-cell cellular D2D equipment | |
| Hussien | PRVNet: Variational autoencoders for massive MIMO CSI feedback | |
| CN111010222B (en) | Deep learning-based large-scale MIMO downlink user scheduling method | |
| CN115278709A (en) | Communication optimization method based on federal learning | |
| CN112836822A (en) | Federal learning strategy optimization method and device based on width learning | |
| CN114501451A (en) | Physical layer channel security authentication method based on limited reorganization data enhancement | |
| Jagatap et al. | Learning relu networks via alternating minimization | |
| Liang et al. | Wireless channel data augmentation for artificial intelligence of things in industrial environment using generative adversarial networks | |
| CN117095217A (en) | Multi-stage comparative knowledge distillation process |
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 |