CN110166089B - Superposition coding CSI feedback method based on deep learning - Google Patents
Superposition coding CSI feedback method based on deep learning Download PDFInfo
- Publication number
- CN110166089B CN110166089B CN201910442100.2A CN201910442100A CN110166089B CN 110166089 B CN110166089 B CN 110166089B CN 201910442100 A CN201910442100 A CN 201910442100A CN 110166089 B CN110166089 B CN 110166089B
- Authority
- CN
- China
- Prior art keywords
- csi
- net1
- det
- sequence
- net2
- 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 38
- 238000013135 deep learning Methods 0.000 title claims abstract description 10
- 238000012549 training Methods 0.000 claims abstract description 22
- 238000013528 artificial neural network Methods 0.000 claims abstract description 9
- 238000012545 processing Methods 0.000 claims abstract description 8
- 230000007480 spreading Effects 0.000 claims description 14
- 239000011159 matrix material Substances 0.000 claims description 12
- 230000005540 biological transmission Effects 0.000 claims description 10
- 239000000126 substance Substances 0.000 claims description 8
- 230000006870 function Effects 0.000 claims description 6
- 230000004913 activation Effects 0.000 claims description 3
- 238000010606 normalization Methods 0.000 claims description 3
- 230000008569 process Effects 0.000 claims description 3
- 230000017105 transposition Effects 0.000 claims description 3
- 238000001228 spectrum Methods 0.000 abstract description 7
- 238000001514 detection method Methods 0.000 abstract description 3
- 238000011084 recovery Methods 0.000 abstract description 3
- 238000005516 engineering process Methods 0.000 description 4
- 238000010586 diagram Methods 0.000 description 2
- 230000009286 beneficial effect Effects 0.000 description 1
- 238000004364 calculation method Methods 0.000 description 1
- 238000004891 communication Methods 0.000 description 1
- 230000006872 improvement Effects 0.000 description 1
- 238000011160 research Methods 0.000 description 1
Images
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
-
- 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
- H04B7/0417—Feedback systems
-
- 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
- H04B7/0613—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 simultaneous transmission
- H04B7/0615—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 simultaneous transmission of weighted versions of same signal
- H04B7/0619—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 simultaneous transmission of weighted versions of same signal using feedback from receiving side
- H04B7/0621—Feedback content
- H04B7/0626—Channel coefficients, e.g. channel state information [CSI]
Landscapes
- Engineering & Computer Science (AREA)
- Computer Networks & Wireless Communication (AREA)
- Signal Processing (AREA)
- Mobile Radio Communication Systems (AREA)
Abstract
The invention discloses a superposition coding CSI feedback method based on deep learning, which comprises the following steps: a, a user side: (a1) a user side reads channel state information and an uplink user sequence; (a2) performing spread spectrum processing on the channel state information to obtain a spread spectrum sequence; (a3) carrying out weighted superposition on the spread spectrum sequence and the uplink user sequence to obtain a superposed sequence; (a4) the client transmits the superimposed sequence. b: a base station end: (b1) a base station receives and obtains a receiving sequence; (b2) and the base station recovers the channel state information and the uplink user sequence by using the multi-task neural network obtained by training. Compared with non-superposition coding CSI feedback, the method completely avoids uplink bandwidth resource occupation; compared with the superposition coding CSI feedback, the method can improve the CSI recovery precision and improve the detection performance of the uplink user sequence; meanwhile, the invention can simplify the system architecture and reduce the processing complexity of the system.
Description
Technical Field
The invention relates to the technical field of superposition feedback of a large-scale MIMO (multiple input multiple output) system, in particular to a deep learning-based superposition coding CSI (channel State information) feedback method.
Background
As a key technology for meeting the efficient spectrum efficiency and energy efficiency of future 5g (radio generation) networks, a large-scale MIMO system can provide wireless data services for more users without increasing the transmission power and system bandwidth through hundreds of antennas deployed at a base station. Meanwhile, many operations (such as multi-user scheduling, rate allocation, transmitting end precoding, and the like) which bring performance improvement in a large-scale MIMO system depend on the acquisition of accurate downlink Channel State Information (CSI). In Frequency Division Duplex (FDD) massive MIMO systems, the reciprocity between channels is no longer applicable, and CSI can only be fed back from the user end to the base station.
The traditional CSI scheme based on the codebook is difficult to apply due to the fact that the number of antennas is large and the dimension of the required codebook is huge; while the Compressed Sensing (CS) feedback technology using the signal sparsity can reduce the feedback overhead of the system to some extent, but occupies a certain spectrum resource in the feedback process.
In recent years, a Superposition (SC) technique has been widely used in various fields of wireless communication due to its characteristic of being able to efficiently use spectrum resources. Meanwhile, deep learning attracts wide attention due to the advantages of high precision, high calculation speed and the like. The invention combines the advantages of SC and CS technology, superimposes the CSI weighting on the information sequence and feeds back the information sequence to the base station, and recovers the CSI at the receiving end by utilizing the deep learning technology. And then, under the condition of not increasing the frequency spectrum overhead, the reconstruction accuracy and the reconstruction rate of the CSI are improved, more implementable schemes are brought for channel feedback research, and the method has great significance.
Disclosure of Invention
The invention aims to provide a superposition coding CSI feedback method based on deep learning, which completely avoids uplink bandwidth resource occupation compared with non-superposition coding CSI feedback; compared with the superposition coding CSI feedback, the method can improve the CSI recovery precision and improve the detection performance of the uplink user sequence; meanwhile, the invention can simplify the system architecture and reduce the processing complexity of the system.
The specific technical scheme is as follows:
the superposition coding CSI feedback method based on deep learning comprises the following steps: :
a, a user side:
(a1) a user side reads channel state information H with the length of N and an uplink user sequence D with the length of M;
(a2) spreading the channel state information H to obtain a spreading sequence H with the length of Mspread;
The matrix P is a walsh spreading matrix and satisfies PTP=MINThe superscript "T" denotes the transposition operation, INRepresenting an N-order identity matrix;
(a3) spreading sequence HspreadCarrying out weighted superposition with the uplink user sequence D to obtain a superposed sequence;
(a4) the user side transmits the superposition sequence;
b: a base station end:
(b1) a base station receives and obtains a receiving sequence R;
(b2) the base station recovers the channel state information H and the uplink user sequence D by utilizing a multi-task neural network obtained by training based on a sub-network-by-sub-network training method;
wherein the weighted overlap-add of the method step (a3) can be expressed as:
wherein rho is [0,1 ]]Denotes the superposition factor, EKRepresenting the user transmission power, N representing the channel state information frame length, and S representing the superposition sequence;
wherein, the "multitasking neural network" in the step (b2) of the method includes four sub-networks: CSI-NET1, CSI-NET2, DET-NET1 and DET-NET 2;
the sub-networks CSI-NET1, CSI-NET2, DET-NET1 and DET-NET2 are connected together in a cascade mode;
the sub-networks CSI-NET1, CSI-NET2, DET-NET1 and DET-NET2 each comprise only an input layer, a hidden layer and an output layer;
the input of the sub-networks CSI-NET1, CSI-NET2, DET-NET1 and DET-NET2 adopts standardized processing;
hidden layer activation functions of the sub-networks CSI-NET1, CSI-NET2, DET-NET1 and DET-NET2 all adopt swish functions;
the numbers of nodes of the input layer, the hidden layer and the output layer of the CSI-NET1 and the CSI-NET2 sub-networks are respectively 2N, 16N and 2N;
the number of nodes of the input layer, the hidden layer and the output layer of the DET-NET1 and the DET-NET2 sub-networks is respectively 2M, 16M and 2M;
between the sub-networks CSI-NET1 and DET-NET1, the superimposed interference of the channel state information is reduced by the following method:
wherein the content of the first and second substances,represents the output of CSI-NET1, ρ ∈ [0,1 ]]Denotes the superposition factor, EKIndicating the user transmission power, N the channel state information frame length, R the receive sequence,represents the output after eliminating the superimposed interference of the channel state information and willAs an input to DET-NET 1;
between the sub-networks DET-NET1 and CSI-NET2, the superimposed interference of the uplink user sequence is reduced by the following method, namely
Wherein the content of the first and second substances,represents the output of DET-NET1, ρ ∈ [0,1 ]]Denotes the superposition factor, EKIndicating the user transmit power, R the receive sequence,represents the output after eliminating the superimposed interference of the uplink user sequence and willAs an input to the CSI-NET 2;
between the sub-networks CSI-NET2 and DET-NET2, the superimposed interference of the channel state information is reduced by the following method
Wherein the content of the first and second substances,represents the output of CSI-NET2, ρ ∈ [0,1 ]]Denotes the superposition factor, EKIndicating the user transmission power, N the channel state information frame length, R the receive sequence,represents the output after eliminating the superimposed interference of the channel state information and willAs an input to DET-NET 2;
the normalization processing formula is norm (x) ═ x- μ)/σ;
wherein x represents a vector to be normalized, μ represents a mean value of x, and σ represents a standard deviation of x;
wherein, the "subnet-by-subnet training method" described in step (b2) of the method:
b21 network parameters (W) for training CSI-NET1h11,bh11,Wh12,bh12);
b22, keeping CSI-NET1 network parameters (W)h11,bh11,Wh12,bh12) Training the DET-NET1 network parameters (W) unchangedd11,bd11,Wd12,bd12);
b23, maintaining CSI-NET1 and DET-NET1 network parameters (W)h11,bh11,Wh12,bh12,Wd11,bd11,Wd12,bd12) Training CSI-NET2 network parameters (W) unchangedh21,bh21,Wh22,bh22);
b24, maintaining CSI-NET1, DET-NET1 and CSI-NET2 network parameters (W)h11,bh11,Wh12,bh12,Wd11,bd11,Wd12,bd12,Wh21,bh21,Wh22,bh22) Training the DET-NET2 network parameters (W) unchangedd21,bd21,Wd22,bd22);
b25, saving network parameters of each layer (W)h11,bh11,Wh12,bh12,Wd11,bd11,Wd12,bd12,Wh21,bh21,Wh22,bh22,Wd21,bd21,Wd22,bd22);
W ishij,Wdij(i-1, 2; j-1, 2) represents a weighting matrix, bhij,bdij(i ═ 1, 2; j ═ 1,2) denotes a bias vector.
The invention has the beneficial effects that:
compared with non-superposition coding CSI feedback, the method completely avoids uplink bandwidth resource occupation; compared with the superposition coding CSI feedback, the method can improve the CSI recovery precision and improve the detection performance of the uplink user sequence; meanwhile, the invention can simplify the system architecture and reduce the processing complexity of the system.
Drawings
FIG. 1 is a schematic flow diagram of the present invention;
FIG. 2 is a schematic structural diagram of the "multitask neural network" according to the present invention.
Detailed Description
The technical solutions of the present invention are further described in detail below with reference to the accompanying drawings, but the scope of the present invention is not limited to the following.
As shown in fig. 1, the method for feedback of superposition coding CSI based on deep learning includes:
a, a user side:
(a1) a user side reads channel state information H with the length of N and an uplink user sequence D with the length of M;
(a2) spreading the channel state information H to obtain a spreading sequence H with the length of Mspread;
The matrix P is a walsh spreading matrix and satisfies PTP=MINThe superscript "T" denotes the transposition operation, INRepresenting an N-order identity matrix;
in this embodiment, the step a2) is exemplified as follows:
suppose that: n-4, M-8, H-0.2 +0.3j,0.4+0.5j,0.6+0.7j,0.8+0.9j,
(a3) spreading sequence HspreadCarrying out weighted superposition with the uplink user sequence D to obtain a superposed sequence;
the weighted overlap-add in step (a3) of the method may be expressed as:
wherein rho is [0,1 ]]Denotes the superposition factor, EKRepresenting the user transmission power, N representing the channel state information frame length, and S representing the superposition sequence;
in this embodiment, the step a3) is exemplified as follows:
suppose that: ρ 0.2, EK100, N-4, the uplink user sequence D ═ 1-1j, -1+1j,1+1j, -1-1j,
spreading sequence Hspread=(0.2+0.3j,0.4+0.5j,0.6+0.7j,0.8+0.9j),
S=(9.3915-8.2735j,-8.0498+10.0623j,10.2859+10.5059j,-7.1554-6.9318j)。
(a4) the user side transmits the superposition sequence;
b: a base station end:
(b1) a base station receives and obtains a receiving sequence R;
(b2) the base station recovers the channel state information H and the uplink user sequence D by utilizing a multi-task neural network obtained by training based on a sub-network-by-sub-network training method;
as shown in fig. 2, in the embodiment of the present application, the "multitasking neural network" described in step b 2):
the multitask neural network comprises four sub-networks: CSI-NET1, CSI-NET2, DET-NET1 and DET-NET 2;
the sub-networks CSI-NET1, CSI-NET2, DET-NET1 and DET-NET2 are connected together in a cascade mode;
the sub-networks CSI-NET1, CSI-NET2, DET-NET1 and DET-NET2 each comprise only an input layer, a hidden layer and an output layer;
the input of the sub-networks CSI-NET1, CSI-NET2, DET-NET1 and DET-NET2 adopts standardized processing;
hidden layer activation functions of the sub-networks CSI-NET1, CSI-NET2, DET-NET1 and DET-NET2 all adopt swish functions;
the numbers of nodes of the input layer, the hidden layer and the output layer of the CSI-NET1 and the CSI-NET2 sub-networks are respectively 2N, 16N and 2N;
the number of nodes of the input layer, the hidden layer and the output layer of the DET-NET1 and the DET-NET2 sub-networks is respectively 2M, 16M and 2M;
between the sub-networks CSI-NET1 and DET-NET1, the superimposed interference of the channel state information is reduced by the following method
Wherein the content of the first and second substances,represents the output of CSI-NET1, ρ ∈ [0,1 ]]Denotes the superposition factor, EKIndicating the user transmission power, N the channel state information frame length, R the receive sequence,represents the output after eliminating the superimposed interference of the channel state information and willAs an input to DET-NET 1;
between the sub-networks DET-NET1 and CSI-NET2, the superimposed interference of the uplink user sequence is reduced by the following method, namely
Wherein the content of the first and second substances,represents the output of DET-NET1, ρ ∈ [0,1 ]]Denotes the superposition factor, EKIndicating the user transmit power, R the receive sequence,represents the output after eliminating the superimposed interference of the uplink user sequence and willAs an input to the CSI-NET 2;
between the sub-networks CSI-NET2 and DET-NET2, the superimposed interference of the channel state information is reduced by the following method
Wherein the content of the first and second substances,represents the output of CSI-NET2, ρ ∈ [0,1 ]]Denotes the superposition factor, EKIndicating the user transmission power, N the channel state information frame length, R the receive sequence,represents the output after eliminating the superimposed interference of the channel state information and willAs an input to DET-NET 2;
the normalization process formula is norm (x) ═ (x- μ)/σ,
wherein x represents a vector to be normalized, μ represents a mean value of x, and σ represents a standard deviation of x;
the "training method by sub-network" in step b2) of the method is characterized in that:
b21 network parameters (W) for training CSI-NET1h11,bh11,Wh12,bh12);
b22, keeping CSI-NET1 network parameters (W)h11,bh11,Wh12,bh12) Training the DET-NET1 network parameters (W) unchangedd11,bd11,Wd12,bd12);
b23, maintaining CSI-NET1 and DET-NET1 network parameters (W)h11,bh11,Wh12,bh12,Wd11,bd11,Wd12,bd12) Training CSI-NET2 network parameters (W) unchangedh21,bh21,Wh22,bh22);
b24, maintaining CSI-NET1, DET-NET1 and CSI-NET2 network parameters (W)h11,bh11,Wh12,bh12,Wd11,bd11,Wd12,bd12,Wh21,bh21,Wh22,bh22) Training the DET-NET2 network parameters (W) unchangedd21,bd21,Wd22,bd22);
b25, saving network parameters of each layer
(Wh11,bh11,Wh12,bh12,Wd11,bd11,Wd12,bd12,Wh21,bh21,Wh22,bh22,Wd21,bd21,Wd22,bd22);
W ishij,Wdij(i-1, 2; j-1, 2) represents a weighting matrix, bhij,bdij(i ═ 1, 2; j ═ 1,2) denotes a bias vector.
It is to be understood that the embodiments described herein are for the purpose of assisting the reader in understanding the manner of practicing the invention and are not to be construed as limiting the scope of the invention to such particular statements and embodiments. Those skilled in the art can make various other specific changes and combinations based on the teachings of the present invention without departing from the spirit of the invention, and these changes and combinations are within the scope of the invention.
Claims (1)
1. The superposition coding CSI feedback method based on deep learning is characterized by comprising the following steps:
a, a user side:
(a1) a user side reads channel state information H with the length of N and an uplink user sequence D with the length of M;
(a2) spreading the channel state information H to obtain a spreading sequence H with the length of Mspread;
The matrix P is a walsh spreading matrix, and satisfies PTP=MINThe superscript "T" denotes the transposition operation, INRepresenting an N-order identity matrix;
(a3) spreading sequence HspreadCarrying out weighted superposition with the uplink user sequence D to obtain a superposed sequence;
the weighted overlap-add of step (a 3):
wherein rho is [0,1 ]]Denotes the superposition factor, EKRepresenting the user transmission power, N representing the channel state information frame length, and S representing the superposition sequence;
(a4) the user side transmits the superposition sequence;
b: a base station end:
(b1) a base station receives and obtains a receiving sequence R;
(b2) the base station recovers the channel state information H and the uplink user sequence D by utilizing a multi-task neural network obtained by training based on a sub-network-by-sub-network training method;
the "multitasking neural network" in the step (b2), comprising four sub-networks: CSI-NET1, CSI-NET2, DET-NET1 and DET-NET 2;
the sub-networks CSI-NET1, CSI-NET2, DET-NET1 and DET-NET2 are connected together in a cascade mode;
the sub-networks CSI-NET1, CSI-NET2, DET-NET1 and DET-NET2 each comprise only an input layer, a hidden layer and an output layer;
the input of the sub-networks CSI-NET1, CSI-NET2, DET-NET1 and DET-NET2 adopts standardized processing;
hidden layer activation functions of the sub-networks CSI-NET1, CSI-NET2, DET-NET1 and DET-NET2 all adopt swish functions;
the numbers of nodes of the input layer, the hidden layer and the output layer of the CSI-NET1 and the CSI-NET2 sub-networks are respectively 2N, 16N and 2N;
the number of nodes of the input layer, the hidden layer and the output layer of the DET-NET1 and the DET-NET2 sub-networks is respectively 2M, 16M and 2M;
between the sub-networks CSI-NET1 and DET-NET1, the superimposed interference of the channel state information is reduced by the following method:
wherein the content of the first and second substances,represents the output of CSI-NET1, ρ ∈ [0,1 ]]Denotes the superposition factor, EKIndicating the user transmission power, N the channel state information frame length, R the receive sequence,represents the output after eliminating the superimposed interference of the channel state information and willAs an input to DET-NET 1;
between the sub-networks DET-NET1 and CSI-NET2, the superimposed interference of the uplink user sequence is reduced by the following method:
wherein the content of the first and second substances,represents the output of DET-NET1, ρ ∈ [0,1 ]]Denotes the superposition factor, EKIndicating the user transmit power, R the receive sequence,represents the output after eliminating the superimposed interference of the uplink user sequence and willAs an input to the CSI-NET 2;
between the sub-networks CSI-NET2 and DET-NET2, the superimposed interference of the channel state information is reduced by the following method:
whereinRepresents the output of CSI-NET2, ρ ∈ [0,1 ]]Denotes the superposition factor, EKIndicating the user transmission power, N the channel state information frame length, R the receive sequence,represents the output after eliminating the superimposed interference of the channel state information and willAs an input to DET-NET 2;
the normalization process formula is norm (x) ═ x- μ)/σ;
wherein x represents a vector to be normalized, μ represents a mean value of x, and σ represents a standard deviation of x;
the sub-network-by-sub-network training method comprises the following steps:
b21 network parameter W of training CSI-NET1h11,bh11,Wh12,bh12;
b22, keeping CSI-NET1 network parameter Wh11,bh11,Wh12,bh12Training the DET-NET1 network parameter W without changed11,bd11,Wd12,bd12;
b23, keeping CSI-NET1 and DET-NET1 network parameters Wh11,bh11,Wh12,bh12,Wd11,bd11,Wd12,bd12Training the CSI-NET2 network parameter W without changeh21,bh21,Wh22,bh22;
b24, keeping CSI-NET1, DET-NET1 and CSI-NET2 network parameters Wh11,bh11,Wh12,bh12,Wd11,bd11,Wd12,bd12,Wh21,bh21,Wh22,bh22Training the DET-NET2 network parameter W without changed21,bd21,Wd22,bd22;
b25, saving network parameters of each layer
Wh11,bh11,Wh12,bh12,Wd11,bd11,Wd12,bd12,Wh21,bh21,Wh22,bh22,Wd21,bd21,Wd22,bd22;
W ishij,Wdij(i-1, 2; j-1, 2) represents a weighting matrix, bhij,bdij(i ═ 1, 2; j ═ 1,2) denotes a bias vector.
Priority Applications (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910442100.2A CN110166089B (en) | 2019-05-24 | 2019-05-24 | Superposition coding CSI feedback method based on deep learning |
Applications Claiming Priority (1)
Application Number | Priority Date | Filing Date | Title |
---|---|---|---|
CN201910442100.2A CN110166089B (en) | 2019-05-24 | 2019-05-24 | Superposition coding CSI feedback method based on deep learning |
Publications (2)
Publication Number | Publication Date |
---|---|
CN110166089A CN110166089A (en) | 2019-08-23 |
CN110166089B true CN110166089B (en) | 2021-06-04 |
Family
ID=67632589
Family Applications (1)
Application Number | Title | Priority Date | Filing Date |
---|---|---|---|
CN201910442100.2A Active CN110166089B (en) | 2019-05-24 | 2019-05-24 | Superposition coding CSI feedback method based on deep learning |
Country Status (1)
Country | Link |
---|---|
CN (1) | CN110166089B (en) |
Families Citing this family (5)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN111193535B (en) * | 2020-01-14 | 2022-05-31 | 西华大学 | Feedback method based on ELM superposition CSI in FDD large-scale MIMO system |
CN111597877A (en) * | 2020-04-02 | 2020-08-28 | 浙江工业大学 | Fall detection method based on wireless signals |
CN113765830B (en) * | 2020-06-03 | 2022-12-27 | 华为技术有限公司 | Method for acquiring channel information and communication device |
CN112564757A (en) * | 2020-12-03 | 2021-03-26 | 西华大学 | Deep learning 1-bit compression superposition channel state information feedback method |
CN113472412B (en) * | 2021-07-13 | 2023-10-24 | 西华大学 | Enhanced ELM-based superposition CSI feedback method |
Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107070588A (en) * | 2016-12-28 | 2017-08-18 | 深圳清华大学研究院 | A kind of multiple access of simplification accesses the receiver and method of reseptance of Transmission system |
CN108390706A (en) * | 2018-01-30 | 2018-08-10 | 东南大学 | A kind of extensive mimo channel state information feedback method based on deep learning |
CN108737032A (en) * | 2018-05-22 | 2018-11-02 | 西华大学 | A kind of compression superposition sequence C SI feedback methods |
CN109687897A (en) * | 2019-02-25 | 2019-04-26 | 西华大学 | Superposition CSI feedback method based on the extensive mimo system of deep learning |
Family Cites Families (1)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
WO2018204917A1 (en) * | 2017-05-05 | 2018-11-08 | Ball Aerospace & Technologies Corp. | Spectral sensing and allocation using deep machine learning |
-
2019
- 2019-05-24 CN CN201910442100.2A patent/CN110166089B/en active Active
Patent Citations (4)
Publication number | Priority date | Publication date | Assignee | Title |
---|---|---|---|---|
CN107070588A (en) * | 2016-12-28 | 2017-08-18 | 深圳清华大学研究院 | A kind of multiple access of simplification accesses the receiver and method of reseptance of Transmission system |
CN108390706A (en) * | 2018-01-30 | 2018-08-10 | 东南大学 | A kind of extensive mimo channel state information feedback method based on deep learning |
CN108737032A (en) * | 2018-05-22 | 2018-11-02 | 西华大学 | A kind of compression superposition sequence C SI feedback methods |
CN109687897A (en) * | 2019-02-25 | 2019-04-26 | 西华大学 | Superposition CSI feedback method based on the extensive mimo system of deep learning |
Also Published As
Publication number | Publication date |
---|---|
CN110166089A (en) | 2019-08-23 |
Similar Documents
Publication | Publication Date | Title |
---|---|---|
CN110166089B (en) | Superposition coding CSI feedback method based on deep learning | |
CN109687897B (en) | Superposition CSI feedback method based on deep learning large-scale MIMO system | |
Liao et al. | CSI feedback based on deep learning for massive MIMO systems | |
Jang et al. | Deep autoencoder based CSI feedback with feedback errors and feedback delay in FDD massive MIMO systems | |
Perlaza et al. | From spectrum pooling to space pooling: Opportunistic interference alignment in MIMO cognitive networks | |
CN103117970B (en) | The system of selection of full-duplex antenna in mimo system | |
CN103209051B (en) | The two step method for precoding of a kind of coordinate multipoint joint transmission system under multi-user scene | |
CN105337651A (en) | User selection method of non-orthogonal multiple access system downlink under limited feedback | |
CN109714091B (en) | Iterative hybrid precoding method based on hierarchical design in millimeter wave MIMO system | |
CN107086886B (en) | Double-layer precoding design for large-scale MIMO system fusion zero forcing and Taylor series expansion | |
CN103220024A (en) | Beam forming algorithm of multi-user pairing virtual multi-input multi-output (MIMO) system | |
Elbir et al. | Federated learning for physical layer design | |
CN110289898A (en) | A kind of channel feedback method based on the perception of 1 bit compression in extensive mimo system | |
CN112564757A (en) | Deep learning 1-bit compression superposition channel state information feedback method | |
CN109167618B (en) | FDD large-scale MIMO downlink channel reconstruction and multi-user transmission method | |
Boukhedimi et al. | LMMSE receivers in uplink massive MIMO systems with correlated Rician fading | |
CN114219354A (en) | Resource allocation optimization method and system based on federal learning | |
CN110808764A (en) | Joint information estimation method in large-scale MIMO relay system | |
Xiao et al. | AI enlightens wireless communication: Analyses, solutions and opportunities on CSI feedback | |
CN102404031A (en) | Self-adaptive user scheduling method based on maximum throughput | |
CN113726376B (en) | 1bit compression superposition CSI feedback method based on feature extraction and mutual-difference fusion | |
CN111193535B (en) | Feedback method based on ELM superposition CSI in FDD large-scale MIMO system | |
CN104158575A (en) | Method of user scheduling of multi-cell MIMO (Multiple Input Multiple Output) system under ZF (Zero Frequency) pre-coding strategy | |
CN111865844B (en) | Channel estimation method and device for large-scale MIMO full-duplex relay system | |
CN109039402B (en) | MIMO topological interference alignment method based on user compression |
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 | ||
EE01 | Entry into force of recordation of patent licensing contract | ||
EE01 | Entry into force of recordation of patent licensing contract |
Application publication date: 20190823 Assignee: Chengdu Tiantongrui Computer Technology Co.,Ltd. Assignor: XIHUA University Contract record no.: X2023510000028 Denomination of invention: Stacked Encoding CSI Feedback Method Based on Deep Learning Granted publication date: 20210604 License type: Common License Record date: 20231124 |