CN112153615A - Deep learning-based user association method in multi-cell cellular D2D equipment - Google Patents

Deep learning-based user association method in multi-cell cellular D2D equipment Download PDF

Info

Publication number
CN112153615A
CN112153615A CN202010964458.4A CN202010964458A CN112153615A CN 112153615 A CN112153615 A CN 112153615A CN 202010964458 A CN202010964458 A CN 202010964458A CN 112153615 A CN112153615 A CN 112153615A
Authority
CN
China
Prior art keywords
neural network
cellular
user association
channel gain
base station
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Granted
Application number
CN202010964458.4A
Other languages
Chinese (zh)
Other versions
CN112153615B (en
Inventor
李君�
朱明浩
仲星
王秀敏
李正权
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.)
Ictehi Technology Development Co ltd
Binjiang College of Nanjing University of Information Engineering
Original Assignee
Binjiang College of Nanjing University of Information Engineering
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 Binjiang College of Nanjing University of Information Engineering filed Critical Binjiang College of Nanjing University of Information Engineering
Priority to CN202010964458.4A priority Critical patent/CN112153615B/en
Publication of CN112153615A publication Critical patent/CN112153615A/en
Application granted granted Critical
Publication of CN112153615B publication Critical patent/CN112153615B/en
Active legal-status Critical Current
Anticipated expiration legal-status Critical

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W4/00—Services specially adapted for wireless communication networks; Facilities therefor
    • H04W4/70—Services for machine-to-machine communication [M2M] or machine type communication [MTC]
    • 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
    • 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
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/391—Modelling the propagation channel
    • H04B17/3911—Fading models or fading generators
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/04—Arrangements for maintaining operational condition
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04W—WIRELESS COMMUNICATION NETWORKS
    • H04W24/00—Supervisory, monitoring or testing arrangements
    • H04W24/06—Testing, supervising or monitoring using simulated traffic
    • Y—GENERAL TAGGING OF NEW TECHNOLOGICAL DEVELOPMENTS; GENERAL TAGGING OF CROSS-SECTIONAL TECHNOLOGIES SPANNING OVER SEVERAL SECTIONS OF THE IPC; TECHNICAL SUBJECTS COVERED BY FORMER USPC CROSS-REFERENCE ART COLLECTIONS [XRACs] AND DIGESTS
    • Y02—TECHNOLOGIES OR APPLICATIONS FOR MITIGATION OR ADAPTATION AGAINST CLIMATE CHANGE
    • Y02D—CLIMATE CHANGE MITIGATION TECHNOLOGIES IN INFORMATION AND COMMUNICATION TECHNOLOGIES [ICT], I.E. INFORMATION AND COMMUNICATION TECHNOLOGIES AIMING AT THE REDUCTION OF THEIR OWN ENERGY USE
    • Y02D30/00—Reducing energy consumption in communication networks
    • Y02D30/70—Reducing energy consumption in communication networks in wireless communication networks

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Theoretical Computer Science (AREA)
  • Signal Processing (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Data Mining & Analysis (AREA)
  • Life Sciences & Earth Sciences (AREA)
  • Artificial Intelligence (AREA)
  • General Physics & Mathematics (AREA)
  • Evolutionary Computation (AREA)
  • General Engineering & Computer Science (AREA)
  • Computing Systems (AREA)
  • Biomedical Technology (AREA)
  • General Health & Medical Sciences (AREA)
  • Computational Linguistics (AREA)
  • Mathematical Physics (AREA)
  • Software Systems (AREA)
  • Biophysics (AREA)
  • Molecular Biology (AREA)
  • Health & Medical Sciences (AREA)
  • Bioinformatics & Cheminformatics (AREA)
  • Evolutionary Biology (AREA)
  • Computer Vision & Pattern Recognition (AREA)
  • Bioinformatics & Computational Biology (AREA)
  • Electromagnetism (AREA)
  • Mobile Radio Communication Systems (AREA)

Abstract

The invention discloses a user association method based on deep learning in multi-cell cellular D2D equipment, which comprises the steps of firstly, collecting channel gain information of terminal equipment in an environment; secondly, obtaining an optimal cellular user association strategy under corresponding channel gain samples by using an exhaustion method, and collecting a training data set; then, constructing a convolutional neural network framework and initializing neural network parameters; finally, the neural network is trained: inputting a training data set into a neural network, constructing MSE between the output of the neural network and a label as a loss function, and selecting a random gradient descent algorithm to update the weight of the neural network; and when the loss function is smaller than a preset value or reaches the iteration times, the training of the neural network is considered to be finished, and the neural network is stored. The invention overcomes the problem of interference of D2D equipment to cellular equipment in the environment, and the convolutional neural network approaches the traditional algorithm in a supervised learning mode to learn the mapping relation between the channel gain and the optimal user association strategy.

Description

Deep learning-based user association method in multi-cell cellular D2D equipment
Technical Field
The invention relates to a communication system physical layer technology, in particular to a wireless communication system resource allocation technology, and particularly relates to a deep learning-based user association method in multi-cell D2D equipment.
Background
Currently, with the gradual popularization of 5G communication technology and the diversification of communication network forms, resource allocation of communication networks becomes a hot issue of research organizations. Device-to-device (D2D) is a technology with improved spectral and energy efficiency, and because of these advantages, D2D has been reused as a key technology for 5G wireless networks, in D2D links, two neighboring devices can transmit data directly without the help of an access point. With the gradual advance of the base station densification technology, more and more base stations appear in people's lives, which also means that the base stations are closer to our distance, in a hybrid network of cellular devices and D2D devices, the simultaneous communication of the two devices tends to cause an interference problem, and how to reduce the interference from the D2D device through the appropriate adjustment of the association state between the base stations and the cellular users, so that the rate performance of the system is maximized to become a new research direction, which is also the core problem of the present invention.
The traditional user association algorithm in the multi-cell network environment is an exhaustive method, and generally, an objective function is determined first, then the performance of all possible strategies is compared through the possibility of exhaustively associating all strategies, and finally the strategy which leads to the optimal performance is selected as a final strategy. The optimal solution meeting the maximum performance of all users in the system can be found, but the algorithm has the typical defects of poor convergence performance, and very high calculation complexity caused by slow convergence, so that the calculation time of the exhaustive algorithm is exponentially increased in a large-scale network, particularly a communication network with a large number of terminal devices, online decision cannot be made, and the application field of the algorithm is limited.
At present, deep learning techniques have been widely studied in various fields such as image processing and speech recognition in recent years, using the function approximation characteristic of a neural network. Although training of the neural network may take some time, since the training process can be performed off-line, the computational complexity is small, and the neural network is very suitable for real-time operation. And due to the convenience and timeliness of channel sample collection in a wireless communication network, the application of deep learning in wireless communication is more and more advantageous. The neural network in deep learning can achieve good nonlinear approximation for the traditional algorithm, so that extensive research is carried out, and the currently commonly used neural networks comprise a Deep Neural Network (DNN), a Convolutional Neural Network (CNN) and a Recurrent Neural Network (RNN).
Much research has been focused on using neural networks to maximize or minimize the objective function through link scheduling or user association. For example, a deep neural network is used for realizing the maximization of the Spectral Efficiency (SE) or the maximization of the Energy Efficiency (EE) of a system through an approximation sub-gradient algorithm, or the maximization of the system and the rate (SR) is realized through an approximation iterative optimization algorithm, and in the aspect of a convolutional neural network, the local features are extracted through a convolutional filter, so that the maximization of the SE, the EE or the SR is realized in a supervision and learning mode. The invention learns the mapping relation between the channel gain information of the terminal equipment and the optimal cellular user association strategy through the CNN.
Disclosure of Invention
The purpose of the invention is as follows: the invention provides a user association method based on deep learning in multi-cell cellular D2D equipment, which overcomes the problem of interference of D2D equipment to cellular equipment in the environment, and learns the mapping relation between channel gain and an optimal user association strategy by approaching the traditional algorithm in a supervised learning mode through a convolutional neural network.
The technical scheme is as follows: the invention relates to a user association method based on deep learning in multi-cell cellular D2D equipment, which specifically comprises the following steps:
(1) collecting channel gain information of terminal equipment in an environment;
(2) optimal cellular user association strategy Z under corresponding channel gain samples obtained by applying exhaustion method*Collecting a training data set;
(3) determining the division ratio of the training set and the test set;
(4) constructing a convolutional neural network model, and initializing neural network parameters;
(5) inputting a training data set into a neural network, constructing MSE between the output of the neural network and a label as a loss function, and selecting a random gradient descent algorithm to update the weight of the neural network;
(6) when the loss function is smaller than a preset value or reaches the iteration times, the neural network training is considered to be finished, and the neural network is stored;
(7) the test phase tests performance with the test set as input data.
Further, the training data set in step (2) includes channel gain information samples and corresponding optimal user associated label data.
Further, the convolutional neural network model in step (4) is constructed as follows:
one-time distributed generation of 16 channel state information hij＝gijαijThe achievable rate is:
Figure BDA0002681713800000021
the multi-cell cellular system and rate maximization problem is:
Figure BDA0002681713800000031
wherein, base station index J belongs to J, cellular device index I belongs to I, D2D device index D belongs to D, and the correlation state and channel gain between cellular device I and base station J are represented as Zij、hijIf Z isijWhen 1 denotes a cellular device i and a base stationj association, data service provided by this base station, otherwise ZijThe transmission power between 0 is denoted as pij(ii) a The channel gain of the D2D link is denoted as hdWith a transmission power of pdBy default all devices transmit at maximum power, i.e. pij＝pd＝pmaxReachable Rate R of cellular device i to base station jijAnd W in the formula is the bandwidth,
Figure BDA0002681713800000032
is the ambient background noise;
the optimization problem contains two constraint terms:
Figure BDA0002681713800000033
and
Figure BDA0002681713800000034
respectively, that each cellular device can be provided with communication services by only one base station, and that each base station can simultaneously serve terminals not exceeding the upper limit θ at most.
Further, the structure of the convolutional neural network in the step (4) comprises 1 convolutional layer, 3 hidden layers and 1 output layer; the convolutional layer is used as an input layer to receive channel gain signals, 2 convolutional kernels with the size of 3 multiplied by 3 are used for generating a characteristic diagram, and then the characteristic diagram passes through a hidden layer formed by 3 full-connected layers, wherein the number of neurons is respectively 100, 80 and 50; the hidden layer selects a ReLU activation function to provide a nonlinear capability, and the output layer dense layer generates a user association strategy after being activated by a sigmoid function.
Further, the loss function in step (5) is constructed as follows:
Figure BDA0002681713800000035
wherein, a neural network output prediction association strategy Z is constructedijAnd labels Z obtained by exhaustion*As a loss function L; adopting a small batch gradient descent algorithm, wherein each batch comprises M samples, the training period is gamma-300, and an optimizerAnd selecting a random gradient descent algorithm to update the weight and the bias of the neural network.
Has the advantages that: compared with the prior art, the invention has the beneficial effects that: the method of the invention considers the problem of the interference between links caused by the density of the base station and the coexistence of multiple users in the multi-cell network environment, and relieves the problem by properly adjusting the association strategy of the cellular users and the base station in the system; based on channel gains of equipment in a multi-cell environment, learning a mapping relation between channel state information and an optimal user association strategy by using a convolutional neural network in a supervised learning mode; the trained convolutional neural network can be used for making an on-line decision, and a user association scheme with good performance and reliability is provided in real time.
Drawings
FIG. 1 is a structural design of a convolutional neural network;
FIG. 2 is a diagram of a convolutional neural network architecture used in the embodiment;
FIG. 3 is a flow chart of a small batch gradient descent training in the training phase;
fig. 4 is a test phase learning flow diagram.
Detailed Description
The present invention will be described in further detail with reference to the accompanying drawings.
The structure of the convolutional neural network used in the present invention is shown in fig. 1, and the convolutional neural network is responsible for learning the mapping relationship between the input signal (channel gain information of the device) and the output signal (optimal user association policy). The whole process comprises the collection of a training data set, the training of a neural network and the verification of a test stage. The specific implementation steps are as follows:
step 1: channel gain information of terminal devices within the environment is collected.
Assuming that in a 1km x 1km square area, there are 3 base stations providing communication services for 8 terminal devices within the environment, including 4 cellular subscribers and 4 pairs of D2D device pairs, the distance between the transceivers is 5,65]Medium and uniform distribution, under the channel environment of Rayleigh fading, the bandwidth W is 10MHz, N0-174dBm/Hz, interference upper limit IS＝103·N0W。
Step 2: operation exhaustion method to obtain optimal cellular user association strategy Z under corresponding channel gain sample*Collecting a training data set comprising channel gain information samples and corresponding optimal user association labels; and determining the segmentation proportion of the training set and the test set.
And step 3: and (4) constructing a convolutional neural network model and initializing neural network parameters.
The structure for constructing the convolutional neural network comprises 1 convolutional layer, 3 hidden layers and 1 output layer, and is shown in fig. 2. The channel gain of the devices in the environment is input into a convolutional neural network as an input signal, a first part convolutional layer of the convolutional neural network is used as an input layer to receive a channel gain signal, 2 convolutional kernels with the size of 3 multiplied by 3 are used for generating a feature map, and then the feature map passes through a hidden layer formed by 3 layers of full-connected layers, wherein the number of neurons is respectively 100, 80 and 50. The hidden layer selects a ReLU activation function to provide nonlinear capability, and finally the output layer dense layer generates a user association strategy after being activated by a sigmoid function.
The channel between links only takes into account the effects of fast fading, where the fading coefficient gij＝1，
Figure BDA0002681713800000051
One distribution can generate 16 pieces of channel state information hij＝gijαijThe achievable rate can be expressed as:
Figure BDA0002681713800000052
the multi-cell cellular system and rate maximization problem is:
Figure BDA0002681713800000053
wherein, base station index J belongs to J, cellular device index I belongs to I, D2D device index D belongs to D, and the correlation state and channel gain between cellular device I and base station J are represented as Zij、hijIf Z isijWhen 1 representsCellular device i is associated with base station j, which provides data service, otherwise ZijThe transmission power between 0 is denoted as pij(ii) a The channel gain of the D2D link is denoted as hdWith a transmission power of pdBy default all devices transmit at maximum power, i.e. pij＝pd＝pmaxReachable Rate R of cellular device i to base station jijAnd W in the formula is the bandwidth,
Figure BDA0002681713800000054
is the ambient background noise;
the optimization problem contains two constraint terms:
Figure BDA0002681713800000055
and
Figure BDA0002681713800000056
respectively, that each cellular device can be provided with communication services by only one base station, and that each base station can simultaneously serve terminals not exceeding the upper limit θ at most. Simultaneously, running an exhaustive algorithm to obtain a corresponding optimal user association strategy Z*Repeating for 10 ten thousand times to obtain 10 ten thousand data sets with channel gain and corresponding labels, and then determining the division ratio of the training set to the test set to be 4: 1.
And 4, step 4: inputting a training data set into a neural network, constructing MSE between the output of the neural network and a label as a loss function, and selecting a random gradient descent algorithm to update the weight of the neural network; and when the loss function is smaller than a preset value or reaches the iteration times, the training of the neural network is considered to be finished, and the neural network is stored.
And (3) constructing the output value of the neural network and the label mean square error as loss functions during training:
Figure BDA0002681713800000061
wherein, a neural network output prediction association strategy Z is constructedijAnd labels Z obtained by exhaustion*Mean square error ofThe difference is taken as a loss function L; with a small batch gradient descent algorithm, each batch containing 40 samples, i.e., M40, and a training period γ 300, the optimizer selects a random gradient descent algorithm to update the weights and biases of the neural network. The method comprises the following specific steps:
1) collecting a training data set, wherein channel gains of equipment in an environment are collected, and an exhaustive algorithm is operated to obtain an optimal cellular user association strategy under the corresponding channel gains;
2) dividing the training data into a plurality of batches by adopting a small batch gradient descent algorithm;
3) constructing a convolutional neural network framework, and initializing convolutional neural network parameters;
4) traversing all batches of training data to construct a loss function;
5) updating the weight of the neural network by using a random gradient descent algorithm until the loss function is smaller than a preset threshold value;
6) and storing the trained neural network.
And 5: the test phase tests performance with the test set as input data.
And in the testing stage, the segmented samples are used as a testing set, the error magnitude of the output value of the convolutional neural network and the label is compared through the convolutional neural network, and the reliability of the convolutional neural network for realizing the optimal user association strategy is verified.
As shown in fig. 3 and 4, a channel model is established in a rayleigh fading environment, and channel gain information of a D2D device and a cellular device in the environment is collected. And running an exhaustive algorithm to obtain the optimal user associated label under the corresponding sample, collecting the sample and the label and forming a training set with 10 thousands of data. The training data set and test data set segmentation ratio was set to 4: 1. A convolutional neural network learning framework is constructed, and the convolutional neural network learning framework is composed of 1 convolutional layer, 3 hidden layers and 1 output layer, wherein 2 convolutional kernels of 3 x 3 are used for convolution calculation in the convolutional layer. The training data set is fed into the neural network, and the mean square error of the output and label of the neural network is constructed as a loss function
Figure BDA0002681713800000062
The training data is divided into 2000 batches by a small batch gradient descent algorithm, the number of samples M in each batch is 40, and the random gradient descent algorithm is used as an optimizer to update the weight of the neural network. And stopping iteration when the loss function is less than 0.01 or meets 300 iteration cycles, and storing the neural network. And in the testing stage, a testing data set is input into the trained convolutional neural network, and the error between the power distribution result and the label is verified to be less than 0.01, so that the reliability of the method is proved.
In a multi-cell multi-device hybrid network environment, the problem of interference caused by D2D device communication on cellular devices is solved, the characteristic of large-scale network characteristics is captured by using a convolutional neural network, the convolutional neural network is applied to learning the mapping relation between channel gain information of devices in the environment and an optimal cellular association strategy, and the learning process is carried out in a supervised learning mode. The trained neural network can be used for real-time online decision making in practical application, and meanwhile, higher system performance is guaranteed.

Claims (5)

1. A method for deep learning based user association in a multi-cell cellular D2D device, comprising the steps of:
(1) collecting channel gain information of terminal equipment in an environment;
(2) optimal cellular user association strategy Z under corresponding channel gain samples obtained by applying exhaustion method*Collecting a training data set;
(3) determining the division ratio of the training set and the test set;
(4) constructing a convolutional neural network model, and initializing neural network parameters;
(5) inputting a training data set into a neural network, constructing MSE between the output of the neural network and a label as a loss function, and selecting a random gradient descent algorithm to update the weight of the neural network;
(6) when the loss function is smaller than a preset value or reaches the iteration times, the neural network training is considered to be finished, and the neural network is stored;
(7) the test phase tests performance with the test set as input data.
2. The deep learning based user association method in a multi-cell cellular D2D device according to claim 1, wherein the training data set in step (2) includes channel gain information samples and corresponding optimal user association label data.
3. The deep learning based user association method in multi-cell cellular D2D device according to claim 1, wherein the convolutional neural network model in step (4) is constructed as follows:
one-time distributed generation of 16 channel state information hij＝gijαijThe achievable rate is:
Figure FDA0002681713790000011
the multi-cell cellular system and rate maximization problem is:
Figure FDA0002681713790000012
wherein, base station index J belongs to J, cellular device index I belongs to I, D2D device index D belongs to D, and the correlation state and channel gain between cellular device I and base station J are represented as Zij、hijIf Z isijWhen 1, it means that the cellular device i is associated with the base station j, and the base station provides data service, otherwise, it is ZijThe transmission power between 0 is denoted as pij(ii) a The channel gain of the D2D link is denoted as hdWith a transmission power of pdBy default all devices transmit at maximum power, i.e. pij＝pd＝pmaxReachable Rate R of cellular device i to base station jijAnd W in the formula is the bandwidth,
Figure FDA0002681713790000021
is the ambient background noise;
the optimization problem contains two constraint terms:
Figure FDA0002681713790000022
and
Figure FDA0002681713790000023
respectively, that each cellular device can be provided with communication services by only one base station, and that each base station can simultaneously serve terminals not exceeding the upper limit θ at most.
4. The deep learning-based user association method in multi-cell cellular D2D equipment according to claim 1, wherein the convolutional neural network structure of step (4) comprises 1 convolutional layer, 3 hidden layers, and 1 output layer; the convolutional layer is used as an input layer to receive channel gain signals, 2 convolutional kernels with the size of 3 multiplied by 3 are used for generating a characteristic diagram, and then the characteristic diagram passes through a hidden layer formed by 3 full-connected layers, wherein the number of neurons is respectively 100, 80 and 50; the hidden layer selects a ReLU activation function to provide a nonlinear capability, and the output layer dense layer generates a user association strategy after being activated by a sigmoid function.
5. The method for deep learning based user association in a multi-cell cellular D2D device according to claim 1, wherein the loss function in step (5) is constructed as follows:
Figure FDA0002681713790000024
wherein, a neural network output prediction association strategy Z is constructedijAnd labels Z obtained by exhaustion*As a loss function L; and (3) adopting a small batch gradient descent algorithm, wherein each batch contains M samples, the training period is gamma 300, and the optimizer selects a random gradient descent algorithm to update the weight and the bias of the neural network.
CN202010964458.4A 2020-09-15 2020-09-15 Deep learning-based user association method in multi-cell cellular D2D equipment Active CN112153615B (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN202010964458.4A CN112153615B (en) 2020-09-15 2020-09-15 Deep learning-based user association method in multi-cell cellular D2D equipment

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN202010964458.4A CN112153615B (en) 2020-09-15 2020-09-15 Deep learning-based user association method in multi-cell cellular D2D equipment

Publications (2)

Publication Number Publication Date
CN112153615A true CN112153615A (en) 2020-12-29
CN112153615B CN112153615B (en) 2022-07-12

Family

ID=73893513

Family Applications (1)

Application Number Title Priority Date Filing Date
CN202010964458.4A Active CN112153615B (en) 2020-09-15 2020-09-15 Deep learning-based user association method in multi-cell cellular D2D equipment

Country Status (1)

Country Link
CN (1) CN112153615B (en)

Cited By (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113242066A (en) * 2021-05-10 2021-08-10 东南大学 Multi-cell large-scale MIMO communication intelligent power distribution method
CN113518457A (en) * 2021-04-22 2021-10-19 南京信息工程大学滨江学院 Power distribution strategy based on one-dimensional deep convolutional neural network
CN114367988A (en) * 2022-02-11 2022-04-19 煤炭科学研究总院有限公司 Mechanical arm motion planning method and device on autonomous inspection platform of coal mine
CN114531726A (en) * 2021-06-30 2022-05-24 南京信息工程大学滨江学院 Power control strategy based on dense wireless network
WO2024187483A1 (en) * 2023-03-16 2024-09-19 Huawei Technologies Co., Ltd. Methods, apparatus and medium for training an articifical intelligence or machine learning model

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108009525A (en) * 2017-12-25 2018-05-08 北京航空航天大学 A kind of specific objective recognition methods over the ground of the unmanned plane based on convolutional neural networks
CN109639377A (en) * 2018-12-13 2019-04-16 西安电子科技大学 Dynamic spectrum resource management method based on deeply study
WO2019190036A1 (en) * 2018-03-27 2019-10-03 경상대학교산학협력단 Method and apparatus for transmission power control based on neural network in wireless communication system

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108009525A (en) * 2017-12-25 2018-05-08 北京航空航天大学 A kind of specific objective recognition methods over the ground of the unmanned plane based on convolutional neural networks
WO2019190036A1 (en) * 2018-03-27 2019-10-03 경상대학교산학협력단 Method and apparatus for transmission power control based on neural network in wireless communication system
CN109639377A (en) * 2018-12-13 2019-04-16 西安电子科技大学 Dynamic spectrum resource management method based on deeply study

Non-Patent Citations (2)

* Cited by examiner, † Cited by third party
Title
李孜恒等: "基于深度强化学习的无线网络资源分配算法", 《通信技术》 *
谢添等: "基于强化学习的定向无线通信网络抗干扰资源调度算法", 《电波科学学报》 *

Cited By (9)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN113518457A (en) * 2021-04-22 2021-10-19 南京信息工程大学滨江学院 Power distribution strategy based on one-dimensional deep convolutional neural network
CN113518457B (en) * 2021-04-22 2024-06-07 南京信息工程大学滨江学院 Power distribution strategy based on one-dimensional deep convolutional neural network
CN113242066A (en) * 2021-05-10 2021-08-10 东南大学 Multi-cell large-scale MIMO communication intelligent power distribution method
CN113242066B (en) * 2021-05-10 2022-07-26 东南大学 Multi-cell large-scale MIMO communication intelligent power distribution method
CN114531726A (en) * 2021-06-30 2022-05-24 南京信息工程大学滨江学院 Power control strategy based on dense wireless network
CN114531726B (en) * 2021-06-30 2024-03-29 南京信息工程大学滨江学院 Power control method based on dense wireless network
CN114367988A (en) * 2022-02-11 2022-04-19 煤炭科学研究总院有限公司 Mechanical arm motion planning method and device on autonomous inspection platform of coal mine
CN114367988B (en) * 2022-02-11 2023-11-17 煤炭科学研究总院有限公司 Mechanical arm motion planning method and device on autonomous inspection platform of coal mine
WO2024187483A1 (en) * 2023-03-16 2024-09-19 Huawei Technologies Co., Ltd. Methods, apparatus and medium for training an articifical intelligence or machine learning model

Also Published As

Publication number Publication date
CN112153615B (en) 2022-07-12

Similar Documents

Publication Publication Date Title
CN112153615B (en) Deep learning-based user association method in multi-cell cellular D2D equipment
CN112153616B (en) Power control method in millimeter wave communication system based on deep learning
Huang et al. Deep learning for physical-layer 5G wireless techniques: Opportunities, challenges and solutions
Wang et al. Federated learning for automatic modulation classification under class imbalance and varying noise condition
CN112153617B (en) Terminal equipment transmission power control method based on integrated neural network
WO2020253691A1 (en) Deep learning signal detection method based on conjugate gradient descent method
CN113473580B (en) User association joint power distribution method based on deep learning in heterogeneous network
CN113435247A (en) Intelligent identification method, system and terminal for communication interference
CN113596785A (en) D2D-NOMA communication system resource allocation method based on deep Q network
Shodamola et al. A machine learning based framework for KPI maximization in emerging networks using mobility parameters
CN107528649A (en) A kind of Forecasting Methodology of the cognitive radio networks frequency spectrum based on Bagging
CN113784359A (en) Dynamic channel access method based on improved BP neural network algorithm
CN114567398B (en) Frequency spectrum sensing method based on convolution long short-time memory neural network
CN111010222B (en) Deep learning-based large-scale MIMO downlink user scheduling method
CN113259289B (en) Single-channel aliasing signal modulation mode identification method based on residual error neural network
Ya et al. Modulation recognition of digital signal based on deep auto-ancoder network
US11489560B2 (en) Method of parameter estimation for a multi-input multi-output system
CN113518457B (en) Power distribution strategy based on one-dimensional deep convolutional neural network
CN116981090A (en) D2D communication resource allocation method based on improved D3QN under multiple base stations
CN114531726B (en) Power control method based on dense wireless network
Wang et al. Joint beamforming and power allocation using deep learning for D2D communication in heterogeneous networks
CN115623575A (en) Power distribution method under CR-NOMA scene
CN113301576B (en) Cellular network resource allocation method based on improved genetic algorithm
Lin Design of intelligent distance music education system based on pan-communication technology
Yang et al. A spectrum prediction method for bursty frequency bands

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
TA01 Transfer of patent application right
TA01 Transfer of patent application right

Effective date of registration: 20210702

Address after: No.333 Xishan Avenue, Xishan District, Wuxi City, Jiangsu Province

Applicant after: Binjiang College of Nanjing University of Information Engineering

Applicant after: ICTEHI TECHNOLOGY DEVELOPMENT Co.,Ltd.

Address before: No.333 Xishan Avenue, Xishan District, Wuxi City, Jiangsu Province

Applicant before: Binjiang College of Nanjing University of Information Engineering

GR01 Patent grant
GR01 Patent grant