CN109660287A - A kind of antenna selecting method based on deep learning - Google Patents
A kind of antenna selecting method based on deep learning Download PDFInfo
- Publication number
- CN109660287A CN109660287A CN201811503785.9A CN201811503785A CN109660287A CN 109660287 A CN109660287 A CN 109660287A CN 201811503785 A CN201811503785 A CN 201811503785A CN 109660287 A CN109660287 A CN 109660287A
- Authority
- CN
- China
- Prior art keywords
- channel
- antenna
- neural network
- state information
- deep learning
- 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
Links
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
- H04B7/06—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station
- H04B7/0602—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the transmitting station using antenna switching
- H04B7/0608—Antenna selection according to transmission parameters
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/309—Measuring or estimating channel quality parameters
- H04B17/336—Signal-to-interference ratio [SIR] or carrier-to-interference ratio [CIR]
-
- H—ELECTRICITY
- H04—ELECTRIC COMMUNICATION TECHNIQUE
- H04B—TRANSMISSION
- H04B17/00—Monitoring; Testing
- H04B17/30—Monitoring; Testing of propagation channels
- H04B17/309—Measuring or estimating channel quality parameters
- H04B17/345—Interference values
-
- 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
-
- 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
- H04B7/08—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the receiving station
- H04B7/0802—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the receiving station using antenna selection
- H04B7/0805—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the receiving station using antenna selection with single receiver and antenna switching
- H04B7/0814—Diversity systems; Multi-antenna system, i.e. transmission or reception using multiple antennas using two or more spaced independent antennas at the receiving station using antenna selection with single receiver and antenna switching based on current reception conditions, e.g. switching to different antenna when signal level is below threshold
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Physics & Mathematics (AREA)
- Electromagnetism (AREA)
- Quality & Reliability (AREA)
- Radio Transmission System (AREA)
Abstract
The invention discloses a kind of antenna selecting methods based on deep learning comprising: obtain the channel state information matrix of each channel in STSK MIMO communication system model；The channel state information matrix is carried out to take amplitude operation, to obtain eigenmatrix xi；The transmission error rates that different antennae combines in each channel are calculated according to channel state information, select the smallest antenna combination of the bit error rate in each channel to mark the channel, and constitute marker samples data set；The partial data in marker samples data set is chosen, and is iterated training with the propagated forward algorithm and back-propagation algorithm of neural network, to obtain the neural network classification model for being used for day line options；The channel for needing day line options is input in neural network classification model, to obtain corresponding antenna combination.Compared with prior art, the selection complexity of system can be greatly reduced in the case where guaranteeing that communication system has higher transmission reliability and power system capacity in method of the invention.
Description
Technical field
The present invention relates to communication and deep learning fields, relate more specifically to a kind of antenna selecting party based on deep learning
Method.
Background technique
In the prior art, MIMO (Multiple-InputMultiple-Output) technology is the key that the following 5G epoch
One of technology.MIMO technology uses multiple transmitting antennas and receiving antenna in transmitting terminal and receiving end respectively, and signal is made to pass through hair
It penetrates the mutiple antennas transmission at end and receiving end and receives, so as to improve communication quality.It can make full use of space resources, by more
A antenna realizes multiple-input multiple-output, because MIMO communication system is configured with more antennas, although increasing antenna amount can obtain more
Excellent transmission reliability, but number of antennas also brings along the rising of system cost and computational complexity.Its line options be exactly
In the case where not influencing its system performance, the complexity of system can be effectively reduced, traditional Antenna Selection Algorithem it is optimal be
The Antenna Selection Algorithem of traversal, but number of antennas is more, operand needed for traversing all antenna combinations rises sharply, and can give in this way
User brings very poor experience, huge challenge is faced in terms of the raising of performance in a conventional manner.
In consideration of it, can guarantee feelings of the communication system with higher transmission reliability and power system capacity it is necessary to provide a kind of
Under condition, be greatly reduced the selection complexity of system based on the antenna selecting method of deep learning to solve drawbacks described above.
Summary of the invention
Technical problem to be solved by the invention is to provide one kind to have higher transmission reliability guaranteeing communication system
In the case where power system capacity, the antenna selecting method based on deep learning of the selection complexity of system is greatly reduced.
In order to solve the above technical problems, the present invention provides a kind of antenna selecting method based on deep learning, this method packet
It includes:
Obtain the channel state information matrix of each channel in STSKMIMO model of communication system；
The channel state information matrix is carried out to take amplitude operation, to obtain eigenmatrix xi；
The transmission error rates that different antennae combines in each channel are calculated according to channel state information, are selected in each channel
The smallest antenna combination of the bit error rate marks the channel, and constitutes marker samples data set；
The partial data in marker samples data set is chosen, and is calculated with the propagated forward algorithm of neural network and backpropagation
Method is iterated training, to obtain the neural network classification model for being used for day line options；
The channel for needing day line options is input in neural network classification model, to obtain corresponding antenna combination.
Its further technical solution are as follows: described that the biography that different antennae combines in each channel is calculated according to channel state information
The defeated bit error rate selects the smallest antenna combination of the bit error rate in each channel to mark the channel, and constitutes marker samples data set packet
It includes:
The transmission error rates that different antennae combines in each channel are calculated according to channel state information；
Regard each antenna combination as a kind of label, wherein label classification be y (y=1,2 ..., s)；
The corresponding label classification y of the smallest antenna combination of the bit error rate in each channel is obtained, and is marked with the label classification y
Corresponding channel information eigenmatrix xi, and constitute marker samples data set (xi,yi)={ (x1,y1),(x2,y2),...,(xm,
ym)}。
Its further technical solution are as follows: the partial data chosen in marker samples data set, and with neural network
Propagated forward algorithm and back-propagation algorithm are iterated training, to obtain the neural network classification model packet for being used for day line options
It includes:
The partial data in marker samples data set is chosen, the other predicted value of tag class is obtained by propagated forward algorithmAnd calculate predicted valueWith true value yiBetween loss function；
Each parameter is updated by the gradient of back-propagation algorithm calculating parameter w and b, then by gradient descent method, with
Minimize loss function；
If loss function converges to minimum value or repetitive exercise reaches preset times, terminate training, obtains and be used for day
The neural network classification model of line options.
Its further technical solution are as follows: the loss function is
Its further technical solution are as follows: the channel status for obtaining each channel in STSK MIMO communication system model
Before information matrix, further includes:
Build STSKMIMO model of communication system, and using maximum Likelihood to the channel in the communication system into
Row estimation.
Its further technical solution are as follows: it is described that the channel for needing day line options is input in neural network classification model, with
After obtaining corresponding antenna combination, further includes: calculate the bit error rate of the antenna combination of acquisition and the size of channel capacity.
Compared with prior art, the present invention can be in the feelings for guaranteeing that communication system has higher transmission reliability and power system capacity
Under condition, the selection complexity of system, i.e. the channel shape by extracting each link of STSK MIMO model is greatly reduced
State information calculates the transmission error rates that different antennae combines in each channel as feature samples, and according to channel state information,
Using the smallest antenna combination marked channels of the bit error rate, and marker samples data set is constituted, then is instructed using the method for deep learning
Practice the partial data in marker samples data set, the neural network classification model for day line options can be obtained, if having new
Data information is input in the neural network classification model, then can obtain optimal antenna combination quickly, reduce system
Workload reduces the selection complexity of system.
Detailed description of the invention
Fig. 1 is the flow diagram of one specific embodiment of antenna selecting method the present invention is based on deep learning.
Fig. 2 is the sub-process schematic diagram of the antenna selecting method based on deep learning in Fig. 1.
Fig. 3 is another sub-process schematic diagram of the antenna selecting method based on deep learning in Fig. 1.
Fig. 4 is using antenna selecting method of the invention and using other antenna selecting methods
The simulation comparison figure of system transmission reliability.
Fig. 5 is using antenna selecting method of the invention and using other antenna selecting methods
The simulation comparison figure for channel capacity of uniting.
Specific embodiment
To make those skilled in the art that the object, technical solutions and advantages of the present invention be more clearly understood, with
Under the present invention is further elaborated in conjunction with the accompanying drawings and embodiments.
STSK (Space-Time ShiftKeying, empty time-shift keying) technology utilizes time-domain and spatial domain
Come, more specifically, STSK system be based on suitably activated in each STSK block duration the space-time being indexed disperse square
Battle array can optimize the dimension of dispersion matrix and the quantity of quantity and transmitting receiving antenna, and can use flexible diversity technique
Resist the loss of spatial multiplexing gain.
Referring to Fig.1, Fig. 1 is that the present invention is based on the signals of the process of one specific embodiment of antenna selecting method of deep learning
Figure.The described method includes:
S101, STSKMIMO model of communication system is built, and using maximum Likelihood in the communication system
Channel is estimated.
In the step, because channel state information is known dependent on sender and recipient in the basis of day line options
(Channel State Information, CSI), therefore need to estimate channel before progress day line options every time,
The then receiving end in STSKMIMO model of communication system, will be using the side maximal possibility estimation (Maximum likelihood, ML)
Method estimation receives signal.
S102, the channel state information matrix for obtaining each channel in STSKMIMO model of communication system.
In the step, transmitting antenna number Nt, receiving antenna number is NrWhen, then wireless channelIt can indicate
Are as follows:
Wherein, i=1,2,3 ... M
M=5 × 10 can be randomly generated according to rayleigh distributed for wireless channel quantity in M4A channel is as training set.
S103, the channel state information matrix is carried out to take amplitude operation, to obtain eigenmatrix xi。
In the step, eigenmatrix xi=[x1,x2,...,xn]=[| H (1,1) | ..., | H (Nt,Nr) |], wherein n=
Nt*Nr。
S104, the transmission error rates that different antennae combines in each channel are calculated according to channel state information, selected each
The smallest antenna combination of the bit error rate marks the channel in channel, and constitutes marker samples data set.
In the step, the problem of can solve classification because of machine learning and predict, deep learning is in machine learning
A method of based on representative learning is carried out to data, and the problem of day line options, can also regard polytypic problem as, pass through
The channel state information of each link of STSKMIMO model is extracted as feature samples, then with data in each transmission
The smallest antenna combination of transmission error rates is key index marked channels, constitutes marker samples data set, so that it may with depth
The method of habit goes to train the sample data in marker samples data set.
Specifically, referring to Fig. 2, which includes following sub-step S1041-S1043:
S1041, the transmission error rates that different antennae combines in each channel are calculated according to channel state information.
In the step, when signal is transmitted in different channels using different antennae combination, calculate every in each channel
The transmission error rates of one antenna combination.
S1042, regard each antenna combination as a kind of label, wherein label classification be y (y=1,2 ..., s).
In the step,M, n are the antenna amount selected in transmitting terminal and receiving end respectively.
S1043, the corresponding label classification y of the smallest antenna combination of the bit error rate in each channel is obtained, and with the tag class
Other y marks corresponding channel information eigenmatrix xi, and constitute marker samples data set (xi,yi)={ (x1,y1),(x2,
y2),...,(xm,ym)}。
In the step, using the bit error rate as the key index to channel label label, transmission error code in each channel is found out
The smallest antenna combination of rate, and the label classification corresponding to the smallest antenna combination of transmission error rates in channel goes to mark this
The channel information eigenmatrix of channel, all channel information eigenmatrixes and corresponding label classification constitute a marker samples number
According to collection.
Partial data in S105, selection marker samples data set, and with the propagated forward algorithm of neural network and reversely
Propagation algorithm is iterated training, to obtain the neural network classification model for being used for day line options.
In the step, gone to train the sample data in marker samples data set with the method for deep learning, and training
Cheng Zhong, the problem of in order to avoid over-fitting, the method that regularization can be used in the training process, by limiting the size of weight,
Prevent model is from the random noise that is arbitrarily fitted in training data, it will be appreciated that ground, the model is using existing deep layer nerve
Network model, details are not described herein.Specifically, referring to Fig. 3, which includes following sub-step S1051-S1053:
It is other pre- to obtain tag class by propagated forward algorithm for partial data in S1051, selection marker samples data set
Measured valueAnd calculate predicted valueWith true value yiBetween loss function.
Wherein, w and b is the parameter for wanting neural network model to need to learn, and the loss function is
S1052, each ginseng is updated by the gradient of back-propagation algorithm calculating parameter w and b, then by gradient descent method
Number, to minimize loss function.
In the step, the gradient of the parameter w and b passes through respectivelyCalculate institute
, and pass through wi+1=wi-ηdwi,bi+1=bi- η db removes undated parameter w and b, wherein η=0.001, is learning rate.Specifically,
This method minimizes loss function using gradient descent method, and what is chosen by training every time is part in marker samples data set
Data then calculate the loss function of sub-fraction training data every time, and training speed is than comparatively fast, being greatly reduced needed for convergence
The number of iterations, while convergent result can be made to be more nearly the effect that gradient declines.
If S1053, loss function converge to minimum value or repetitive exercise reaches preset times, terminate training, obtains
Neural network classification model for day line options.
S106, the channel for needing day line options is input in neural network classification model, to obtain corresponding antenna combination.
In the step, selection channel is gone to pass using the neural network classification model for day line options that above-mentioned steps are established
Defeated optimal antenna combination.A label classification is obtained after needing the channel of day line options to input neural network classification model,
This label classification is mapped as a receiving antenna and emitting antenna combination [a, b].Assuming that need to select m root transmitting antenna with
N root receiving antenna, wherein a indicates the receiving antenna index vector selected, length m；B indicates the transmitting antenna selected
Index vector, length n, the then channel gone out using deep learning algorithms selection
In some other embodiments, the antenna selecting method based on deep learning further includes having: calculating acquisition
The bit error rate of antenna combination and the size of channel capacity.Specifically, the bit error rate and channel of the antenna combination of the acquisition are calculated
The size of capacity is described as follows:
(1) the dispersion matrix A of STSK system is generatedQ, Q=1,2,3,4, Q indicate the number of dispersion matrix.
(2) precoding is carried out using bit data of the Space-Time ShiftKeying method for precoding to input, this
Invention uses QPSK modulation system, therefore 4 bit datas can be regarded as to one group, and the first two input bit is as dispersion matrix
AQIndex, latter two input bit carry out QPSK modulate to obtain QPSK symbol s, then this 4 input bits can be precoded into
One transmission block S, wherein S=AQs.Such as: when the bit data of input is [1001] bit=, because of two input bits one
Share 4 kinds of combinations: 00,01,10 and 11,00 represents A1, and 01 represents A2, and 10 represent A3, and 11 represent A4, then inputs ratio by front two
A should be selected known to spy3Carry out precoding；Two bits can be modulated to s=1-j afterwards (j represents plural number).
(3) channel by being gone out using deep learning algorithms selectionKnown to the mode of signal beIts
Middle V represents additive white Gaussian noise.It is legal there are one in recipient when recipient is decoded using maximal possibility estimation
Codeword table:If the number for dispersing matrix in the present invention is 4, Q=4；L
Indicate modulation system, L~PSK, such as L=2 in BPSK (2PSK)；The present invention uses QPSK (4PSK), so L=4.It then can be with
The mode of signal is rewritten are as follows:Wherein y=vector [Y],I indicates unit matrix,
V=vector [V], θ=[vector [A1]…vector[AQ]]。
(4) according to the mode formula of above-mentioned rewriting, maximum Likelihood can be used and estimate obtained signal,
And its position is found in legal-code table, that is, pass through formulaInFind conjunction
Corresponding bit in method codeword table, then by channel decode, revert to information bit r_bit, by the information bit r_bit with
Transmitted bit bit is compared, to obtain the bit number of mistake, to calculate the bit error rate.And it can be by formulaComputing system channel capacity, I therein are unit matrix, and P is that (this system will for transmission power
Power normalization, i.e. P=1),Indicate noise power,
Fig. 4 is to be transmitted using antenna selecting method of the invention and using other antenna selecting methods to signal
BER (Bit Error Ratio, bit error probability) curve graph, the value of BER are a search time interim error bit
Number and transmission total number of bits ratio, usually as a percentage.Fig. 5 be using antenna selecting method of the invention and
The power system capacity curve graph that signal is transmitted using other antenna selecting methods, referring to Fig. 4 and Fig. 5 it is found that using this hair
Bright antenna selecting method based on deep learning and signal is transmitted using optimal antenna selection method (traversal)
BER curve is almost overlapped, and power system capacity curve is also almost overlapped, i.e., its bit error rate and power system capacity are close.Because of traditional day
Line options algorithm it is optimal be traversal Antenna Selection Algorithem, then method of the invention can guarantee communication system have higher biography
In the case where defeated reliability and power system capacity, the selection complexity of system is greatly reduced.
In conclusion the present invention can guarantee communication system have higher transmission reliability and power system capacity in the case where,
The selection complexity of system, i.e. the channel status letter by extracting each link of STSK MIMO model is greatly reduced
Breath is used as feature samples, and calculates the transmission error rates that different antennae combines in each channel according to channel state information, uses
The smallest antenna combination marked channels of the bit error rate, and marker samples data set is constituted, then using the method training mark of deep learning
Remember the partial data that sample data is concentrated, the neural network classification model for day line options can be obtained, if there are new data
Information input can then obtain optimal antenna combination quickly, reduce the work of system into the neural network classification model
Amount, reduces the selection complexity of system.
The above description is only a preferred embodiment of the present invention, rather than does limitation in any form to the present invention.This field
Technical staff can impose various equivalent changes and improvement, all institutes within the scope of the claims on the basis of the above embodiments
The equivalent variations or modification done, should all fall under the scope of the present invention.
Claims (6)
1. a kind of antenna selecting method based on deep learning, which is characterized in that the antenna selecting method includes:
Obtain the channel state information matrix of each channel in STSK MIMO communication system model；
The channel state information matrix is carried out to take amplitude operation, to obtain eigenmatrix xi；
The transmission error rates that different antennae combines in each channel are calculated according to channel state information, select error code in each channel
The smallest antenna combination of rate marks the channel, and constitutes marker samples data set；
Choose the partial data in marker samples data set, and with the propagated forward algorithm of neural network and back-propagation algorithm into
Row iteration training, to obtain the neural network classification model for being used for day line options；
The channel for needing day line options is input in neural network classification model, to obtain corresponding antenna combination.
2. as described in claim 1 based on the antenna selecting method of deep learning, it is characterised in that: described according to channel shape
State information calculates the transmission error rates that different antennae combines in each channel, selects the smallest antenna sets of the bit error rate in each channel
It closes and marks the channel, and constitute marker samples data set and include:
The transmission error rates that different antennae combines in each channel are calculated according to channel state information；
Regard each antenna combination as a kind of label, wherein label classification be y (y=1,2 ..., s)；
The corresponding label classification y of the smallest antenna combination of the bit error rate in each channel is obtained, and is marked and is corresponded to the label classification y
Channel information eigenmatrix xi, and constitute marker samples data set (xi,yi)={ (x1,y1),(x2,y2),...,(xm,ym)}。
3. as described in claim 2 based on the antenna selecting method of deep learning, it is characterised in that: the selection marks sample
The partial data that notebook data is concentrated, and it is iterated training with the propagated forward algorithm and back-propagation algorithm of neural network, with
It obtains and includes: for the neural network classification model of day line options
The partial data in marker samples data set is chosen, the other predicted value of tag class is obtained by propagated forward algorithmAnd calculate predicted valueWith true value yiBetween loss function；
Each parameter is updated by the gradient of back-propagation algorithm calculating parameter w and b, then by gradient descent method, with minimum
Change loss function；
If loss function converges to minimum value or repetitive exercise reaches preset times, terminate training, obtains and be used for day line selection
The neural network classification model selected.
4. as described in claim 3 based on the antenna selecting method of deep learning, it is characterised in that: the loss function is
5. as described in claim 1 based on the antenna selecting method of deep learning, it is characterised in that: the acquisition STSK
In MIMO communication system model before the channel state information matrix of each channel, further includes:
STSK MIMO communication system model is built, and the channel in the communication system is carried out using maximum Likelihood
Estimation.
6. as described in claim 1 based on the antenna selecting method of deep learning, it is characterised in that: described to need day line selection
The channel selected is input in neural network classification model, after obtaining corresponding antenna combination, further includes: calculate the day of acquisition
The bit error rate of line combination and the size of channel capacity.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811503785.9A CN109660287B (en) | 2018-12-10 | 2018-12-10 | Antenna selection method based on deep learning |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201811503785.9A CN109660287B (en) | 2018-12-10 | 2018-12-10 | Antenna selection method based on deep learning |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN109660287A true CN109660287A (en) | 2019-04-19 |
| CN109660287B CN109660287B (en) | 2022-02-11 |
Family
ID=66113099
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201811503785.9A Active CN109660287B (en) | 2018-12-10 | 2018-12-10 | Antenna selection method based on deep learning |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN109660287B (en) |
Cited By (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110113119A (en) * | 2019-04-26 | 2019-08-09 | 国家无线电监测中心 | A kind of Wireless Channel Modeling method based on intelligent algorithm |
| CN112202510A (en) * | 2020-09-27 | 2021-01-08 | 广州大学 | Signal detection method based on regularization stream neural network |
| CN112380903A (en) * | 2020-10-14 | 2021-02-19 | 东北电力大学 | Human activity identification method based on WiFi-CSI signal enhancement |
| CN112737650A (en) * | 2020-12-30 | 2021-04-30 | 玉林师范学院 | MIMO system transmitting antenna selection method based on machine learning |
| WO2023231639A1 (en) * | 2022-06-02 | 2023-12-07 | 中兴通讯股份有限公司 | Transmit antenna selection method and apparatus, and computer device and computer storage medium |
Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080132281A1 (en) * | 2006-08-21 | 2008-06-05 | Byoung-Hoon Kim | Approach to a unified su-mimo/mu-mimo operation |
| CN107332598A (en) * | 2017-06-26 | 2017-11-07 | 浙江理工大学 | A kind of precoding of mimo system joint and antenna selecting method based on deep learning |
| CN107592144A (en) * | 2017-08-29 | 2018-01-16 | 广西师范大学 | EH MIMO collection of energy and the node antennas system of selection of multiple antenna communication and device |
| CN108923828A (en) * | 2018-07-06 | 2018-11-30 | 西北工业大学 | A kind of emitting antenna selecting method of the MIMO tapping channel based on deeply study |
-
2018
- 2018-12-10 CN CN201811503785.9A patent/CN109660287B/en active Active
Patent Citations (4)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US20080132281A1 (en) * | 2006-08-21 | 2008-06-05 | Byoung-Hoon Kim | Approach to a unified su-mimo/mu-mimo operation |
| CN107332598A (en) * | 2017-06-26 | 2017-11-07 | 浙江理工大学 | A kind of precoding of mimo system joint and antenna selecting method based on deep learning |
| CN107592144A (en) * | 2017-08-29 | 2018-01-16 | 广西师范大学 | EH MIMO collection of energy and the node antennas system of selection of multiple antenna communication and device |
| CN108923828A (en) * | 2018-07-06 | 2018-11-30 | 西北工业大学 | A kind of emitting antenna selecting method of the MIMO tapping channel based on deeply study |
Non-Patent Citations (4)
| Title |
|---|
| WEI JIANG: "Neural Network-Based Channel Prediction and Its Performance in Multi-Antenna Systems", 《2018 IEEE 88TH VEHICULAR TECHNOLOGY CONFERENCE (VTC-FALL)》 * |
| YOUBING HU: "Optimal Transmit Antenna Selection Strategy for MIMO Wiretap Channel Based on Deep Reinforcement Learning", 《2018 IEEE/CIC INTERNATIONAL CONFERENCE ON COMMUNICATIONS IN CHINA (ICCC)》 * |
| ZHIBIN XIE: "Effective antenna selection in MIMO systems under spatial correlated fading", 《2008 7TH WORLD CONGRESS ON INTELLIGENT CONTROL AND AUTOMATION》 * |
| 董健: "基于改进反向传播神经网络代理模型的快速多目标天线设计", 《电子与信息学报》 * |
Cited By (7)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| CN110113119A (en) * | 2019-04-26 | 2019-08-09 | 国家无线电监测中心 | A kind of Wireless Channel Modeling method based on intelligent algorithm |
| CN112202510A (en) * | 2020-09-27 | 2021-01-08 | 广州大学 | Signal detection method based on regularization stream neural network |
| CN112380903A (en) * | 2020-10-14 | 2021-02-19 | 东北电力大学 | Human activity identification method based on WiFi-CSI signal enhancement |
| CN112380903B (en) * | 2020-10-14 | 2024-02-02 | 东北电力大学 | Human body activity recognition method based on WiFi-CSI signal enhancement |
| CN112737650A (en) * | 2020-12-30 | 2021-04-30 | 玉林师范学院 | MIMO system transmitting antenna selection method based on machine learning |
| CN112737650B (en) * | 2020-12-30 | 2022-02-22 | 玉林师范学院 | MIMO system transmitting antenna selection method based on machine learning |
| WO2023231639A1 (en) * | 2022-06-02 | 2023-12-07 | 中兴通讯股份有限公司 | Transmit antenna selection method and apparatus, and computer device and computer storage medium |
Also Published As
| Publication number | Publication date |
|---|---|
| CN109660287B (en) | 2022-02-11 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN109660287A (en) | A kind of antenna selecting method based on deep learning | |
| Honkala et al. | DeepRx: Fully convolutional deep learning receiver | |
| JP5265660B2 (en) | Fast square root algorithm for MIMO equalization | |
| US8537922B2 (en) | Methods and systems for providing feedback for beamforming and power control | |
| CN102598523B (en) | Tomlinson harashima precoding with additional receiver processing in a multi-user multiple-input multiple-output wireless transmission system | |
| CN105591717A (en) | Low-complexity detection method for MIMO-OFDM system with subcarrier index modulation | |
| Mandloi et al. | Congestion control based ant colony optimization algorithm for large MIMO detection | |
| Song et al. | Benchmarking end-to-end learning of MIMO physical-layer communication | |
| Alawad et al. | A deep learning-based detector for IM-MIMO-OFDM | |
| CN103312422B (en) | Signal blind detection method based on artificial fish swarm algorithm | |
| CN106452662B (en) | A kind of method for precoding and device | |
| CN112039568A (en) | Large-scale MIMO system cross-layer design scheme based on incomplete channel state information | |
| Valente et al. | LR-aided MIMO detectors under correlated and imperfectly estimated channels | |
| KR102355383B1 (en) | Deep learning-based signal detection technique for high reliability in massive MIMO systems | |
| Lee | Minimum Euclidean distance evaluation using deep neural networks | |
| Nandi et al. | Avenues to improve channel estimation using optimized CP in STBC coded MIMO-OFDM systems—a global optimization approach | |
| US8369461B2 (en) | Method and arrangement relating to telecommunications | |
| US6693976B1 (en) | Method of wireless communication using structured unitary space-time signal constellations | |
| US11387870B2 (en) | MIMO detector selection | |
| Ma et al. | Design of DBN based Demodulator in Underwater Wireless Optical Communications | |
| Ramasamy et al. | Evolutionary gravitational Neocognitron neural network based differential spatial modulation detection scheme for uplink multiple user huge MIMO systems | |
| Katwal et al. | Improved Channel Equalization using Deep Reinforcement Learning and Optimization | |
| CN107493123A (en) | Low complex degree detection method based on precoding auxiliary generalized orthogonal spatial modulation | |
| CN101640583B (en) | Method for transmitting preprocessing | |
| CN108429574B (en) | Method for selecting transmitting antenna of large-scale MIMO system |
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 |