CN114844537B - Deep learning auxiliary robust large-scale MIMO receiving and transmitting combined method - Google Patents
Deep learning auxiliary robust large-scale MIMO receiving and transmitting combined method Download PDFInfo
- Publication number
- CN114844537B CN114844537B CN202210473641.3A CN202210473641A CN114844537B CN 114844537 B CN114844537 B CN 114844537B CN 202210473641 A CN202210473641 A CN 202210473641A CN 114844537 B CN114844537 B CN 114844537B
- Authority
- CN
- China
- Prior art keywords
- robust
- optimal
- power
- channel
- transmitting
- Prior art date
- Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
- Active
Links
- 238000000034 method Methods 0.000 title claims abstract description 40
- 238000013135 deep learning Methods 0.000 title claims abstract description 18
- 239000013598 vector Substances 0.000 claims abstract description 44
- 238000013528 artificial neural network Methods 0.000 claims abstract description 28
- 238000004422 calculation algorithm Methods 0.000 claims abstract description 28
- 239000011159 matrix material Substances 0.000 claims abstract description 16
- 230000005540 biological transmission Effects 0.000 claims abstract description 8
- 238000000354 decomposition reaction Methods 0.000 claims abstract description 7
- 238000012549 training Methods 0.000 claims description 16
- 230000008569 process Effects 0.000 claims description 8
- 230000000295 complement effect Effects 0.000 claims description 3
- 238000004088 simulation Methods 0.000 claims description 3
- 239000000284 extract Substances 0.000 claims description 2
- 238000013461 design Methods 0.000 description 5
- 238000005457 optimization Methods 0.000 description 4
- 238000004364 calculation method Methods 0.000 description 3
- 230000006870 function Effects 0.000 description 3
- 230000003044 adaptive effect Effects 0.000 description 2
- 238000010606 normalization Methods 0.000 description 2
- ORILYTVJVMAKLC-UHFFFAOYSA-N Adamantane Natural products C1C(C2)CC3CC1CC2C3 ORILYTVJVMAKLC-UHFFFAOYSA-N 0.000 description 1
- 230000004913 activation Effects 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 238000010586 diagram Methods 0.000 description 1
- 230000009977 dual effect Effects 0.000 description 1
- 230000005251 gamma ray Effects 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000011423 initialization method Methods 0.000 description 1
- 238000012804 iterative process Methods 0.000 description 1
- 239000000463 material Substances 0.000 description 1
- 238000012986 modification Methods 0.000 description 1
- 230000004048 modification Effects 0.000 description 1
- 238000012545 processing Methods 0.000 description 1
- 238000001228 spectrum Methods 0.000 description 1
Classifications
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
- H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
- H04B7/0413—MIMO systems
-
- 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
- H04B7/00—Radio transmission systems, i.e. using radiation field
- H04B7/02—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas
- H04B7/04—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas
- H04B7/0413—MIMO systems
- H04B7/0426—Power distribution
- H04B7/043—Power distribution using best eigenmode, e.g. beam forming or beam steering
-
- 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)
- 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)
- Signal Processing (AREA)
- Computer Networks & Wireless Communication (AREA)
- Power Engineering (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses a deep learning-assisted large-scale MIMO downlink transmission robust transceiving combined beamforming method, wherein a transmitting end (base station) obtains the optimal beamforming vectors of the transmitting end and the receiving end under a mobility enhanced channel with low complexity under certain transmitting power and service quality constraint based on imperfect channel state information of each receiving end (user end). Firstly, calculating a beam forming direction vector of a receiving end based on a heuristic algorithm of channel matrix eigenvalue decomposition; then, directly predicting an optimal Lagrangian multiplier from a channel matrix through a pre-trained deep neural network structure; and finally, respectively solving the direction and the power of the beamforming vector of the transmitting end by utilizing an optimal solution structure, introducing the constraint of the transmitting power into a linear equation set of the beamforming power of the transmitting end, and calculating an optimal robust factor. The invention can realize the robust joint beamforming adapting to the mobility enhanced channel while keeping the complexity of the online stage low.
Description
Technical Field
The invention belongs to the field of transmit-receive combined beam forming in the field of wireless communication, and particularly relates to a deep learning auxiliary robust large-scale MIMO transmit-receive combined method.
Background
In recent years, a massive MIMO system can simultaneously transmit multiple sets of parallel data with different degrees of spatial freedom by using the same time and frequency resources, so as to improve power efficiency and spectrum efficiency, and is receiving extensive attention in the industry. With the continued improvement of the fifth generation (5G) systems, the large number of antennas at the transmitting end (base station) and the receiving end (user) provides a great degree of freedom for joint design of transceivers. When the receiving end is equipped with a plurality of antennas, the design of beamforming vectors of the receiving end and the transmitting end plays a crucial role.
A common optimization problem in transmit-receive joint beam design is to minimize the system transmission power under conditions that meet the quality of service (Quality of Service, qoS) requirements. The optimal solution of the problem can be obtained through an MMSE-SOCP algorithm or an iterative algorithm such as a UDD algorithm utilizing an uplink and downlink dual theory. In the iterative algorithm, the MMSE-DUAL integrates the MMSE-SOCP algorithm and the UDD algorithm, and the global optimization problem is converted into a series of simplified sub-problems, so that good performance can be obtained. However, in the scenario of high-speed movement, the base station can only acquire imperfect channel state information, which makes the solution of the conventional iterative algorithm based on perfect channel state information unable to meet the QoS constraint of the user end, and measures need to be taken to adapt the algorithm to the mobility enhanced channel.
In a large-scale MIMO system, beamforming vectors of a receiving end and a transmitting end often have very high dimensionality, so that the calculation time of a traditional iterative algorithm is extremely high, and the application of the algorithm in an actual scene is limited.
Disclosure of Invention
The invention aims to provide a deep learning auxiliary robust large-scale MIMO receiving and transmitting combined method, which aims to solve the technical problems that beamforming vectors of a receiving end and a transmitting end often have very high dimensionality, so that the calculation time of a traditional iterative algorithm is extremely high, and the application of the algorithm in an actual scene is limited.
In order to solve the technical problems, the specific technical scheme of the invention is as follows:
a deep learning aided robust massive MIMO transceiving combined method, based on imperfect channel state information, using a deep neural network to maximize a robust factor with low complexity under transmit power and quality of service constraints, performing the steps of:
directly performing heuristic eigenvalue decomposition on a channel matrix, and taking an eigenvector corresponding to the maximum eigenvalue as a beam forming vector of a receiving end;
step two, calculating a power distribution vector of a downlink receiving end, namely a Lagrangian multiplier through a deep neural network;
step three, using the obtained beamforming vector and Lagrangian multiplier of the receiving end as input, and calculating the direction of the beamforming vector of the transmitting end by utilizing an optimal solution structure;
step four, solving an optimal robust factor by using the constraint of the transmission power;
calculating the beam forming power of a transmitting end by solving a linear system containing an optimal robust factor and a service quality constraint;
wherein, training sets used for training the deep neural network structure are generated by a data generation algorithm.
Further, the maximization of the robust factor under the constraint of the transmitting power and the service quality is equivalent to the maximization of the service quality satisfaction rate under the constraint of the Gaussian channel noise model and the total power of the transmitting end.
Further, the robustness factor is a complementary term of noise variance introduced based on an offset maximization method, and the factor characterizes the upper bound of a channel noise module, namely, the worst case of channel estimation.
Further, the main process of establishing the training set is as follows: the base station side processes the received pilot signal into imperfect channel state information, and records the imperfect channel state information in a training set together with Lagrangian multipliers generated by a traditional joint beamforming iterative algorithm; the data generation algorithm comprises the following steps:
step 1, generating a large number of simulation channel matrixes under different channel noise power, receiving end moving speed and equipment distribution environments;
step 2, operating a traditional iterative algorithm to calculate a feasible optimal Lagrangian multiplier;
and step 3, combining the channel matrix and the Lagrangian multiplier into a data sample.
Further, the deep neural network structure comprises the step of inputting the pre-trained deep neural network to calculate Lagrange multipliers by using the channel matrix.
Furthermore, the deep neural network structure utilizes a convolution module to extract low-dimensional semantic feature vectors from the high-dimensional channel state information, and inputs the feature vectors into a full-connection layer for regression prediction.
Further, the process of solving the optimal robust factor is as follows: after the beamforming direction vector and Lagrange multiplier of the transmitting end are obtained, the optimal robust factor is directly calculated by taking the conditions of the service quality constraint and the transmitting power constraint.
The deep learning auxiliary robust large-scale MIMO receiving and transmitting combined method has the following advantages:
1. the invention assumes that the channel noise obeys Gaussian distribution, and utilizes an offset maximization method to equate the maximization QoS satisfaction rate under the total power limit of the transmitting end to the optimization problem of the maximization robust factor under certain transmitting power and service quality constraint, and obtains the beamforming vector of the receiving end and the transmitting end corresponding to the optimal solution of the optimization problem, thereby being capable of realizing the receiving-transmitting combined method under the high mobility channel.
2. The invention uses the feature vector corresponding to the maximum feature value obtained by heuristic feature decomposition as the beam forming direction vector of each receiving end, separates the robust factors and avoids the high complexity introduced by the loop iteration robust factors.
3. According to the invention, the Lagrange multiplier which is relatively insensitive is predicted by using the deep neural network, and multiple iterations are avoided while the original performance of the algorithm is maintained, so that the calculation complexity is reduced, and the time consumption of offline training can be flexibly controlled according to the requirement of prediction precision.
4. The deep neural network training set is generated by a traditional joint beamforming iterative algorithm, and is convenient to integrate into an existing system.
Drawings
FIG. 1 is a flow chart of a method according to an embodiment of the present invention;
fig. 2 is a schematic diagram of an overall framework of the deep learning-assisted robust massive MIMO transceiver joint method of the present invention.
Detailed Description
In order to better understand the purpose, structure and function of the present invention, the following describes in further detail a deep learning auxiliary robust massive MIMO transceiver combination method according to the present invention with reference to the accompanying drawings.
In the deep learning auxiliary robust large-scale MIMO receiving and transmitting combined method disclosed by the embodiment of the invention, a transmitting end (base station) provided with a unified plane array provides service for a plurality of multi-antenna receiving ends (user ends) in a cell, and based on imperfect channel state information, the beam forming vectors of the optimal transmitting end and the optimal receiving end under a mobility enhanced channel are obtained with low complexity by maximizing a robust factor through a properly trained deep neural network structure under a certain transmission power and service quality constraint.
The complexity is reduced by using the deep neural network structure, and the following steps are executed:
step one: directly performing heuristic eigenvalue decomposition on the channel matrix, and taking an eigenvector corresponding to the maximum eigenvalue as a beam forming vector of the receiving end;
step two: calculating a power distribution vector of a downlink receiving end, namely a Lagrangian multiplier through a deep neural network;
step three: using the obtained beamforming vector and Lagrangian multiplier of the receiving end as input, and calculating the direction of the beamforming vector of the transmitting end by utilizing an optimal solution structure;
step four: solving an optimal robust factor by using a transmission power constraint;
step five: and calculating the beam forming power of the transmitting end by solving a linear system containing the optimal robust factor and the service quality constraint.
Wherein, training sets used for training the deep neural network structure are generated by a data generation algorithm.
Under certain transmitting power and service quality constraint, the maximization of the robust factor is equivalent to the maximization of the QoS satisfaction rate under the Gaussian channel noise model and the total power limit of the transmitting end.
The robustness factor is a complementary term of noise variance introduced based on an offset maximization method, and the factor characterizes the upper bound of a channel noise module, namely the worst case of channel estimation.
The main process for establishing the training set is as follows: and the base station side processes the received pilot signal into imperfect channel state information, and records the imperfect channel state information in a training set together with Lagrangian multipliers generated by a traditional joint beamforming iterative algorithm. The data generation algorithm comprises the following steps:
1. generating a large number of simulation channel matrixes under different channel noise power, receiving end moving speed and equipment distribution environments;
2. operating a traditional iterative algorithm to calculate a feasible optimal Lagrangian multiplier;
3. the channel matrix and the lagrangian multiplier are combined into one data sample.
The robustness factor is introduced based on an offset maximization method, and the factor characterizes the upper bound of a channel noise module, namely the worst case of channel estimation.
The deep neural network structure comprises the step of inputting a pre-trained deep neural network to calculate Lagrange multipliers by using a channel matrix.
The deep neural network structure utilizes a convolution module to extract low-dimensional semantic feature vectors from high-dimensional channel state information, and inputs the feature vectors into a full-connection layer for regression prediction.
The process for solving the optimal robust factor is as follows: after the beamforming direction vector and Lagrange multiplier of the transmitting end are obtained, the optimal robust factor is directly calculated by taking the QoS constraint and the transmitting power constraint and other conditions.
The method of the invention is not limited to specific scenes, and for other implementations outside the exemplary scenes of the invention, a person skilled in the art can adapt to the specific scenes by utilizing the prior knowledge according to the technical thought of the invention. The method of the embodiment of the invention is further described below in connection with a specific implementation scenario.
1) System configuration
Consider a multi-user MIMO system consisting of a single base station and K users, with the base station as the transmitting end and the users as the receiving end. Wherein a Base Station (BS) is equipped with m=m v ×M h Uniform panel antenna array (uniform planar array, UPA), M v Representing the number of antennas per vertical column, M h Representing each horizontal lineIs used for the number of antennas. Each user is equipped with N > 1 antennas.
2) Signal model
Suppose that the base station will uniflow signal s k Transmitted to the intended receiver user k using a transmit beamformer v k ∈C M ×1 To transmit a data stream, and after passing through the channel matrix H, the received signal of the kth user is y k ∈C N×1 The method comprises the following steps:
wherein H is k ∈C N×M Representing a channel matrix, v j ∈C M×1 Beamforming vector s for transmitting end corresponding to jth user j Single stream signal transmitted by base station to jth user, n k ∈C N×1 Satisfying independent co-distribution for elementsNoise vector of>The noise power of the kth user, K is the total number of users, and N is the number of antennas equipped for each user.
At the receiving end, the kth user can utilize the received beam forming vectorTo estimate the data stream it transmits, estimated data stream +.>Can be expressed as:
3) Problem modeling
For general cases, the goal is to design transmit-receive beamformingDevice u 1 ，...，u K ，v 1 ，…，v K Minimizing the transmission rate under quality of service constraints.
In the case of Time-division Duplex (TDD), the BS can only obtain incomplete channel state information in the uplink training phase. To introduce uncertainty, the channel is modeled as:
wherein,for estimation, E k For mobility enhanced channel estimation errors, obeying complex gaussian distribution, variance is +.>
QoS constraints are expressed in terms of signal-to-interference-and-noise ratio (signal to interference plus noise ratio, SINR).
SINR (signal to interference plus noise ratio) of user k k The definition is as follows:
our objective is to design transmit and receive beamformers that minimize the outage probability delta over a certain total transmit power k ，k＝1，2，...，K。
The above problem can be expressed as a problem of minimizing the out-of-bounds probability:
wherein delta k The probability that the kth does not meet the QoS constraint of the receiving end (user), namely the out-of-bounds probability; gamma ray k > 0 is the SINR constraint for user k; u epsilon C M×K Representing a set of receive beamformers, denoted asU＝[u 1 ，...，u K ]；V∈C N×K Representing a set of transmit beamformers, denoted v= [ V ] 1 ，...，v K ]；A H Representing the conjugate transpose of matrix a; the A and a represent L2 norms of the matrix A and the vector a respectively; the |x| represents modulo the complex number x; x is x 2 Representing squaring the real number x.
4) Deep learning auxiliary robust large-scale MIMO receiving and transmitting combined method
The input P is provided to the computer with a logic circuit,k＝1，2，...，K，γ k the signal-to-interference-and-noise ratio of user K is represented, K being the total number of users. Setting the fixed point iteration number N > 0, initializing the transmit beamformer vk, k=1, …, K, lagrangian multiplier λ= [ λ ] 1 ...，λ K ]∈C K×1 Uplink power allocation vector μ= [ μ ] 1 ...，μ K ]∈C K×1 Is a random value. The following steps are repeated until the convergence condition is satisfied.
Step one: directly performing heuristic eigenvalue decomposition on the channel matrix, and taking the eigenvector corresponding to the maximum eigenvalue as a beamforming vector of the receiving end:
order theCalculating the beamforming vector u of the receiving end by utilizing eigenvalue decomposition k ：
Wherein I represents a unit array.
Then, normalization processing is performed
Step two: calculating a power distribution vector of a downlink receiving end through a deep neural network, namely Lagrangian multiplier:
λ＝g(X) (8)
λ∈C K×1 representing a vector of lagrange multipliers. Representing the deep neural network, and converting the iterative process into a parameter prediction process. The deep neural network structure is shown in fig. 2. X is the input of the deep neural network, which is the imperfect CSI part, namely:
X＝[Re(H)，Im(H)] (9)
where Re represents the real part and Im represents the imaginary part. We extract features using a column vector convolution module and input the full connection layer. The convolution module is a combination of a convolution layer, a batch normalization layer, and an activation layer. Wherein the active layer is a linear rectification function (ReLU):
minimizing the mean square error (mean squared error, MSE) loss function:
wherein the method comprises the steps ofRepresenting a prediction of the ith sample by the DNN structure. />Representing the total number of training samples.
The deep neural network weight parameters are initialized using a Kaiming initialization method. The dropout with a probability of 0.5 is introduced before the output layer, i.e. the fully connected layer, to suppress overfitting. An adaptive moment estimation (adaptive moment estimation, ADAM) algorithm is employed to train the neural network.
Step three: using the obtained beamforming vector and Lagrangian multiplier of the receiving end as input, and calculating the beamforming vector direction of the transmitting end by using an optimal solution structure:
step four: solving the optimal robust factor by using the transmission power constraint:
r ★ ←(P-1 T A -1 c)/(1 T A -1 1)
step five: and calculating the beam forming power of the transmitting end by solving a linear system containing the optimal robust factor and the service quality constraint.
Calculating an optimal solution of a noise variance robust factor r of a specific constant satisfying the original SINR constraint:
mu is set k ＝||v k || 2 And μ= [ μ ] 1 ，μ 2 ，...，μ K ] T . According to the following:
Aμ＝c+r 0 1 (14)
wherein the method comprises the steps of
As a result of:
so the optimal solution for r can be found as:
r ★ ←(P-1 T A -1 c)/(1 T A -1 1) (16)
solving the linear equation set of (14) to obtain mu, and further obtaining the beamforming vector v of the transmitting end k ：
It will be understood that the invention has been described in terms of several embodiments, and that various changes and equivalents may be made to these features and embodiments by those skilled in the art without departing from the spirit and scope of the invention. In addition, many modifications may be made to adapt a particular situation or material to the teachings of the invention without departing from the essential scope thereof. Therefore, it is intended that the invention not be limited to the particular embodiment disclosed, but that the invention will include all embodiments falling within the scope of the appended claims.
Claims (7)
1. The deep learning auxiliary robust large-scale MIMO receiving and transmitting combined method is characterized in that the method is based on imperfect channel state information, and under the constraint of transmitting power and service quality, a deep neural network is utilized to maximize a robust factor with low complexity, and the following steps are executed:
directly performing heuristic eigenvalue decomposition on a channel matrix, and taking an eigenvector corresponding to the maximum eigenvalue as a beam forming vector of a receiving end;
step two, calculating a power distribution vector of a downlink receiving end, namely a Lagrangian multiplier through a deep neural network;
step three, using the obtained beamforming vector and Lagrangian multiplier of the receiving end as input, and calculating the direction of the beamforming vector of the transmitting end by utilizing an optimal solution structure;
step four, solving an optimal robust factor by using the constraint of the transmission power;
calculating the beam forming power of a transmitting end by solving a linear system containing an optimal robust factor and a service quality constraint;
wherein, training sets used for training the deep neural network structure are generated by a data generation algorithm.
2. The deep learning aided robust massive MIMO transmit-receive joint method of claim 1, wherein maximizing the robust factor under transmit power and quality of service constraints is equivalent to maximizing the quality of service satisfaction rate under transmit-side total power constraints.
3. The deep learning aided robust massive MIMO transmit-receive joint method of claim 1, wherein the robust factor is a complementary term to the noise variance introduced based on the offset maximization method, and the factor characterizes the upper bound of the channel noise model, i.e., the worst case of the channel estimation.
4. The deep learning aided robust massive MIMO transceiving combined method according to claim 1, wherein the main procedure for establishing the training set is as follows: the base station side processes the received pilot signal into imperfect channel state information, and records the imperfect channel state information in a training set together with Lagrangian multipliers generated by a traditional joint beamforming iterative algorithm; the data generation algorithm comprises the following steps:
step 1, generating a large number of simulation channel matrixes under different channel noise power, receiving end moving speed and equipment distribution environments;
step 2, operating a traditional iterative algorithm to calculate a feasible optimal Lagrangian multiplier;
and step 3, combining the channel matrix and the Lagrangian multiplier into a data sample.
5. The deep learning aided robust massive MIMO transmit-receive joint method of claim 1, wherein the deep neural network structure comprises inputting a pre-trained deep neural network to compute lagrangian multipliers using a channel matrix.
6. The deep learning aided robust large-scale MIMO transceiving combined method of claim 1, wherein the deep neural network structure extracts a low-dimensional semantic feature vector from the high-dimensional channel state information by using a convolution module, and inputs the feature vector into a full-connection layer for regression prediction.
7. The deep learning aided robust large-scale MIMO transceiving combined method according to claim 1, wherein the solving the optimal robust factor comprises the following steps: after the beamforming direction vector and Lagrange multiplier of the transmitting end are obtained, the optimal robust factor is directly calculated by taking the conditions of the service quality constraint and the transmitting power constraint.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210473641.3A CN114844537B (en) | 2022-04-29 | 2022-04-29 | Deep learning auxiliary robust large-scale MIMO receiving and transmitting combined method |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN202210473641.3A CN114844537B (en) | 2022-04-29 | 2022-04-29 | Deep learning auxiliary robust large-scale MIMO receiving and transmitting combined method |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN114844537A CN114844537A (en) | 2022-08-02 |
| CN114844537B true CN114844537B (en) | 2024-02-02 |
Family
ID=82566947
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN202210473641.3A Active CN114844537B (en) | 2022-04-29 | 2022-04-29 | Deep learning auxiliary robust large-scale MIMO receiving and transmitting combined method |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN114844537B (en) |
Families Citing this family (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN116032332B (en) * | 2022-12-30 | 2024-04-12 | 哈尔滨工程大学 | Large-scale MIMO system detection model construction method suitable for changeable channel state information |
Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111865378A (en) * | 2020-05-28 | 2020-10-30 | 东南大学 | Large-scale MIMO downlink precoding method based on deep learning |
| CN113765553A (en) * | 2021-09-09 | 2021-12-07 | 东南大学 | Multi-beam satellite communication system robust precoding method based on machine learning |
| CN113765567A (en) * | 2021-09-08 | 2021-12-07 | 东南大学 | Large-scale MIMO (multiple input multiple output) transceiving combined beam forming method based on deep learning |
-
2022
- 2022-04-29 CN CN202210473641.3A patent/CN114844537B/en active Active
Patent Citations (3)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN111865378A (en) * | 2020-05-28 | 2020-10-30 | 东南大学 | Large-scale MIMO downlink precoding method based on deep learning |
| CN113765567A (en) * | 2021-09-08 | 2021-12-07 | 东南大学 | Large-scale MIMO (multiple input multiple output) transceiving combined beam forming method based on deep learning |
| CN113765553A (en) * | 2021-09-09 | 2021-12-07 | 东南大学 | Multi-beam satellite communication system robust precoding method based on machine learning |
Also Published As
| Publication number | Publication date |
|---|---|
| CN114844537A (en) | 2022-08-02 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| Pan et al. | Multicell MIMO communications relying on intelligent reflecting surfaces | |
| Qi et al. | Integrating sensing, computing, and communication in 6G wireless networks: Design and optimization | |
| WO2023044982A1 (en) | Sending method for satellite on the basis of integrated sensing and communication with massive mimo | |
| CN111865378B (en) | Large-scale MIMO downlink precoding method based on deep learning | |
| Tolli et al. | Decentralized minimum power multi-cell beamforming with limited backhaul signaling | |
| CN113765553B (en) | Multi-beam satellite communication system robust precoding method based on machine learning | |
| CN111147113B (en) | Multi-beam satellite communication robust precoding method for energy efficiency guarantee | |
| CN111092641A (en) | Hybrid precoding design method based on millimeter wave MIMO system deep learning | |
| CN112260737B (en) | Multi-beam satellite communication robust precoding method with total energy efficiency and minimum energy efficiency balanced | |
| CN110138427B (en) | Large-scale multi-input multi-output hybrid beam forming algorithm based on partial connection | |
| CN113193893B (en) | Millimeter wave large-scale MIMO intelligent hybrid beam forming design method | |
| CN111901812A (en) | Full-duplex cellular communication network base station and intelligent reflecting surface combined control method | |
| CN114844537B (en) | Deep learning auxiliary robust large-scale MIMO receiving and transmitting combined method | |
| CN110881010A (en) | Statistical CSI-assisted multi-user NOMA downlink transmission method | |
| Palhares et al. | Iterative AP selection, MMSE precoding and power allocation in cell‐free massive MIMO systems | |
| CN113765567B (en) | Large-scale MIMO receiving and transmitting combined beam forming method based on deep learning | |
| CN114567358B (en) | Large-scale MIMO robust WMMSE precoder and deep learning design method thereof | |
| CN115412134A (en) | Off-line reinforcement learning-based user-centered non-cellular large-scale MIMO power distribution method | |
| CN102291727B (en) | Distributed cooperative beam forming and designing method | |
| Mohamadi et al. | Low-complexity ADMM-based algorithm for robust multi-group multicast beamforming in large-scale systems | |
| Ni et al. | Partially distributed beamforming design for RIS-aided cell-free networks | |
| CN102740325B (en) | Method, device for acquiring channel information and method, device for optimizing beam forming | |
| CN109818887A (en) | Half-blind channel estimating method based on EVD-ILSP | |
| CN112188605B (en) | Design method of network-assisted full-duplex system transceiver in wireless energy-carrying communication scene | |
| CN115865148B (en) | Honeycomb-removing MIMO robust wave beam forming method under non-ideal channel |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| PB01 | Publication | ||
| PB01 | Publication | ||
| SE01 | Entry into force of request for substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |