CN105790813B - Code book selection method based on deep learning under a kind of extensive MIMO - Google Patents
Code book selection method based on deep learning under a kind of extensive MIMO Download PDFInfo
- Publication number
- CN105790813B CN105790813B CN201610327115.0A CN201610327115A CN105790813B CN 105790813 B CN105790813 B CN 105790813B CN 201610327115 A CN201610327115 A CN 201610327115A CN 105790813 B CN105790813 B CN 105790813B
- Authority
- CN
- China
- Prior art keywords
- channel
- code book
- neural network
- information
- pilot
- 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
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
- H04B7/0456—Selection of precoding matrices or codebooks, e.g. using matrices antenna weighting
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The present invention relates to the code book selection methods based on deep learning under a kind of extensive MIMO (Multiple-Input Multiple-Output, multiple-input and multiple-output), belong to wireless communication technology field.This method includes：The pilot frequency information in collecting test area builds pilot training sequence, and then obtains pilot tone training sample；Neural network iterative learning is carried out to pilot tone training sample, obtains final network weight weight values；According to the channel of the neural network output after study, optimal code word is selected from complete code book.Unknown area and test section are subjected to channel information matching later, obtain its wireless channel, and then obtain code word corresponding with wireless channel.The present invention can effectively, accurately and rapidly establish wireless channel model and code book is inquired, and avoid the channel estimation of unknown area and greatly reduce the complexity of unknown area channel selection code book.
Description
Technical field
The invention belongs to wireless communication technology fields, are related to a kind of extensive MIMO (Multiple-Input
Multiple-Output, multiple-input and multiple-output) under the code book selection method based on deep learning.
Background technology
Any one communication system, channel are essential component parts.Wireless channel is typical " variable-parameter channel ",
The characteristic and communication environments of wireless channel, such as：Landform, atural object, climate characteristic, electromagnetic interference situation, communication body movement speed and
The frequency range used etc. is closely related.Communication capacity, the service quality (Quality of Service, QoS) of wireless communication system
It is closely related Deng the quality all with radio channel performance.Therefore, to high quality, great Rong as far as possible on limited frequency spectrum resource
Amount transmits useful information, it is necessary to the characteristic for grasping wireless channel well, especially in the big data epoch, while will also be as far as possible
Ensure that the error rate of the wireless channel obtained is smaller.
Wireless channel model is after having to wireless propagation environment and propagation characteristic and fully understanding, to one of wireless channel
Abstractdesription can reflect some critical natures of wireless propagation environment well.The foundation of wireless channel model depends on
Channel detection.Currently, the existing method for establishing wireless channel propagation model has：Statistical model, deterministic models and half are really
Qualitative model.
But the above-mentioned existing method for establishing wireless channel propagation model is according to electromagnetism there are some disadvantages, such as these methods
Theory of wave propagation analyzes the method for building up for obtaining wireless channel model under some simplified conditions.And practical mobile circumstances
It is ever-changing, significantly limits the application range of these notional results, some specific environment, single can only be directed to
Link carries out, and it is not comprehensive enough accurate to be described to the characteristic of channel, the directionality characteristic of channel under high-speed mobile scene.Another party
Face, the method for building up of existing channel model need fully to excavate the causality of sending and receiving end.Its letter by acquiring sending and receiving end
Number, analysis sending and receiving signal establishes the causality of transmitting-receiving two-end.Because of the Finite Samples of acquisition, and based on the assumption that condition, so as to get
Result can be impacted.In the small data epoch, computer capacity is insufficient, and most of analysis is only limitted to seek simply linear close
System.
In extensive mimo system, because number of antennas is huge so that channel battle array H dimensions become larger rapidly, based on the pre- of non-code book
Coding techniques is no longer applicable in, and based on the Linear Precoding of code book at focus of attention.Currently used generation code book
Method have：Based on Grassmannian subspace packing, DFT etc..But the former is generally being looked for most with exhaustive search
Excellent code word, such as random search alternately predict that the computation burden of Selwyn Lloyd iterative algorithm, these algorithms will be with transmitting antenna number
Increase and increased dramatically.And DFT provides high chordal distance, but average error rate between precoding vectors with the mode of system
Easily influence is received when transmitting antenna is by high spatial correlation.In view of the above problems, a kind of low calculation amount of proposition, anti-space
The method for precoding of correlation has become active demand.
Invention content
In view of this, the purpose of the present invention is to provide a kind of extensive MIMO (Multiple-Input Multiple-
Output, multiple-input and multiple-output) under the code book selection method based on deep learning, this method can effectively establish wireless channel
Propagation model and code book inquiry.
In order to achieve the above objectives, the present invention provides the following technical solutions：
Code book selection method based on deep learning under a kind of extensive MIMO, this approach includes the following steps：
S1：Information collection step：By the pilot frequency information of user terminal in information acquisition system collecting test area；
S2：Obtain training sample：Pilot training sequence is built according to pilot frequency information, and then obtains pilot tone training sample；
S3：Initialize neural network：Initialize neural network model parameter；
S4：Neural network learning：Neural network deep learning is carried out by pilot tone training sample, obtains final network weight
Value；
S5：Construct complete code book：With improved DFT, (Discrete Fourier Transform, discrete fourier become
Changing) method construct is suitble to the code books of all channel status；
S6：Codeword selection：According to the channel of the neural network output after study, codeword selection is carried out from complete code book；
S7：The foundation of correlativity：It is related between characteristics of radio channels information to build the pilot frequency information of the test section
Relationship；
S8：Channel matched step：The channel of unknown area is matched with existing wireless channel, and then selects the letter of unknown area
The corresponding code word in road.
Further, in step sl, four classes are divided into test when carrying out information collection：Suburb macrocell
(suburban macro), urban district macrocell (urban macro, UM-a), urban microcell (urban micro, UM-i) and
High-speed Circumstance (high rise scenario).
Further, in step s3, the initialization neural network model parameter specifically includes：Learning rate η, bias δ,
The weight coefficient ω of input layer i-node and hidden layer j nodesijThe weight coefficient of ∈ (0,1), hidden layer j and output node layer l
ωjl∈ (0,1), wherein i, j, k ∈ N+ and ∑ | ω |=M (M is constant), maximum iteration lmax, Initial value e=0, god
Threshold function table, linear function or Sigmoid functions are used through first activation primitive f ().
Further, in step s 4, the neural network deep learning specifically includes：
S41：Inputs of the pilot tone training sample P as neural network, H=[H0,H1,...,HN] be neural network estimation
Desired value,For the estimation output valve of neural network；
S42：Parameter is carried out by error, maximum iteration and the weighted value constraints between model output and desired value
Depth is trained, until obtaining meeting required precision；
S43：It often carries out once, iterations add 1 i.e. l=l+1；As iterations l≤lmaxOr e (l)≤τmaxWhen terminate
It trains, otherwise return to step S42；
S44：The weight coefficient of target update is obtained after step S41, S42, S43；After the completion of the study stage, neural network
Estimated using the pilot tone P of test sectionAnd it willIt is stored into the Shark databases based on Spark clusters, Shark databases
Provide the inquiry service of channel information to the user.
Further, in step s 5, described constructed with improved DFT method is suitble to the code book of all channel status as follows：
F=WFDFT
Wherein, W (∈ Mt×Mt) it is unitary matrice, meet U=W ∑s VH(U∈(Mt×Mt), element obeys CN (0,1)).
Further, in step s 6, the codeword selection includes：After the completion of neural network learning, the output of neural network
Value is the channel estimated using pilot tone training sample；Codeword selection is carried out according to code selection criterion, and the optimum code selected
Word is placed in Shark databases, provides the inquiry service of codeword information to the user.
Further, in the step s 7, it by channel information, builds between the pilot frequency information of test section and characteristics of radio channels information
Correlativity, specifically include：According to the pilot correlation feature of test section, the pilot tone in test section is divided into multiple with generation
The reference pilot pattern of table；The channel letter of the test section internal reference pilot frequency design is obtained by the wireless channel model of test section
Breath obtains the corresponding channel characteristics of each reference pattern, and the corresponding channel characteristics of reference pattern is stored in reference channel letter
It ceases in database.
Further, in step s 8, the channel by unknown area is matched with existing wireless channel, and then is selected unknown
The corresponding code word of channel in area specifically includes：
S81：Characteristic matching is carried out according to the reference pilot pattern in the pilot frequency information and test section of the zone of ignorance；
S82：Judge whether the similarity between the pilot frequency information of zone of ignorance and the reference pilot pattern characteristics of test section is small
In the threshold value of setting, if the successful match less than if；Otherwise, reference pilot pattern is chosen again, until meeting the threshold less than setting
Value；
S83：After the pilot frequency design characteristic matching in the pilot frequency information and test section in unknown area is successful, by reference channel
The corresponding channel characteristics of the pattern are determined as the channel characteristics of unknown area in information database, and channel characteristics are integrated, are obtained
To the wireless channel in the unknown area；
S84：Optimal code word is obtained in the Shark databases for storing optimal codeword information according to the wireless channel, and
It is fed back to the base station (BS) in the unknown area.
The beneficial effects of the present invention are：The present invention can effectively, accurately and rapidly establish wireless channel model and code book is looked into
It askes, avoids the channel estimation of unknown area and greatly reduce the complexity of unknown area channel selection code book.
Description of the drawings
In order to keep the purpose of the present invention, technical solution and advantageous effect clearer, the present invention provides following attached drawing and carries out
Explanation：
Fig. 1 is the flow diagram of the method for the invention；
Fig. 2 is the wireless channel Establishing process figure of test section；
Fig. 3 is the step flow chart of neural network deep learning；
Fig. 4 is codebook precoding method flow diagram under extensive MIMO；
Matching Model flow charts of the Fig. 5 between unknown area pilot frequency information and characteristics of radio channels information.
Specific implementation mode
Below in conjunction with attached drawing, the preferred embodiment of the present invention is described in detail.
Fig. 1 is the flow diagram of the method for the invention, as shown, this method specifically includes following steps：
S1：Information collection step：By the pilot frequency information of user terminal in information acquisition system collecting test area；
S2：Obtain training sample：Pilot training sequence is built according to pilot frequency information, and then obtains pilot tone training sample；
S3：Initialize neural network：Initialize neural network model parameter；
S4：Neural network learning：Neural network deep learning is carried out by pilot tone training sample, obtains final network weight
Value；
S5：Construct complete code book：With improved DFT, (Discrete Fourier Transform, discrete fourier become
Changing) method construct is suitble to the code books of all channel status；
S6：Codeword selection：According to the channel of the neural network output after study, codeword selection is carried out from complete code book；
S7：The foundation of correlativity：It is related between characteristics of radio channels information to build the pilot frequency information of the test section
Relationship；
S8：Channel matched step：The channel of unknown area is matched with existing wireless channel, and then selects the letter of unknown area
The corresponding code word in road.
Fig. 2 is the wireless channel Establishing process figure of test section, including：
Four classes are divided into test：Suburb macrocell, urban district macrocell, urban microcell, High-speed Circumstance, by taking UM-i as an example
Information collection is carried out, remaining is similar；
Four classes are divided into test：Suburb macrocell (suburban macro), urban district macrocell (urban macro,
UM-a), urban microcell (urban micro, UM-i), High-speed Circumstance (high rise scenario) are with UM-i here
Example is analyzed, remaining is similar；
The pilot frequency information of user terminal in information acquisition system collecting test area；
Pilot training sequence is built by pilot frequency information, and then obtains pilot tone training sample；
Input of the pilot tone training sample as neural network carries out neural network deep learning, god is obtained after study
Through network output valve, the channel that as estimates.
Fig. 3 is the step flow chart of neural network deep learning, and the neural network deep learning specifically includes：
S41：Inputs of the pilot tone training sample P as neural network, H=[H0,H1,...,HN] be neural network estimation
Desired value,For the estimation output valve of neural network；
S42：Parameter is carried out by error, maximum iteration and the weighted value constraints between model output and desired value
Depth is trained, until obtaining meeting required precision；
S43：It often carries out once, iterations add 1 i.e. l=l+1；As iterations l≤lmaxOr e (l)≤τmaxWhen terminate
It trains, otherwise return to step S42；
S44：The weight coefficient of target update is obtained after step S41, S42, S43；After the completion of the study stage, neural network
Estimated using the pilot tone P of test sectionAnd it willIt is stored into the Shark databases based on Spark clusters, Shark databases
Provide the inquiry service of channel information to the user.
Fig. 4 is codebook precoding method flow diagram under extensive MIMO, constructs code book by improved DFT method, and by code
Originally it is placed on transmitting-receiving two-end：
F=WFDFT
Wherein, W (∈ Mt×Mt) it is unitary matrice, meet U=W Σ VH(U∈(Mt×Mt), element obeys CN (0,1)).
The index of optimal code word that neural network exports in the codebook is fed back to base station end (BS) by user terminal, and will be optimal
In the information storage of code word to Shark databases, which provides the inquiry service of optimal code word to the user.
Matching Model flow charts of the Fig. 5 between unknown area pilot frequency information and characteristics of radio channels information, in the present embodiment,
Detailed process is as follows：
According to the pilot correlation feature of the test sections UM-i, the pilot tone in test section is divided into multiple representative ginsengs
Examine pilot frequency design；
The channel information that the test section internal reference pilot frequency design is obtained by the wireless channel of the test sections UM-i, obtains each ginseng
The corresponding channel characteristics of pattern are examined, and the corresponding channel characteristics of reference pattern are stored in reference channel information database.
Characteristic matching is carried out according to the reference pilot pattern in the pilot frequency information of the unknown area and the test sections UM-i；
Judge whether the similarity between the pilot frequency information of zone of ignorance and the reference pilot pattern characteristics of test section is less than to set
Fixed threshold value, if the successful match less than if；Otherwise, reference pilot pattern is chosen again, until meeting the threshold value less than setting；
After the pilot frequency design characteristic matching in the pilot frequency information and test section in unknown area is successful, by reference channel information
The corresponding channel characteristics of the pattern are determined as the channel characteristics of unknown area in database, and channel characteristics are integrated, are somebody's turn to do
Wireless channel in unknown area；
According to the wireless channel, in the Shark databases for storing optimal codeword information, obtain optimal code word, and by its
Feed back to the BS in the unknown area.
Finally illustrate, preferred embodiment above is merely illustrative of the technical solution of the present invention and unrestricted, although logical
It crosses above preferred embodiment the present invention is described in detail, however, those skilled in the art should understand that, can be
Various changes are made to it in form and in details, without departing from claims of the present invention limited range.
Claims (8)
1. the code book selection method based on deep learning under a kind of extensive MIMO, it is characterised in that：This method includes following step
Suddenly：
S1：Information collection step：By the pilot frequency information of user terminal in information acquisition system collecting test area；
S2：Obtain training sample：Pilot training sequence is built according to pilot frequency information, and then obtains pilot tone training sample；
S3：Initialize neural network：Initialize neural network model parameter；
S4：Neural network learning：Neural network deep learning is carried out by pilot tone training sample, obtains final network weight weight values；
S5：Construct complete code book：With the sides improved DFT (Discrete Fourier Transform, discrete Fourier transform)
Method construction is suitble to the code book of all channel status；
S6：Codeword selection：According to the channel of the neural network output after study, codeword selection is carried out from complete code book；
S7：The foundation of correlativity：Build the correlativity between the pilot frequency information of the test section and characteristics of radio channels information；
S8：Channel matched step：The channel of unknown area is matched with existing wireless channel, and then selects the channel pair of unknown area
The code word answered.
2. the code book selection method based on deep learning under a kind of extensive MIMO according to claim 1, feature exist
In：In step sl, four classes are divided into test when carrying out information collection：Suburb macrocell (suburban macro), city
Area's macrocell (urban macro, UM-a), urban microcell (urban micro, UM-i) and High-speed Circumstance (high rise
scenario)。
3. the code book selection method based on deep learning under a kind of extensive MIMO according to claim 2, feature exist
In：In step s3, the initialization neural network model parameter specifically includes：Learning rate η, bias δ, input layer i-node
With the weight coefficient ω of hidden layer j nodesijThe weight coefficient ω of ∈ (0,1), hidden layer j and output node layer kjl∈ (0,1),
Wherein i, j, k ∈ N+, N+For positive integer, and ∑ | ω |=M (M is constant), maximum iteration lmax, Initial value e=0, god
Threshold function table, linear function or Sigmoid functions are used through first activation primitive f ().
4. the code book selection method based on deep learning under a kind of extensive MIMO according to claim 3, feature exist
In：In step s 4, the neural network deep learning specifically includes：
S41：Inputs of the pilot tone training sample P as neural network, channel battle array H=[H0,H1,...,HN] it is estimating for neural network
Desired value is counted,For the estimation output valve of neural network；
S42：By the worst error τ between model output and desired valuemax, maximum iteration and weighted value constraints joined
Several depth training, until obtaining meeting required precision；
S43：It often carries out once, iterations add 1 i.e. l=l+1；As iterations l≤lmaxOr e (l)≤τmax(τmaxMaximum is accidentally
Difference) when terminate to train, otherwise return to step S42；
S44：The weight coefficient of target update is obtained after step S41, S42, S43；After the completion of the study stage, neural network utilizes
The pilot tone P of test section estimatesAnd it willIt is stored into the Shark databases based on Spark clusters, Shark databases are
User provides the inquiry service of channel information.
5. the code book selection method based on deep learning under a kind of extensive MIMO according to claim 4, feature exist
In：In step s 5, described constructed with improved DFT method is suitble to the code book of all channel status as follows：
F=WFDFT
Wherein, FDFTFor Fourier transform code book, F is modified Fourier transform code book, MtTo emit number of days, W (∈ Mt×
Mt) it is unitary matrice, meet U=W ∑s VH(U∈(Mt×Mt) singular value decomposition, element u can be carried outkObey 0 mean value, 1 variance
Independent identically distributed multiple Gauss distribution, i.e. uk~CN (0,1)).
6. the code book selection method based on deep learning under a kind of extensive MIMO according to claim 5, feature exist
In：In step s 6, the codeword selection includes：After the completion of neural network learning, the output valve of neural network is to utilize to lead
The channel that frequency training sample estimates；Codeword selection is carried out according to code selection criterion, and the optimal code word selected is placed on Shark numbers
According to the inquiry service in library, providing codeword information to the user.
7. the code book selection method based on deep learning under a kind of extensive MIMO according to claim 6, feature exist
In：In the step s 7, by channel information, the correlativity between the pilot frequency information of test section and characteristics of radio channels information is built,
It specifically includes：According to the pilot correlation feature of test section, the pilot tone in test section is divided into multiple representative references
Pilot frequency design；The channel information that the test section internal reference pilot frequency design is obtained by the wireless channel model of test section, obtains each
The corresponding channel characteristics of reference pattern, and the corresponding channel characteristics of reference pattern are stored in reference channel information database.
8. the code book selection method based on deep learning under a kind of extensive MIMO according to claim 7, feature exist
In：In step s 8, the channel by unknown area is matched with existing wireless channel, and then the channel for selecting unknown area corresponds to
Code word specifically include：
S81：Characteristic matching is carried out according to the reference pilot pattern in the pilot frequency information and test section of the unknown area；
S82：Judge whether the similarity between the pilot frequency information of unknown area and the reference pilot pattern characteristics of test section is less than setting
Threshold value, if less than if successful match；Otherwise, reference pilot pattern is chosen again, until meeting the threshold value less than setting；
S83：After the pilot frequency design characteristic matching in the pilot frequency information and test section in unknown area is successful, by reference channel information
The corresponding channel characteristics of the pattern are determined as the channel characteristics of unknown area in database, and channel characteristics are integrated, are somebody's turn to do
Wireless channel in unknown area；
S84：According to the wireless channel, in the Shark databases for storing optimal codeword information, obtain optimal code word, and by its
Feed back to the base station (BS) in the unknown area.
Priority Applications (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610327115.0A CN105790813B (en) | 2016-05-17 | 2016-05-17 | Code book selection method based on deep learning under a kind of extensive MIMO |
Applications Claiming Priority (1)
| Application Number | Priority Date | Filing Date | Title |
|---|---|---|---|
| CN201610327115.0A CN105790813B (en) | 2016-05-17 | 2016-05-17 | Code book selection method based on deep learning under a kind of extensive MIMO |
Publications (2)
| Publication Number | Publication Date |
|---|---|
| CN105790813A CN105790813A (en) | 2016-07-20 |
| CN105790813B true CN105790813B (en) | 2018-11-06 |
Family
ID=56379965
Family Applications (1)
| Application Number | Title | Priority Date | Filing Date |
|---|---|---|---|
| CN201610327115.0A Active CN105790813B (en) | 2016-05-17 | 2016-05-17 | Code book selection method based on deep learning under a kind of extensive MIMO |
Country Status (1)
| Country | Link |
|---|---|
| CN (1) | CN105790813B (en) |
Families Citing this family (15)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US10552738B2 (en) * | 2016-12-15 | 2020-02-04 | Google Llc | Adaptive channel coding using machine-learned models |
| CN107332598B (en) * | 2017-06-26 | 2020-06-16 | 浙江理工大学 | MIMO system joint precoding and antenna selection method based on deep learning |
| CN108183736B (en) * | 2017-12-28 | 2021-02-23 | 北京邮电大学 | Transmitter codeword selection method and device based on machine learning and transmitter |
| CN108390706B (en) * | 2018-01-30 | 2020-10-27 | 东南大学 | Large-scale MIMO channel state information feedback method based on deep learning |
| CN108566257B (en) * | 2018-04-27 | 2020-08-18 | 电子科技大学 | Signal recovery method based on back propagation neural network |
| CN110611627B (en) * | 2018-06-15 | 2021-03-19 | 维沃移动通信有限公司 | Signal detection method and receiving end |
| CN108923828B (en) * | 2018-07-06 | 2019-06-07 | 西北工业大学 | A kind of emitting antenna selecting method of the MIMO tapping channel based on deeply study |
| CN109617584B (en) * | 2019-01-08 | 2021-12-21 | 南京邮电大学 | MIMO system beam forming matrix design method based on deep learning |
| CN110300075B (en) * | 2019-04-30 | 2020-10-02 | 北京科技大学 | Wireless channel estimation method |
| CN110378467A (en) * | 2019-06-17 | 2019-10-25 | 浙江大学 | A kind of quantization method for deep learning network parameter |
| CN111447620B (en) * | 2020-03-19 | 2022-05-17 | 重庆邮电大学 | Millimeter wave heterogeneous network resource allocation joint optimization method |
| CN112803976B (en) * | 2020-12-24 | 2022-07-08 | 浙江香农通信科技有限公司 | Large-scale MIMO precoding method and system and electronic equipment |
| CN113114323B (en) * | 2021-04-22 | 2022-08-12 | 桂林电子科技大学 | Signal receiving method of MIMO system |
| CN113242069B (en) * | 2021-05-10 | 2022-06-17 | 东南大学 | Codebook design method based on neural network |
| CN113660015B (en) * | 2021-08-11 | 2022-08-16 | 东南大学 | Online wireless channel acquisition optimization method under assistance of environment knowledge base |
Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2005031501B1 (en) * | 2003-09-22 | 2008-10-23 | Kim Hyeung-Yun | Sensors and systems for structural health monitoring |
| CN102739383A (en) * | 2012-05-28 | 2012-10-17 | 上海交通大学 | Method for allocating union resource based on limited feedback OFDM-AF (Orthogonal Frequency Division Multiplexing-Audio Frequency) system |
| CN104618045A (en) * | 2015-01-27 | 2015-05-13 | 北京交通大学 | Collected data-based wireless channel transmission model establishing method and system |
| CN104993914A (en) * | 2015-05-18 | 2015-10-21 | 北京交通大学 | Wireless channel estimation method and device based on wireless environment map |
| WO2016041463A1 (en) * | 2014-09-18 | 2016-03-24 | Huawei Technologies Co., Ltd. | Common broadcast channel low papr signaling in massive mimo systems |
Family Cites Families (1)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| US9231676B2 (en) * | 2014-05-12 | 2016-01-05 | Nokia Solutions And Networks Oy | Low effort massive MIMO antenna arrays and their use |
-
2016
- 2016-05-17 CN CN201610327115.0A patent/CN105790813B/en active Active
Patent Citations (5)
| Publication number | Priority date | Publication date | Assignee | Title |
|---|---|---|---|---|
| WO2005031501B1 (en) * | 2003-09-22 | 2008-10-23 | Kim Hyeung-Yun | Sensors and systems for structural health monitoring |
| CN102739383A (en) * | 2012-05-28 | 2012-10-17 | 上海交通大学 | Method for allocating union resource based on limited feedback OFDM-AF (Orthogonal Frequency Division Multiplexing-Audio Frequency) system |
| WO2016041463A1 (en) * | 2014-09-18 | 2016-03-24 | Huawei Technologies Co., Ltd. | Common broadcast channel low papr signaling in massive mimo systems |
| CN104618045A (en) * | 2015-01-27 | 2015-05-13 | 北京交通大学 | Collected data-based wireless channel transmission model establishing method and system |
| CN104993914A (en) * | 2015-05-18 | 2015-10-21 | 北京交通大学 | Wireless channel estimation method and device based on wireless environment map |
Also Published As
| Publication number | Publication date |
|---|---|
| CN105790813A (en) | 2016-07-20 |
Similar Documents
| Publication | Publication Date | Title |
|---|---|---|
| CN105790813B (en) | Code book selection method based on deep learning under a kind of extensive MIMO | |
| Zhang et al. | NAS-AMR: Neural architecture search-based automatic modulation recognition for integrated sensing and communication systems | |
| WO2020253690A1 (en) | Deep learning beam domain channel estimation method based on approximate message passing algorithm | |
| CN109617584B (en) | MIMO system beam forming matrix design method based on deep learning | |
| Guo et al. | CAnet: Uplink-aided downlink channel acquisition in FDD massive MIMO using deep learning | |
| Elijah et al. | Intelligent massive MIMO systems for beyond 5G networks: An overview and future trends | |
| CN113595666B (en) | Channel prediction method and system of large-scale MIMO system based on time-frequency joint correlation | |
| Jha et al. | Online downlink multi-user channel estimation for mmWave systems using Bayesian neural network | |
| Wang et al. | Jointly learned symbol detection and signal reflection in RIS-aided multi-user MIMO systems | |
| Kim et al. | Massive data generation for deep learning-aided wireless systems using meta learning and generative adversarial network | |
| Mai et al. | A UAV air-to-ground channel estimation algorithm based on deep learning | |
| Li et al. | A variational learning approach for concurrent distance estimation and environmental identification | |
| Chen et al. | Deep Learning Based Multi-User Positioning in Wireless FDMA Cellular Networks | |
| Sun et al. | Environment information-based channel prediction method assisted by graph neural network | |
| CN116634358A (en) | Terminal positioning method and device and nonvolatile storage medium | |
| CN114915526B (en) | Communication signal modulation identification method, device and system | |
| Haider et al. | GAN-based Channel Estimation for IRS-aided Communication Systems | |
| Boas et al. | Machine learning for CSI recreation in the digital twin based on prior knowledge | |
| Bano et al. | A federated channel modeling system using generative neural networks | |
| Li et al. | Piecewise-drl: Joint beamforming optimization for ris-assisted mu-miso communication system | |
| Ma et al. | Green wireless local area network received signal strength dimensionality reduction and indoor localization based on fingerprint algorithm | |
| Li et al. | CWGAN-Based Channel Modeling of Convolutional Autoencoder-Aided SCMA for Satellite-Terrestrial Communication | |
| Yang et al. | Deep-learning-based bouncing-order prediction for propagation channel characterization using graph modeling | |
| Baur et al. | Evaluation Metrics and Methods for Generative Models in the Wireless PHY Layer | |
| Gutierrez-Estevez et al. | Hybrid model and data driven algorithm for online learning of any-to-any path loss maps |
Legal Events
| Date | Code | Title | Description |
|---|---|---|---|
| C06 | Publication | ||
| PB01 | Publication | ||
| C10 | Entry into substantive examination | ||
| SE01 | Entry into force of request for substantive examination | ||
| GR01 | Patent grant | ||
| GR01 | Patent grant |