WO2022121797A1 - 一种传输数据的方法和装置 - Google Patents

一种传输数据的方法和装置 Download PDF

Info

Publication number
WO2022121797A1
WO2022121797A1 PCT/CN2021/135367 CN2021135367W WO2022121797A1 WO 2022121797 A1 WO2022121797 A1 WO 2022121797A1 CN 2021135367 W CN2021135367 W CN 2021135367W WO 2022121797 A1 WO2022121797 A1 WO 2022121797A1
Authority
WO
WIPO (PCT)
Prior art keywords
information
decoding
csi
neural network
encoding
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.)
Ceased
Application number
PCT/CN2021/135367
Other languages
English (en)
French (fr)
Inventor
王四海
毕双凯声
李雪茹
秦城
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Huawei Technologies Co Ltd
Original Assignee
Huawei Technologies Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Huawei Technologies Co Ltd filed Critical Huawei Technologies Co Ltd
Priority to EP21902504.6A priority Critical patent/EP4243492A4/en
Priority to US18/265,495 priority patent/US20240030980A1/en
Publication of WO2022121797A1 publication Critical patent/WO2022121797A1/zh
Anticipated expiration legal-status Critical
Ceased legal-status Critical Current

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/003—Arrangements for allocating sub-channels of the transmission path
    • H04L5/0053—Allocation of signalling, i.e. of overhead other than pilot signals
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/003—Arrangements for allocating sub-channels of the transmission path
    • H04L5/0053—Allocation of signalling, i.e. of overhead other than pilot signals
    • H04L5/0057—Physical resource allocation for CQI
    • G—PHYSICS
    • G06—COMPUTING OR CALCULATING; COUNTING
    • G06N—COMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00—Computing arrangements based on biological models
    • G06N3/02—Neural networks
    • G06N3/04—Architecture, e.g. interconnection topology
    • G06N3/045—Combinations of networks
    • G06N3/0455—Auto-encoder networks; Encoder-decoder networks
    • 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
    • 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]
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L25/00—Baseband systems
    • H04L25/02—Details ; arrangements for supplying electrical power along data transmission lines
    • H04L25/0202—Channel estimation
    • H04L25/0204—Channel estimation of multiple channels
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/003—Arrangements for allocating sub-channels of the transmission path
    • H04L5/0048—Allocation of pilot signals, i.e. of signals known to the receiver
    • H04L5/005—Allocation of pilot signals, i.e. of signals known to the receiver of common pilots, i.e. pilots destined for multiple users or terminals
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04L—TRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L5/00—Arrangements affording multiple use of the transmission path
    • H04L5/0091—Signalling for the administration of the divided path, e.g. signalling of configuration information
    • H04L5/0094—Indication of how sub-channels of the path are allocated

Definitions

  • the present application relates to the field of wireless communication, and more particularly, to a method and apparatus for transmitting data.
  • the terminal device needs to feed back the estimated channel state information (CSI) to the access network device, so that the access network device can use the CSI to perform multi-antenna precoding.
  • CSI channel state information
  • CSI-reference signal CSI-reference signal
  • CSI-RS channel state information reference signal
  • the present application provides a method and apparatus for transmitting data, which can reduce the occupation of resources.
  • a first aspect provides a method for transmitting data, the method can be performed by an access network device or a chip in the access network device, the method includes: the access network device performs a first step on a channel state information reference signal CSI-RS a code to generate first information, the CSI-RS is used to obtain the channel state information CSI corresponding to the channel between the access network device and the terminal device; the access network device sends the information to the terminal device through N radio frequency links the first information, where N is a positive integer; the access network device receives third information from the terminal device, the third information is generated by the terminal device performing second encoding on the second information, and the second information is generated by the terminal device based on The first information and the CSI-RS are determined; the access network device performs first decoding on the third information to generate the CSI, where the first decoding includes decoding corresponding to the first encoding.
  • CSI-RS channel state information reference signal
  • the terminal device observes the first coded CSI-RS to determine the implicit CSI , where M is a positive integer.
  • the access network device sends the CSI-RS for obtaining CSI to the terminal device.
  • the CSI-RS can be encoded and sent to the terminal device, that is, the first information is sent to the terminal device.
  • the second information may be generated according to the first information and the CSI-RS, where the CSI-RS may be pre-configured in the terminal device, that is, the terminal device receives the transmission through the channel
  • the first information of the CSI is observed, so as to obtain the second information that implicitly includes the CSI.
  • the terminal device performs second encoding on the second information to generate third information.
  • the access network device obtains the CSI after receiving the third information and decoding it.
  • the access network device sends the encoded CSI-RS to the terminal device through N radio frequency links, and the terminal device does not need to perform a decoding operation on the encoded CSI-RS, but then performs a decoding operation.
  • the one-time encoding is sent to the access network device to be decoded by the access network device to obtain the CSI, which can reduce resource occupation during the process of obtaining the CSI.
  • the method further includes: performing, by the access network device, a second decoding on the third information, where the second decoding corresponds to the second encoding; the accessing
  • the network device performing the first decoding on the third information includes: the access network device performing the first decoding on the third information subjected to the second decoding.
  • the access network device When the access network device receives the twice-encoded information, it can first perform the second decoding on the second encoding, and then perform the first decoding on the first encoding on the third information after the second decoding, that is, two-level decoding is adopted.
  • the structure of decodes the third information to obtain CSI.
  • the first decoding further includes decoding corresponding to the second encoding.
  • the third information received by the access network is the information that has been encoded twice, but the decoding structures corresponding to the two encodings can be merged into one structure, that is, the third information can be decoded first, and the The first decoding includes both decoding corresponding to the first encoding and decoding corresponding to the second encoding, and the required CSI is obtained through one decoding.
  • the access network device can use a fusion structure to decode the twice-encoded information, which improves the decoding accuracy.
  • the second code is a code based on a first neural network, and the parameters of the first neural network are related to the number of sampling points F of the channel and the occupation of the first information.
  • the number of resources M is related to M, where M and F are positive integers, M ⁇ N, and the resources include at least one of the following: time domain resources, frequency domain resources or code domain resources.
  • the parameter of the first neural network is the dimension of the input matrix of the first neural network.
  • the terminal device can perform encoding again based on the neural network and then feed back the CSI to the network device, which further saves the resource overhead of CSI feedback.
  • the first encoding is a compressed sensing-based encoding
  • the first encoding uses a first matrix
  • the dimensions of the first matrix are related to the M and the N .
  • the dimension of the first matrix is M ⁇ N.
  • the first matrix is in the form of multiplying multiple matrices, and the dimension after multiplying the multiple matrices is M ⁇ N.
  • the access network device can use compressed sensing to reserve only M groups of resources for sending CSI-RS, which reduces the occupation of sending CSI-RS resources.
  • the first code is a code based on a second neural network
  • the second neural network includes a fully connected linear layer
  • the parameters of the fully connected linear layer are the same as the M related to this N.
  • the fully connected linear layer may be in the form of at least one matrix multiplication, and the dimension of the formed matrix is M ⁇ N.
  • the access network device may only reserve M groups of resources for sending CSI-RS by including the second neural network of the fully connected linear layer, which reduces the need for sending CSI-RS. Occupation of resources.
  • the first decoding is decoding based on a third neural network, and the parameters of the third neural network are related to the N, the M and the F.
  • the third neural network is only used for decoding corresponding to the first code, and the parameters of the third neural network are related to N, M and F.
  • the third neural network is used not only for decoding the corresponding first code, but also for decoding the second code, and the parameters of the third neural network are related to N, M and F.
  • the training method of the first neural network is obtained by concatenating the first matrix, the first neural network and the third neural network through unified training.
  • the training method of the first neural network is obtained by concatenating the second neural network, the first neural network and the third neural network through unified training.
  • the third neural network at this time is a fusion decoding structure corresponding to the first decoding and the second decoding.
  • the second decoding is decoding based on a fourth neural network method
  • the training method of the first neural network is to connect the first neural network and the fourth neural network in series.
  • the training method of the network is either the training method of concatenating the first matrix, the first neural network, the fourth neural network and the third neural network, or the series of the second neural network, the first neural network, the fourth neural network and the third neural network. How the neural network is trained.
  • the access network device sends the encoded CSI-RS to the terminal device through N radio frequency links, and the terminal device does not need to perform a decoding operation on the encoded CSI-RS, but then performs a decoding operation. After encoding once, it is sent to the access network device for decoding by the access network device to obtain the CSI, which can reduce the occupation of resources in the process of obtaining the CSI.
  • a method for transmitting data can be executed by a terminal device or a chip in the terminal device, and the method includes: the terminal device receives data sent by an access network device to the terminal device through N radio frequency links first information, the first information is generated by the access network device performing first coding on the channel state information reference signal CSI-RS, and the CSI-RS is used to obtain the channel information CSI; the terminal device is based on the first information and the The CSI-RS determines the second information; the terminal device performs second coding on the second information to generate the third information; the terminal device sends the third information to the access network device, and the third information is used The access network device performs first decoding on the third information to generate the CSI, where the first decoding includes decoding corresponding to the first encoding.
  • the access network device sends the CSI-RS for obtaining CSI to the terminal device.
  • the CSI-RS can be encoded and sent to the terminal device, that is, the first information is sent to the terminal device.
  • the second information may be generated according to the first information and the CSI-RS, where the CSI-RS may be pre-configured in the terminal device, that is, the terminal device receives the transmission through the channel
  • the first information of the CSI is observed, so as to obtain the second information that implicitly includes the CSI.
  • the terminal device performs second encoding on the second information to generate third information.
  • the access network device obtains the CSI after receiving the third information and decoding it.
  • the access network device sends the encoded CSI-RS to the terminal device through N radio frequency links, and the terminal device does not need to perform a decoding operation on the encoded CSI-RS, but then performs a decoding operation.
  • the one-time encoding is sent to the access network device to be decoded by the access network device to obtain the CSI, which can reduce resource occupation during the process of obtaining the CSI.
  • the third information is used for the access network device to perform the first decoding on the second decoded third information to generate the CSI, and the third information is used for generating the CSI.
  • the second decoding corresponds to the second encoding.
  • the first decoding further includes decoding corresponding to the second encoding.
  • the second code is a code based on a first neural network, and the parameters of the first neural network are related to the number of sampling points F of the channel and the occupation of the first information.
  • the number of resources M is related to M, where M and F are positive integers, M ⁇ N, and the resources include at least one of the following: time domain resources, frequency domain resources or code domain resources.
  • the parameter of the first neural network is the dimension of the input matrix of the first neural network.
  • the first encoding is a compressed sensing-based encoding
  • the first encoding uses a first matrix
  • the dimensions of the first matrix are related to the M and the N .
  • the first encoding is an encoding based on a second neural network
  • the second neural network includes a fully connected linear layer
  • the parameters of the fully connected linear layer are the same as the M related to this N.
  • the first decoding is decoding based on a third neural network, and the parameters of the third neural network are related to the N, the M and the F.
  • an apparatus for transmitting data includes: a processing unit configured to perform first encoding on a channel state information reference signal CSI-RS to generate first information, where the CSI-RS is used to obtain a connection Channel state information CSI corresponding to the channel between the network access device and the terminal device; the transceiver unit is used to send the first information to the terminal device through N radio frequency links, where N is a positive integer; the third information is generated by the The terminal device performs second encoding on the second information to generate, and the second information is determined by the terminal device based on the first information and the CSI-RS; the processing unit is further configured to perform a first decoding on the third information, to The CSI is generated, and the first decoding includes decoding corresponding to the first encoding.
  • CSI-RS channel state information reference signal
  • the access network device sends the encoded CSI-RS to the terminal device through N radio frequency links, and the terminal device does not need to perform a decoding operation on the encoded CSI-RS, but then performs a decoding operation.
  • the one-time encoding is sent to the access network device to be decoded by the access network device to obtain the CSI, which can reduce resource occupation during the process of obtaining the CSI.
  • the processing unit is further configured to perform a second decoding on the third information, where the second decoding corresponds to the second encoding; the processing unit is further configured to perform a second decoding on the third information.
  • the first decoding is performed on the third information, which is specifically used for performing the first decoding on the third information subjected to the second decoding.
  • the first decoding further includes decoding corresponding to the second encoding.
  • the second coding is a coding based on a first neural network, and the parameters of the first neural network are related to the number of sampling points F of the channel and the occupation of the first information.
  • the number of resources M is related to M, where M and F are positive integers, M ⁇ N, and the resources include at least one of the following: time domain resources, frequency domain resources or code domain resources.
  • the parameter of the first neural network is the dimension of the input matrix of the first neural network.
  • the first encoding is a compressed sensing-based encoding
  • the first encoding uses a first matrix
  • the dimensions of the first matrix are related to the M and the N .
  • the first code is a code based on a second neural network
  • the second neural network includes a fully connected linear layer
  • the parameters of the fully connected linear layer are the same as the M related to this N.
  • the first decoding is decoding based on a third neural network, and the parameters of the third neural network are related to the N, the M and the F.
  • an apparatus for transmitting data includes: a transceiver unit configured to receive first information sent by an access network device through N radio frequency links, where the first information is sent by the access network device to a channel
  • the state information reference signal CSI-RS is used to generate the first code, and the CSI-RS is used to obtain the channel information CSI
  • the processing unit is used to determine the second information based on the first information and the CSI-RS
  • the processing unit is further used for Perform second encoding on the second information
  • the transceiver unit is further configured to send third information to the access network device, where the third information is used for the access network device to first decode the third information to generate
  • the CSI the first decoding includes decoding corresponding to the first encoding.
  • the access network device sends the encoded CSI-RS to the terminal device through N radio frequency links, and the terminal device does not need to perform a decoding operation on the encoded CSI-RS, but then performs a decoding operation.
  • the one-time encoding is sent to the access network device to be decoded by the access network device to obtain the CSI, which can reduce resource occupation during the process of obtaining the CSI.
  • the third information is used for the access network device to perform the first decoding on the third information that has undergone the second decoding, so as to generate the CSI, and the third information is used for generating the CSI.
  • the second decoding corresponds to the second encoding.
  • the first decoding further includes decoding corresponding to the second encoding.
  • the second coding is a coding based on a first neural network, and the parameters of the first neural network are related to the number of sampling points F of the channel and the occupation of the first information.
  • the number of resources M is related to M, where M and F are positive integers, M ⁇ N, and the resources include at least one of the following: time domain resources, frequency domain resources or code domain resources.
  • the parameter of the first neural network is the dimension of the input matrix of the first neural network.
  • the first encoding is a compressed sensing-based encoding
  • the first encoding uses a first matrix
  • the dimensions of the first matrix are related to the M and the N .
  • the first code is a code based on a second neural network
  • the second neural network includes a fully connected linear layer
  • the parameters of the fully connected linear layer are the same as the M related to this N.
  • the first decoding is decoding based on a third neural network, and the parameters of the third neural network are related to the N, the M and the F.
  • a method for transmitting data comprising: performing dimensionality reduction processing on first information to generate second information; encoding the second information based on a first neural network, the second information is The dimensions correspond to the dimensions of the information that the first neural network can process.
  • the information is first subjected to dimensionality reduction processing to obtain second information with a lower dimension, and then the second information is encoded based on the first neural network, which ensures the quality of the encoding and greatly reduces the computational complexity of the encoding.
  • the first neural network is obtained by training based on first training data, and the first training data is data that has undergone dimensionality reduction processing.
  • the first training data is the second information.
  • the first neural network can be trained on data that has undergone dimensionality reduction processing, so that data corresponding to the dimension of the second information can be processed, thereby improving coding performance.
  • the training method of the first neural network is to concatenate the dimensionality reduction processing and the unified training of the first neural network.
  • the dimensionality reduction process and the first neural network can be cascaded for unified training, and this training method can reduce the dependence of the neural network on the training data, and also have better performance on data other than the training set. performance, so as to obtain a certain generalization ability.
  • performing dimensionality reduction processing on the first information includes: performing dimensionality reduction processing on the first information based on a compressed sensing method that uses a first matrix , the size of the first matrix corresponds to the dimension of the information that the first neural network can process.
  • the compressed sensing method can finally be expressed as a layer of matrix, and the compressed sensing method can achieve the effect of reducing the dimension of the channel information. Therefore, the first neural network encodes the channel information that has been reduced in dimension, which can reduce the overall encoding calculation amount.
  • the composite coding method combining compression coding and neural network coding can significantly improve the quality of channel information compared with the method relying solely on compression coding.
  • dimensionality reduction processing is performed on the first information
  • dimensionality reduction processing is performed on the first information to generate second information, including: based on the first neural network
  • the first to N layers perform dimensionality reduction processing on the first information, the first to N layers of the first neural network are the first fully connected linear layers, the N is an integer, and N ⁇ 1; based on the first neural network, the The encoding of the second information includes: encoding the second information based on layers N+1 to M of the first neural network, the size of the first fully connected linear layer being the same as the size of the N+th layer of the first neural network
  • the dimensions of the information that can be processed by layers 1 to M correspond, where M is an integer, and M ⁇ N+1.
  • performing dimensionality reduction processing on the first information to generate the second information, and encoding the second information based on the first neural network can be expressed as a coding structure of a neural network.
  • the first to N layers of the neural network are fully-connected linear layers, that is, the channel information is dimensionally reduced through the fully-connected linear layer, and then the latter layers are used to encode the dimension-reduced channel information.
  • This application is different from the current coding method based on neural network.
  • the current neural network usually fills the data first, and then reduces the dimensionality in the later layers. In this application, the dimensionality reduction is performed first, which greatly reduces the calculation of coding. quantity.
  • performing dimensionality reduction processing on the first information, and performing dimensionality reduction processing on the first information to generate the second information includes: based on the second neural network Dimensionality reduction processing is performed on the first information to generate second information, the second neural network includes a second fully connected linear layer, and the second fully connected linear layer corresponds to the dimension of the information that can be processed by the first neural network.
  • the fully connected linear layer can mathematically be expressed as non-sparse matrix multiplication, therefore, the channel information can be dimensionally reduced through the fully connected linear layer.
  • the first information is channel state information.
  • Perform dimensionality reduction processing on the first information to generate second information that is, perform dimensionality reduction processing on channel state information to generate dimensionality-reduced channel state information, and perform dimensionality reduction on the dimensionality-reduced channel state information based on the first neural network. encoding to generate the encoded channel state information to be transmitted.
  • an apparatus for transmitting data comprising: a transceiver unit for receiving the first information; a processing unit for performing dimension reduction processing on the first information to generate the second information; the processing unit The unit is further configured to encode the second information based on the first neural network, and the dimension of the second information corresponds to the dimension of the information that can be processed by the first neural network.
  • the device first performs dimensionality reduction processing on the channel information to obtain second information with a lower dimension, and then encodes the second information based on the first neural network, which can greatly reduce the computational complexity of the encoding and ensure the quality of the encoding.
  • the first neural network is obtained by training based on first training data, and the first training data is data that has undergone dimensionality reduction processing.
  • the first training data is the second information.
  • the training method of the first neural network is to concatenate the dimensionality reduction process and the unified training of the first neural network.
  • the processing unit is specifically configured to: perform dimensionality reduction processing on the first information based on a compressed sensing method, the compressed sensing method uses a first matrix, and the first The size of the matrix corresponds to the dimension of the information that the first neural network can process.
  • the processing unit is specifically configured to: perform dimensionality reduction processing on the first information based on layers 1 to N of the first neural network, and the first neural network
  • the 1st to Nth layers are the first fully connected linear layers, and the N is an integer, N ⁇ 1;
  • the second information is encoded based on the N+1 to Mth layers of the first neural network, and the first fully connected
  • the size of the linear layer corresponds to the dimension of information that can be processed by the N+1 to Mth layers of the first neural network, where M is an integer, and M ⁇ N+1.
  • the processing unit is specifically configured to: perform dimension reduction processing on the first information based on a second neural network to generate second information, where the second neural network includes The second fully connected linear layer corresponds to the dimension of the information that the first neural network can process.
  • the first information is channel state information.
  • a method for transmitting data comprising: receiving third information; decoding the third information based on a third neural network to generate fourth information, the dimension of the third information being the same as the dimension of the third information
  • the dimensions of the information that can be processed by the three neural networks correspond to each other; the fourth information is subjected to restoration processing, and the restoration processing corresponds to the dimensionality reduction processing.
  • the training method of the third neural network is that the dimensionality reduction processing is connected in series, and the first neural network, the third neural network and the recovery processing are uniformly trained.
  • recovery processing is performed on the fourth information based on compressed sensing.
  • the third information is encoded channel state information.
  • an apparatus for transmitting data includes a transceiver unit for receiving the third information; a processing unit for decoding the third information based on a third neural network, the The dimension of the third information corresponds to the dimension of the information that the first neural network can process.
  • the training method of the third neural network is that the dimensionality reduction processing is connected in series, and the first neural network, the third neural network and the recovery processing are uniformly trained.
  • recovery processing is performed on the fourth information based on a compressed sensing manner.
  • the third information is encoded channel state information.
  • a communication apparatus may include a processing unit, a sending unit and a receiving unit.
  • the sending unit and the receiving unit may also be transceiver units.
  • the processing unit may be a processor, and the sending unit and the receiving unit may be transceivers; the apparatus may further include a storage unit, and the storage unit may be a memory; the storage unit is used for storing instruction, the processing unit executes the instruction stored in the storage unit, so that the access network device executes any method of the first aspect, the second aspect, the fifth aspect or the seventh aspect.
  • the processing unit may be a processor, the sending unit and the receiving unit may be input/output interfaces, pins or circuits, etc.; the processing unit executes the instructions stored in the storage unit , so that the chip performs the method of the first aspect, the second aspect, the fifth aspect or the seventh aspect.
  • the storage unit is used to store instructions, and the storage unit may be a storage unit in the chip (for example, a register, a cache, etc.), or a storage unit in the access network device located outside the chip (for example, a read-only unit) memory, random access memory, etc.).
  • the storage unit may be a storage unit in the chip (for example, a register, a cache, etc.), or a storage unit in the access network device located outside the chip (for example, a read-only unit) memory, random access memory, etc.).
  • the processing unit may be a processor, and the sending unit and the receiving unit may be transceivers; the apparatus may further include a storage unit, and the storage unit may be a memory; the storage unit is used for storing instructions, The processing unit executes the instructions stored in the storage unit, so that the terminal device executes any method of the first aspect, the second aspect, the fifth aspect or the seventh aspect.
  • the processing unit may be a processor, and the sending unit and the receiving unit may be input/output interfaces, pins or circuits, etc.; the processing unit executes the instructions stored in the storage unit to The chip is caused to perform the method of the first aspect, the second aspect, the fifth aspect or the seventh aspect.
  • the storage unit is used to store instructions, and the storage unit may be a storage unit in the chip (for example, a register, a cache, etc.), or a storage unit in the terminal device located outside the chip (for example, a read-only memory, random access memory, etc.).
  • a tenth aspect provides a communication device, comprising a processor and an interface circuit, the interface circuit is configured to receive signals from other communication devices other than the communication device and transmit to the processor or send signals from the processor
  • the processor is used to implement any method of the first aspect to the fifth aspect through a logic circuit or executing code instructions.
  • a computer-readable storage medium is provided, and a computer program or instruction is stored in the computer-readable storage medium.
  • the computer program or instruction is executed, the aforementioned first aspect, second aspect, and first aspect are realized. Any method of the fifth aspect or the seventh aspect.
  • a twelfth aspect provides a computer program product comprising instructions that, when executed, implement any of the methods of the aforementioned first, second, fifth or seventh aspects.
  • a thirteenth aspect provides a computer program comprising code or instructions that, when executed, implements any possible possibility of the first, second, fifth or seventh aspects described above. method in the implementation.
  • a fourteenth aspect provides a chip system, the chip system includes a processor, and may further include a memory, for implementing at least one of the methods described in the first aspect, the second aspect, the fifth aspect or the seventh aspect.
  • the chip system can be composed of chips, and can also include chips and other discrete devices.
  • a fifteenth aspect provides a communication system, which includes the apparatus (eg, an access network device) according to any one of the ninth aspect to the fourteenth aspect.
  • a sixteenth aspect provides a communication system, which includes the apparatus (eg, a terminal device) according to any one of the ninth aspect to the fourteenth aspect.
  • FIG. 1 is a schematic diagram of an example of the data transmission method applicable to the present application.
  • FIG. 2 is a schematic structural diagram of an example of transmission CSI based on compressed sensing
  • FIG. 3 is a schematic flowchart of an example of transmitting CSI according to an embodiment of the present application
  • FIG. 4 is a schematic structural diagram of an example of transmitting CSI according to an embodiment of the present application.
  • FIG. 5 is a schematic structural diagram of another example of transmitting CSI according to an embodiment of the present application.
  • FIG. 6 is a schematic flowchart of an example of acquiring CSI according to an embodiment of the present application.
  • FIG. 7 is a schematic structural diagram of an example of acquiring CSI
  • LTE Long Term Evolution
  • FDD frequency division duplex
  • TDD time division duplex
  • UMTS universal mobile telecommunication system
  • WiMAX worldwide interoperability for microwave access
  • 5G mobile communication system may include a non-standalone (NSA, NSA) and/or an independent network (standalone, SA).
  • NSA non-standalone
  • SA independent network
  • the technical solutions provided in this application can also be applied to machine type communication (MTC), Long Term Evolution-machine (LTE-M), device-to-device (D2D) networks, machine-to-machine communication (Long Term Evolution-machine, LTE-M) To a machine to machine (M2M) network, an internet of things (IoT) network, or other networks.
  • MTC machine type communication
  • LTE-M Long Term Evolution-machine
  • D2D device-to-device
  • M2M machine-to-machine communication
  • M2M machine to machine
  • IoT internet of things
  • the IoT network may include, for example, the Internet of Vehicles.
  • V2X vehicle to other devices
  • V2X vehicle to other devices
  • V2X vehicle to other devices
  • V2X vehicle to other devices
  • the V2X can include: vehicle to vehicle (vehicle to vehicle, V2V) communication, vehicle and vehicle Infrastructure (V2I) communication, vehicle-to-pedestrian (V2P) or vehicle-to-network (V2N) communication, etc.
  • V2V vehicle to vehicle
  • V2I vehicle and vehicle Infrastructure
  • V2P vehicle-to-pedestrian
  • V2N vehicle-to-network
  • the access network device may be any device with a wireless transceiver function.
  • the device includes but is not limited to: evolved Node B (evolved Node B, eNB), radio network controller (radio network controller, RNC), Node B (Node B, NB), base station controller (base station controller, BSC) , base transceiver station (base transceiver station, BTS), home base station (for example, home evolved NodeB, or home Node B, HNB), baseband unit (baseband unit, BBU), wireless fidelity (wireless fidelity, WiFi) system Access point (AP), wireless relay node, wireless backhaul node, transmission point (TP) or transmission and reception point (TRP), etc.
  • evolved Node B evolved Node B
  • RNC radio network controller
  • Node B Node B
  • BSC base station controller
  • base transceiver station base transceiver station
  • BTS home base station
  • home base station for example, home evolved NodeB, or home Node B, HNB
  • It can also be 5G, such as NR , a gNB in the system, or, a transmission point (TRP or TP), one or a group of (including multiple antenna panels) antenna panels of a base station in a 5G system, or, it can also be a network node that constitutes a gNB or a transmission point, Such as baseband unit (BBU), or distributed unit (distributed unit, DU) and so on.
  • BBU baseband unit
  • DU distributed unit
  • a gNB may include a centralized unit (CU) and a DU.
  • the gNB may also include an active antenna unit (AAU).
  • CU implements some functions of gNB
  • DU implements some functions of gNB.
  • CU is responsible for processing non-real-time protocols and services, implementing radio resource control (RRC), and packet data convergence protocol (PDCP) layer function.
  • RRC radio resource control
  • PDCP packet data convergence protocol
  • the DU is responsible for processing physical layer protocols and real-time services, and implementing the functions of the radio link control (RLC) layer, medium access control (MAC) layer, and physical (PHY) layer.
  • RLC radio link control
  • MAC medium access control
  • PHY physical layer.
  • AAU implements some physical layer processing functions, radio frequency processing and related functions of active antennas.
  • the higher-layer signaling such as the RRC layer signaling
  • the access network device may be a device including one or more of a CU node, a DU node, and an AAU node.
  • the CU can be divided into access network equipment in the access network (radio access network, RAN), and the CU can also be divided into the access network equipment in the core network (core network, CN), this application does not Do limit.
  • the access network equipment provides services for the cell, and the terminal equipment communicates with the cell through transmission resources (for example, frequency domain resources, or spectrum resources) allocated by the access network equipment, and the cell may belong to a macro base station (for example, a macro eNB or (macro gNB, etc.), can also belong to the base station corresponding to the small cell, the small cell here can include: urban cell, micro cell, pico cell, femto cell, etc. These small cells have the characteristics of small coverage and low transmission power, suitable for to provide high-speed data transmission services.
  • a macro base station for example, a macro eNB or (macro gNB, etc.
  • the small cell here can include: urban cell, micro cell, pico cell, femto cell, etc.
  • a terminal device may also be referred to as user equipment (user equipment, UE), an access terminal, a subscriber unit, a subscriber station, a mobile station, a mobile station, a remote station, a remote terminal, a mobile device, a user terminal, Terminal, wireless communication device, user agent or user equipment.
  • user equipment user equipment
  • UE user equipment
  • an access terminal a subscriber unit, a subscriber station, a mobile station, a mobile station, a remote station, a remote terminal, a mobile device, a user terminal, Terminal, wireless communication device, user agent or user equipment.
  • the terminal device may be a device that provides voice/data connectivity to the user, such as a handheld device with a wireless connection function, a vehicle-mounted device, and the like.
  • some examples of terminals can be: mobile phones, tablet computers, computers with wireless transceiver functions (such as notebook computers, PDAs, etc.), mobile internet devices (mobile internet devices, MIDs), virtual reality (virtual reality, VR) devices , Augmented reality (AR) equipment, wireless terminals in industrial control, wireless terminals in unmanned driving, wireless terminals in telemedicine, wireless terminals in smart grids, wireless terminals in transportation security, and smart cities wireless terminals, wireless terminals in smart homes, cellular phones, cordless phones, session initiation protocol (SIP) phones, wireless local loop (WLL) stations, personal digital assistants (personal digital assistants) PDA), handheld devices with wireless communication capabilities, computing devices or other processing devices connected to wireless modems, in-vehicle devices, wearable devices, terminal devices in 5G networks or future evolution of public land mobile networks , PLM
  • wearable devices can also be called wearable smart devices, which is a general term for the intelligent design of daily wear and the development of wearable devices using wearable technology, such as glasses, gloves, watches, clothing and shoes.
  • a wearable device is a portable device that is worn directly on the body or integrated into the user's clothing or accessories.
  • Wearable device is not only a hardware device, but also realizes powerful functions through software support, data interaction, and cloud interaction.
  • wearable smart devices include full-featured, large-scale, complete or partial functions without relying on smart phones, such as smart watches or smart glasses, and only focus on a certain type of application function, which needs to cooperate with other devices such as smart phones. Use, such as all kinds of smart bracelets, smart jewelry, etc. for physical sign monitoring.
  • the terminal device may also be a terminal device in an internet of things (Internet of things, IoT) system.
  • IoT Internet of things
  • IoT is an important part of the development of information technology in the future. Its main technical feature is to connect items to the network through communication technology, so as to realize the intelligent network of human-machine interconnection and interconnection of things.
  • IoT technology can achieve massive connections, deep coverage, and terminal power saving through, for example, narrow band (NB) technology.
  • NB narrow band
  • terminal equipment can also include sensors such as smart printers, train detectors, and gas stations.
  • the main functions include collecting data (part of terminal equipment), receiving control information and downlink data of access network equipment, and sending electromagnetic waves to the access network.
  • the device transmits upstream data.
  • FIG. 1 shows a schematic diagram of a communication system 100 suitable for the method provided by this embodiment of the present application.
  • the communication system 100 may include at least one access network device, such as the access network device 101 shown in FIG. 1 ; the communication system 100 may also include at least one terminal device, as shown in FIG. 1 .
  • Terminal devices 102 to 107 may be mobile or stationary.
  • Each of the access network device 101 and one or more of the terminal devices 102 to 107 may communicate over a wireless link.
  • Each access network device can provide communication coverage for a specific geographic area and can communicate with terminal devices located within the coverage area. For example, the access network device may send configuration information to the terminal device, and the terminal device may send uplink data to the access network device based on the configuration information; for another example, the access network device may send downlink data to the terminal device. Therefore, the access network device 101 and the terminal devices 102 to 107 in FIG. 1 constitute a communication system.
  • D2D technology can be used to realize direct communication between terminal devices.
  • D2D technology can be used for direct communication between terminal devices 105 and 106 and between terminal devices 105 and 107 .
  • Terminal device 106 and terminal device 107 may communicate with terminal device 105 individually or simultaneously.
  • the terminal devices 105 to 107 can also communicate with the access network device 101, respectively. For example, it can communicate directly with the access network device 101.
  • the terminal devices 105 and 106 in FIG. 1 can directly communicate with the access network device 101; they can also communicate with the access network device 101 indirectly, such as the terminal device in FIG. 107 communicates with the access network device 101 via the terminal device 105 .
  • FIG. 1 exemplarily shows an access network device, a plurality of terminal devices, and communication links between the communication devices.
  • the communication system 100 may include multiple access network devices, and the coverage of each access network device may include other numbers of terminal devices, such as more or less terminal devices. This application does not limit this.
  • Each of the above communication devices may be configured with multiple antennas.
  • the plurality of antennas may include at least one transmit antenna for transmitting signals and at least one receive antenna for receiving signals.
  • each communication device additionally includes a transmitter chain and a receiver chain, which can be understood by those of ordinary skill in the art, all of which may include multiple components (eg, processors, modulators, multiplexers) related to signal transmission and reception. , demodulator, demultiplexer or antenna, etc.). Therefore, the multi-antenna technology can be used for communication between the access network device and the terminal device.
  • the wireless communication system 100 may further include other network entities such as a network controller, a mobility management entity, and the like, which are not limited in this embodiment of the present application.
  • network entities such as a network controller, a mobility management entity, and the like, which are not limited in this embodiment of the present application.
  • Antenna port referred to as port.
  • An antenna port can be one physical antenna or a weighted combination of multiple physical antennas.
  • the antenna port may include a transmit antenna port and a receive antenna port.
  • the transmit antenna port can be understood as the transmit antenna identified by the receiving end, or the transmit antenna that can be distinguished in space. Signals transmitted by the same transmit antenna port experience the same channel environment. The receiving end can perform channel estimation based on this to demodulate the transmitted signal.
  • the transmit antenna port may be an independent transceiver unit, or may be a reference signal port.
  • One reference signal port may correspond to one reference signal.
  • the reference signal port may include, but is not limited to, a channel state information reference signal (CSI-RS) port, a demodulation reference signal (demodulation reference signal, DMRS) port, and the like. This application does not limit this.
  • CSI-RS channel state information reference signal
  • DMRS demodulation reference signal
  • the receiving antenna port can be understood as the receiving antenna recognized by the receiving end, or the receiving antenna that can be distinguished in space.
  • the receive antenna ports and transmit antenna ports may be used, for example, to subsequently determine the channel matrix. This application also does not limit this.
  • Spatial vector It can also be called angle vector, beam vector, etc.
  • Each element in the spatial vector can be used to represent the weight of each transmit antenna port. Based on the weight of each transmit antenna port represented by each element in the space vector, the transmit energy of each transmit antenna port is linearly superimposed to form a region with strong energy in a certain direction in space.
  • a space vector can be a vector of length T.
  • T can represent the number of transmit antenna ports, and T ⁇ 1 and is an integer.
  • the space vector can be, for example, a column vector or a row vector of length T. This application does not limit this. In the following, for the convenience of understanding and description, it is assumed that the received space vector is a column vector of length T.
  • a spatial vector of length T it contains T spatial weights (or simply, weights), and the T weights can be used to weight the T transmit antenna ports, so that the T transmit antennas
  • the reference signal emitted by the port has a certain spatial directivity, so as to realize beamforming. Therefore, precoding the signal based on different spatial vectors is equivalent to performing beamforming on the transmit antenna port based on different spatial vectors, so that the transmitted signals have different spatial directivities.
  • the spatial vector is a discrete Fourier transform (DFT) vector.
  • DFT vector may refer to a column vector in the DFT matrix.
  • the space vector is the conjugate transpose of the DFT vector.
  • the DFT conjugate transpose vector may refer to a column vector in the conjugate transpose matrix of the DFT matrix.
  • the spatial vector is an oversampled DFT vector.
  • An oversampled DFT vector may refer to a vector in an oversampled DFT matrix.
  • the spatial vector is the conjugate transpose of the oversampled DFT vector.
  • the airspace vector may be, for example, the type II (type II) in the 3rd Generation Partnership Project (3GPP) technical specification (TS) 38.214 release 15 (release 15, R15) or R16.
  • 3GPP 3rd Generation Partnership Project
  • TS technical specification
  • the spatial vector can be a 2D-DFT vector or an oversampled 2D-DFT vector.
  • Frequency domain vector It can also be called delay vector.
  • the frequency domain vector can be used to represent the vector of the channel's variation law in the frequency domain.
  • Each frequency domain vector can represent a variation law. Since the signal is transmitted through the wireless channel, there are multiple paths from the transmitting antenna to the receiving antenna. The multipath delay causes frequency selective fading, which is the change of the frequency domain channel. Therefore, the variation law of the channel in the frequency domain caused by the delay on different transmission paths can be represented by different frequency domain vectors.
  • the frequency domain vector may be a vector of length N.
  • N represents the number of frequency domain units. N ⁇ 1, and is an integer.
  • the frequency domain vector can be, for example, a column vector or row vector of length N. This application does not limit this. In the following, for the convenience of understanding and description, it is assumed that the frequency domain vector is a column vector of length N.
  • the frequency domain vector is a DFT vector or a conjugate transpose of a DFT vector.
  • the frequency domain vector is an oversampled DFT vector or a conjugate transpose vector of an oversampled DFT vector.
  • a DFT vector is a vector in a DFT matrix.
  • the DFT matrix is a set of orthogonal basis vectors, so any two vectors in the DFT matrix can be orthogonal to each other.
  • the oversampled DFT matrix may be obtained by oversampling the DFT matrix.
  • the vectors in the oversampling DFT matrix can be divided into multiple subsets, and the adjacent DFTs in each subset can also be mutually orthogonal, and different subsets can be non-orthogonal.
  • Frequency domain unit It can be used to represent different granularity of frequency domain resources.
  • the frequency domain unit may include, but is not limited to, a subband, a resource block (RB), a subcarrier, a resource block group (RBG), or a precoding resource block group (PRG), etc. .
  • the channel corresponding to the frequency domain unit can be used to determine the precoding corresponding to the frequency domain unit for subsequent data transmission.
  • the channel corresponding to the frequency domain unit can be obtained by measuring the reference signal on the frequency domain unit.
  • the channel corresponding to the frequency domain unit may be determined based on the channel information corresponding to the frequency domain unit fed back by the terminal device, or may be determined based on the channel information corresponding to the frequency domain unit near the frequency domain unit fed back by the terminal device. . This application does not limit this.
  • the channel corresponding to the frequency domain unit may be simply referred to as the channel of the frequency domain unit.
  • Compressed sensing Also known as compressive sampling or sparse sampling, is a technique for finding sparse solutions to underdetermined linear systems for acquiring and reconstructing sparse or compressible signals.
  • the channel information is first converted into sparse time-delay domain channel information through Fourier transform (corresponding to the dictionary matrix in compressed sensing), and then the sparse channel information is multiplied by the observation matrix.
  • the transformed delay-domain channel information is converted into a low-dimensional representation, thereby realizing compression (or coding) from high-dimensional to low-dimensional.
  • Neural network an algorithmic mathematical model that imitates the behavioral characteristics of animal neural networks and performs distributed parallel information processing.
  • a neural network can be composed of neural units, and a neural unit can refer to an operation unit that takes x s and an intercept 1 as input, and the output of the operation unit can be:
  • W s is the weight of x s
  • b is the bias of the neural unit.
  • f is the activation function of the neural unit, which is used to introduce nonlinear characteristics into the neural network to convert the input signal in the neural unit into an output signal.
  • the output signal of the activation function can be used as the input of the next convolutional layer, and the activation function can be sigmoid, Tanh, Relu or LeakyRelu function, etc.
  • a neural network is a network formed by connecting a plurality of the above single neural units together, that is, the output of one neural unit can be the input of another neural unit.
  • the input of each neural unit can be connected with the local receptive field of the previous layer to extract the features of the local receptive field, and the local receptive field can be an area composed of several neural units.
  • the reference signal may be used for channel measurement and then for demodulation.
  • the reference signal in this embodiment of the present invention may include a demodulation reference signal (demodulation reference signal, DMRS).
  • DMRS demodulation reference signal
  • reference signal as a reference signal used for channel measurement or channel estimation, is only an exemplary illustration, and should not constitute any limitation to the embodiments of the present invention, and the present application does not exclude the use of other protocols in existing or future protocols. may be used in place of the reference signal to achieve the same or similar function.
  • FIG. 2 shows an example of a schematic structural diagram of an access network device sending a CSI-RS to obtain CSI.
  • the receiving end estimates and obtains the CSI matrix, it performs Fourier transform on the CSI matrix, that is, converts the CSI matrix into sparse delay information, and then selects the sparse delay information for the sparse delay information. Part of the strong beam, quantized, and through the pre-configured codebook, select the appropriate indicator to feed back to the sender.
  • the codebook-based CSI feedback scheme implements CSI coding through beam selection and quantization, and directly discards secondary information, resulting in the loss of a large amount of useful information, poor CSI recovery accuracy, and negative impact on channel capacity.
  • Figure 2 shows a schematic structural diagram of an example of compressive sensing-based feedback CSI.
  • the compressive sensing coding module is first passed, for example, it is first matched with the dictionary matrix. Multiply, convert the channel information into sparse delay channel information, and then convert the sparse delay information into a low-dimensional representation by multiplying with the observation matrix, so as to realize coding.
  • a compressive sensing decoding module is used, for example, an iterative method is usually used to restore channel information.
  • Commonly used decoding methods include orthogonal matching pursuit (OMP), approximate message passing (approximate message passing, AMP) etc.
  • OMP orthogonal matching pursuit
  • AMP approximate message passing
  • LAMP Learned Approximate Message Passing
  • the feedback CSI scheme based on compressed sensing can map high-dimensional signals into low-dimensional space for signals that are codeable or sparse in a certain transform domain, so that the signal is sampled at a frequency much lower than the Nyquist sampling frequency, Compared with the codebook based scheme, the recovered signal has higher quality.
  • This scheme can obtain better encoding and decoding performance with lower computational cost, and has better adaptability to various channel environments, that is, better generalization ability.
  • This scheme maps the channel information to a low-dimensional representation through linear transformations such as Fourier transform and matrix multiplication at the encoding end. However, the linear transformation has limited ability to extract information features. Under high compression ratios, this scheme has the ability to encode and decode. The performance is poor, and the recovered channel information is distorted.
  • the data-driven artificial intelligence technology represented by deep learning has powerful feature extraction capabilities, and can obtain better results in channel information encoding and decoding.
  • the encoding end uses a layer of The 3x3 convolutional neural network and the CsiNet module of a fully connected linear layer network
  • the decoding end uses multiple RefineNet modules to superimpose, which can exceed the performance of the current channel information encoding and decoding scheme based on compressed sensing.
  • artificial intelligence-based encoding and decoding schemes require a large amount of computation and require a large amount of training data.
  • the trained neural network usually reflects more characteristics of the training data, and performs poorly for data that deviates greatly from the characteristics of the training data.
  • FIG. 3 shows a schematic flowchart of an example of a method 100 for feeding back CSI.
  • dimensionality reduction processing is performed on CSI based on compressed sensing.
  • the CSI is multiplied by a dictionary matrix.
  • the dictionary matrix is used to realize the sparseness of the CSI.
  • the dictionary matrix can be a DFT or FFT transformation matrix, or it can be other specially designed matrix, which is not limited in this application.
  • the observation matrix may be a Gaussian random sampling matrix, a matrix obtained by training data using a gradient descent method in deep learning, or other specially designed matrices, which are not limited in this application.
  • the dictionary matrix and the observation matrix are non-sparse matrices, that is, the number of non-zeros in the matrix is greater than 50%.
  • dimensionality reduction processing is performed on the CSI by means of a neural network, which includes a fully connected linear layer, thereby generating a dimensionality-reduced CSI.
  • the fully connected linear layer is mathematically represented as matrix multiplication, and fully connected means that the matrix is a non-sparse matrix. Therefore, the neural network is used to reduce the dimension of CSI, which can be understood as combining CSI with at least one matrix.
  • Perform linear multiplication For example, the first layer of the neural network is a DFT or fast Fourier transform (fast Fourier transform, FFT) calculation matrix, and the second layer is a Gaussian random sampling matrix.
  • Multiple matrices can also be combined to form a matrix multiplication, and the series connection of multiple fully connected linear layers can also be equivalent to a fully connected linear layer.
  • a truncation operation is included in the dimensionality reduction processing on the CSI.
  • a truncation operation may be performed, and then the linear multiplication with the next matrix may be performed.
  • CSI is multiplied by the dictionary matrix
  • it is truncated and multiplied by the observation matrix.
  • the information linearly multiplied with the dictionary matrix has high sparsity, so that the dimension of the information can be further reduced by means of direct truncation.
  • CSI is input to one of the fully connected linear layers of the neural network, truncated before entering the next layer.
  • the neural network may be obtained based on the dimensionality reduction processing method above the cascade and the neural network is uniformly trained, or it can be understood that the dimensionality reduction processing method above the cascade and the neural network are uniformly trained.
  • the dimension of the CSI after dimensionality reduction is related to the input parameters of the neural network, and the parameters of the matrix used in the above dimensionality reduction processing need to correspond to the dimensions of the input parameters of the neural network.
  • a neural network is used to encode the dimension-reduced CSI.
  • the neural network may include a nonlinear activation function, and the activation function may be Sigmoid, Tanh, Relu or LeakyRelu, etc.
  • the neural network may also include convolution, Full connection, pooling or batch normalization are not limited in this application.
  • the coding structure of the present application can be expressed as a whole trainable
  • the 1st to Nth layers of the neural network are fully connected linear operation layers, and the N+1th to Pth layers may contain nonlinear activation functions.
  • the decoding side of the method 100 for feeding back CSI in the present application may be composed of two cascaded structures, and the order is a decoding structure oriented to a neural network and a decoding structure oriented to dimensionality reduction processing.
  • the neural network may include a nonlinear activation function, and the activation function may be Sigmoid, Tanh, Relu, or LeakyRelu, etc.
  • the neural network may also include convolution, full connection, pooling, or batch normalization, which is not limited in this application.
  • the decoding structure for dimensionality reduction processing can be a traditional OMP algorithm, an AMP algorithm or a data-driven compressed sensing decoding algorithm, that is, a decoding structure formed by an iterative algorithm expansion, such as Learned approximate message-passing LAMP algorithm, the structure can obtain parameters in the mechanism through data training.
  • the encoding and decoding structure at this time can be seen in Figure 4.
  • the CSI sequentially passes through the compressed sensing encoding module and the artificial intelligence encoding module, and the two modules respectively achieve a partial compression ratio.
  • the decoding end goes through the artificial intelligence decoding module and the compressed sensing decoding module in sequence, firstly recovering part of the data encoding by the method of artificial intelligence, and then recovering the data by the method of compressed sensing.
  • the decoding structure for dimension reduction processing can also be a decoding structure based on a neural network, that is, the decoding structure at this time can also be regarded as a whole. Trained neural network.
  • the above neural networks are obtained through training data.
  • the structure of dimensionality reduction processing, the neural network, the decoding structure oriented to neural network and the decoding structure oriented to dimensionality reduction processing can be combined. Connected in series for a unified end-to-end training.
  • FIG. 5 shows a schematic structural diagram of an example of a specific feedback CSI, which is only an example and does not constitute any limitation to the present application.
  • the following takes 32 antennas and 1024 carriers of the access network device as an example, and one receiving antenna of the terminal device as an example, and the time-domain CSI estimated by the terminal device is a three-dimensional matrix of 1024 ⁇ 32 ⁇ 2 as an example.
  • the working process of the feedback CSI structure is illustrated by an example, and the following description does not constitute any limitation to the present application.
  • the 1024 ⁇ 32 ⁇ 2 three-dimensional matrix of the time-domain CSI passes through a compressed sensing coding module, wherein the compressed sensing coding module includes a DFT module, an observation matrix module, and optionally, a matrix truncation module may also be included between the DFT module and the observation module. .
  • the time-domain CSI enters the DFT module for DFT conversion, and becomes a three-dimensional matrix with sparsity, that is, the dimension remains unchanged at this time, which is still 1024 ⁇ 32 ⁇ 2.
  • the three-dimensional matrix transformed by DFT since the three-dimensional matrix transformed by DFT has sparseness, the three-dimensional matrix may be truncated into 32 ⁇ 32 ⁇ 2 by means of truncation.
  • the dimension of the truncated CSI matrix is reshaped into a 2048 ⁇ 1 vector, and then multiplied by the observation matrix, that is, the CSI encoding vector after compressed sensing compression is obtained, which is an example of the second information in this application.
  • the observation matrix can be obtained by Gaussian random sampling. Taking the size of the observation matrix as 2048 ⁇ 512 as an example, the size of the CSI coding vector after compressed sensing coding is 512 ⁇ 1.
  • the CSI encoding vector after compressed sensing encoding enters the AI encoding module.
  • the "(full connection + LeakyRELU) ⁇ 2 + full connection” structure is used as an example to implement the second-level encoding of CSI to form a feedback vector.
  • the first is the decoding module oriented to the AI encoding module.
  • the "(full connection + LeakyRELU) ⁇ 5 + full connection” structure is used as an example to realize the first-level decoding of CSI, and obtain 512 ⁇ A vector of 1.
  • enter the decoding module oriented to the compression coding perception module that is, enter the AMP algorithm module to obtain a 2048 ⁇ 1 vector, which is then resized into a 32 ⁇ 32 ⁇ 2 matrix through dimension reorganization, and then undergoes a zero-filling operation corresponding to the truncation operation at the encoding end.
  • a 1024 ⁇ 32 ⁇ 2 matrix is obtained, and finally the restored time-domain CSI matrix is obtained through IDFT.
  • the CSI-RS is encoded and fed back to the access network device to save resources.
  • the present application proposes an implementation method for encoding CSI-RS, which can save resources and reduce computing power consumption of the terminal device.
  • FIG. 6 shows a schematic flowchart of an example of a method 200 for encoding CSI-RS at the access network device side.
  • the access network device performs first encoding on the CSI-RS to generate first information.
  • the CSI-RS is used to obtain the CSI corresponding to the channel between the access network device and the terminal device.
  • the access network device performs the first encoding on the CSI-RS in a compressed sensing manner.
  • the access network equipment only reserves M groups of resources, where N and M are both positive integers, M can be much smaller than N, and the compressed sensing method encodes the resources.
  • CSI-RS can be understood as mapping the CSI-RS to the radio frequency link through compressed sensing, that is, the M CSI-RS can be mapped through the sensing matrix A of compressed sensing (for example, the dimension of the matrix is M ⁇ N). Mapped to N radio links.
  • the access network device needs to use N resources to transmit the CSI-RS through N radio frequency links.
  • the RS matrix is observed to obtain the CSI matrix.
  • the CSI-RS can be encoded by a sensing matrix A with a dimension of M ⁇ N (M ⁇ N), which is equivalent to multiplying the CSI-RS matrix by a sensing matrix A.
  • Matrix A the terminal equipment observes an M ⁇ F ⁇ 2 CSI-RS matrix, and the access network equipment only needs to use M resources to transmit CSI-RS through N radio frequency links, where F is the number of sampling points of the channel , 2 represents the real part and imaginary part of the CSI matrix element, the above description is only an example, this application reduces the occupation of resources when transmitting CSI-RS by using compressed sensing to encode the CSI-RS, and the parameter F can also be other The combination of parameters is not limited in this application.
  • the perception matrix A may be a Gaussian random sampling matrix, a matrix obtained by training data using a gradient descent method in deep learning, or another specially designed matrix, which is not limited in this application.
  • perception matrix A may also be expressed in the form of multiplying multiple matrices, which is not limited in this application.
  • the access network device uses a neural network to perform the first encoding on the CSI-RS, and the neural network includes a fully connected linear layer to generate the first information.
  • the fully connected linear layer is mathematically represented as matrix multiplication.
  • Fully connected means that the matrix is a non-sparse matrix. Therefore, the neural network is used to encode the CSI-RS first, which can be understood as the CSI-RS. Multiplies linearly with at least one matrix.
  • the first layer of the neural network is a DFT or fast Fourier transform calculation matrix
  • the second layer is a Gaussian random sampling matrix, that is, the CSI-RS is linearly multiplied by the two matrices.
  • Multiple matrices can also be combined to form a matrix obtained by multiplication, and the series connection of multiple fully connected linear layers can also be equivalent to a fully connected linear layer.
  • the access network device sends the first information to the terminal device through N radio frequency links.
  • the terminal device receives the first information sent by the access network device to the terminal device through the N radio frequency links.
  • the information received by the terminal device has implicitly included CSI.
  • the terminal device determines the second information based on the first information and the CSI-RS.
  • the terminal device After the access network device uses M resources to send the first coded CSI-RS to the terminal device through N radio frequency links, the terminal device carries the first coded CSI-RS. observation, thereby determining the second information of the implicit CSI.
  • the terminal device may generate second information according to the first information and CSI-RS, where the CSI-RS may be pre-configured in the terminal device , that is, the terminal device can determine the second information of implicit CSI after receiving the compressed CSI-RS and the known CSI-RS transmitted through the channel.
  • the terminal device feeds back CSI to the access network device, it can be understood as feeding back the observation result of the first information to the access network device, that is, the observation of the CSI-RS using M resources through N radio frequency links As a result, the CSI is obtained by decoding at the access network device side.
  • the terminal device performs second encoding on the second information to generate third information.
  • the second encoding may be encoding based on the first neural network, that is, the terminal device encodes the second information by means of a neural network, that is, the first neural network, to generate the third information.
  • the parameters of the first neural network are related to the number of sampling points F of the channel and the number of resources M occupied by the first information, where F and M are both positive integers, and the resources include at least one of the following resources: time domain resources , frequency domain resources or code domain resources.
  • the dimension of the CSI matrix H may be N ⁇ F ⁇ 2, N is the number of radio frequency links of the access network device, F is the number of sampling points of the channel, M is the number of resources occupied by the compressed first information, 2 represents the real and imaginary parts of the CSI matrix elements. Therefore, it can be seen that the dimension of the matrix Y is M ⁇ F ⁇ 2, that is to say, the dimension of the input matrix of the first neural network is M ⁇ F ⁇ 2.
  • the parameters of the network are related to the number F of sampling points of the channel and the number M of resources occupied by the first information.
  • the fully connected linear layer can also be expressed in the form of a matrix, so the dimension of the input matrix of the first neural network is similar to the above, in This will not be repeated here.
  • the terminal device sends third information to the access network device, and correspondingly, the access network device receives the third information from the terminal device.
  • the access network device performs the second decoding on the third information.
  • the second decoding corresponds to the second encoding of the second information by the terminal device.
  • the terminal device performs the second encoding on the matrix Y with a dimension of M ⁇ F ⁇ 2, and the second decoding can restore the third information to the matrix Y.
  • the second decoding may be a decoding method based on a neural network
  • the neural network may include at least one layer of nonlinear activation function operations, such as sigmoid, Tanh, Relu or LeakyRelu functions, etc.
  • the neural network may also include other operations, For example, convolution, full connection, pooling, batch normalization, etc.
  • the terminal device first decodes the third information.
  • the access network device performs first encoding on the CSI-RS to generate the first information
  • the terminal device determines the second information according to the first information and the CSI-RS
  • the terminal device sends the access network device to The third information, where the second information and the third information are the same
  • the first decoding is the decoding corresponding to the first encoding, for example, the decoding corresponding to the compressed sensing method, or the second neural network method including the fully connected linear layer. Encoded decoding.
  • the decoding corresponding to the compressed sensing method can be a traditional OMP or AMP algorithm, or a data-driven compressed sensing decoding algorithm, that is, a traditional iterative algorithm is expanded to form a decoding structure, and the parameters in the decoding structure can be passed through Data training methods are obtained, such as the LAMP algorithm.
  • the decoding corresponding to the second neural network method can be a decoding method based on a neural network.
  • the neural network can include at least one layer of nonlinear activation function operations, such as sigmoid, Tanh, Relu or LeakyRelu functions. Including other operations, such as convolution, full connection, pooling, batch normalization, etc.
  • the access network device performs first coding on the CSI-RS to generate the first information
  • the terminal device determines the second information according to the first information and the CSI-RS
  • the terminal device performs the second information on the second information.
  • the second encoding is used to generate the third information
  • the access network device performs the second decoding on the third information
  • the first decoding also only corresponds to the decoding of the first encoding
  • the access network device performs the first decoding on the third information
  • the access network device performs first coding on the CSI-RS to generate the first information
  • the terminal device determines the second information according to the first information and the CSI-RS
  • the terminal device performs the second information on the second information.
  • the second encoding is used to generate the third information, but the access network device does not decode the third information corresponding to the second encoding.
  • the first decoding can be understood as the deep fusion of the decoding corresponding to the first encoding and the decoding corresponding to the second encoding , which is expressed as a decoding structure based on a deep neural network, rather than a cascade of two decoding structures.
  • the neural network can be used to decode the data encoded by the compressed sensing method, if the first encoding is the encoding based on the compressed sensing method, the first decoding can also be presented as a fusion decoding structure.
  • the access network device sends the encoded CSI-RS to the terminal device through N radio frequency links, and the terminal device does not need to perform a decoding operation on the encoded CSI-RS, but then performs a decoding operation.
  • the one-time encoding is sent to the access network device to be decoded by the access network device to obtain the CSI, which can reduce resource occupation during the process of obtaining the CSI.
  • FIG. 7 shows a schematic structural diagram of an example of obtaining CSI.
  • the first encoding module on the device side of the access network encodes the CSI-RS, and uses M resources to send the CSI-RS to the terminal device through the N radio frequency links RFC, and the terminal device interprets the information sent through the N radio frequency links.
  • Channel estimation is performed to determine the Y matrix, which implicitly includes the CSI matrix H.
  • the terminal device further includes a second encoding module, configured to perform second encoding on the Y matrix, and send the second encoded Y matrix to the access network device.
  • the access network device includes a first decoding module, and the first decoding module is configured to perform decoding for the first encoding module, or the first decoding module is configured to perform decoding for the first encoding module and
  • the access network device includes a first decoding module and a second decoding module, the first decoding module is used for decoding the first encoding module, and the second decoding module is used for decoding.
  • the Y matrix after the second encoding first enters the second decoding module to restore the Y matrix, and then enters the first decoding module to obtain the CSI matrix H.
  • N, M, etc. in this application can also be understood as N groups and M groups, which are not limited in this application.
  • FIG. 8 is a schematic block diagram of an apparatus for transmitting data provided by an embodiment of the present application.
  • the apparatus 1000 may include a processing unit 1100 and a transceiver unit 1200 .
  • the apparatus 1000 may correspond to the encoding apparatus in the above method embodiments, for example, may be an encoder, or a component (such as a circuit, a chip, or a chip system, etc.) configured in the encoder.
  • a component such as a circuit, a chip, or a chip system, etc.
  • the apparatus 1000 may correspond to the encoding apparatus in the method according to the embodiment of the present application, and the apparatus 1000 may include a unit for executing the method performed by the encoding apparatus in FIG. 3 .
  • each unit in the apparatus 1000 and the above-mentioned other operations and/or functions are respectively for realizing the corresponding flow in FIG. 3 .
  • the processing unit 1100 may be configured to perform the steps in FIG. 3 to perform dimensionality reduction processing on CSI, and may also be configured to perform the neural network-based dimension reduction in FIG. 3 .
  • the CSI is encoded, and the transceiver unit 1200 can be used to receive the CSI. It should be understood that the specific process of each unit performing the above-mentioned corresponding steps has been described in detail in the above-mentioned method embodiments, and is not repeated here for brevity.
  • the communication apparatus 1000 may correspond to a terminal device according to an embodiment of the present application, and the communication apparatus 1000 may include a unit for executing the method performed by the terminal device in the method in FIG. 6 .
  • each unit in the communication apparatus 1000 and the above-mentioned other operations and/or functions are respectively for realizing the corresponding flow in FIG. 6 .
  • the processing unit 1100 may be configured to execute the steps in FIG. 6 to determine the second information based on the first information and the CSI-RS, and the transceiver unit 1200 may be configured to execute the steps in FIG. 6 .
  • the steps in receive the first information sent by the access network equipment through N radio frequency links, the first information is generated by first encoding the channel state information reference signal CSI-RS, and the CSI-RS is used to obtain the channel information CSI. .
  • the processing unit 1100 may also be configured to perform the second encoding of the second information in FIG. 6 to generate the third information.
  • the transceiver unit 1200 may also be configured to perform sending third information to the access network device in FIG.
  • a decoding includes decoding corresponding to the first encoding.
  • the transceiver unit 1200 in the apparatus 1000 may be implemented by a transceiver, for example, it may correspond to the transceiver 2020 in the apparatus 2000 shown in FIG. 9 or the transceiver 2020 shown in FIG. 10 .
  • the transceiver 3020 in the terminal device 3000, the processing unit 1100 in the apparatus 1000 may be implemented by at least one processor, for example, may correspond to the processor 2010 in the apparatus 2000 shown in FIG. 9 or the terminal shown in FIG. 10 Processor 3010 in device 3000.
  • the transceiver unit 1200 in the apparatus 1000 may be implemented through an input/output interface, a circuit, etc.
  • the processing unit 1100 in the apparatus 1000 may be implemented through the Implementation of a processor, microprocessor or integrated circuit integrated on a chip or system of chips.
  • the apparatus 1000 may correspond to the decoding device in the above method embodiments, for example, may be a decoder, or a component (such as a circuit, a chip, or a chip system, etc.) configured in the decoder.
  • the apparatus 1000 may correspond to a decoding apparatus according to an embodiment of the present application, and the apparatus 1000 may include a unit for executing a method performed by the decoding apparatus. Moreover, each unit in the apparatus 1000 and the other operations and/or functions mentioned above are respectively intended to implement the corresponding operations of the method 400 .
  • the processing unit 1100 can be used to decode the encoded CSI based on the neural network for executing the method shown in FIG. 3
  • the transceiver unit 1200 can be used to execute the method shown in FIG. 3 .
  • the received encoded CSI It should be understood that the specific process of each unit performing the above-mentioned corresponding steps has been described in detail in the above-mentioned method embodiments, and for the sake of brevity, it will not be repeated here.
  • the transceiver unit 1200 in the apparatus 1000 may be implemented by a transceiver, for example, it may correspond to the transceiver 2020 in the apparatus 2000 shown in FIG. 9 or the transceiver 2020 shown in FIG. 11 .
  • the RRU 4100 in the access network device 4000, the processing unit 1100 in the apparatus 1000 may be implemented by at least one processor, for example, may correspond to the processor 2010 in the apparatus 2000 shown in FIG. 9 or the processor 2010 shown in FIG. 11 .
  • the transceiver unit 1200 in the apparatus 1000 may be implemented through input/output interfaces, circuits, etc., and the processing unit 1100 in the apparatus 1000 may be implemented through the Implementation of a processor, microprocessor or integrated circuit integrated on a chip or system of chips.
  • the communication apparatus 1000 may correspond to the access network device in FIG. 6 according to an embodiment of the present application, and the communication apparatus 1000 may include a unit for executing the method performed by the access network device in the method in FIG. 6 . Moreover, each unit in the communication apparatus 1000 and the other operations and/or functions mentioned above are respectively for realizing the corresponding flow of the method in FIG. 6 .
  • the processing unit 1100 may be configured to perform the steps in FIG. 6 to perform first encoding on the channel state information reference signal CSI-RS to generate first information, the CSI-RS - The RS is used to obtain the channel state information CSI corresponding to the channel between the access network device and the terminal device.
  • the transceiver unit 1200 can be used for the steps in FIG. 6 to send the first information to the terminal device through N radio frequency links. is a positive integer, the transceiver unit 1200 is also used for the steps in FIG.
  • the processing unit 1100 may be further configured to perform the steps in FIG. 6 to first decode the third information to generate the CSI, where the first decoding includes decoding. It should be understood that the specific process of each unit performing the above-mentioned corresponding steps has been described in detail in the above-mentioned method embodiments, and for the sake of brevity, it will not be repeated here.
  • the transceiver unit 1200 in the communication apparatus 1000 may be implemented by a transceiver, for example, may correspond to the transceiver 2020 in the communication apparatus 2000 shown in FIG.
  • the RRU 4100 in the base station 4000 shown in FIG. 11 the processing unit 1100 in the communication device 1000 may be implemented by at least one processor, for example, may correspond to the processor 2010 in the communication device 2000 shown in FIG. 9 or FIG. 11
  • the transceiver unit 1200 in the communication device 1000 can be implemented through input/output interfaces, circuits, etc., and the processing unit 1100 in the communication device 1000 It can be implemented by a processor, microprocessor or integrated circuit integrated on the chip or chip system.
  • FIG. 9 is another schematic block diagram of an apparatus 2000 provided by an embodiment of the present application.
  • the apparatus 2000 includes a processor 2010 , a transceiver 2020 and a memory 2030 .
  • the processor 2010, the transceiver 2020 and the memory 2030 communicate with each other through an internal connection path, the memory 2030 is used to store instructions, and the processor 2010 is used to execute the instructions stored in the memory 2030 to control the transceiver 2020 to send signals and / or receive signals.
  • the apparatus 2000 may correspond to the terminal device in the above method embodiments, and may be used to execute various steps and/or processes performed by the network device or the terminal device in the above method embodiments.
  • the memory 2030 may include read only memory and random access memory and provide instructions and data to the processor. A portion of the memory may also include non-volatile random access memory.
  • the memory 2030 may be a separate device or may be integrated in the processor 2010 .
  • the processor 2010 may be configured to execute the instructions stored in the memory 2030, and when the processor 2010 executes the instructions stored in the memory, the processor 2010 is configured to execute each of the foregoing method embodiments corresponding to the network device or the terminal device steps and/or processes.
  • the apparatus 2000 is the terminal device in the foregoing embodiment.
  • the apparatus 2000 is the network device in the foregoing embodiment.
  • the transceiver 2020 may include a transmitter and a receiver.
  • the transceiver 2020 may further include antennas, and the number of the antennas may be one or more.
  • the processor 2010, the memory 2030 and the transceiver 2020 may be devices integrated on different chips.
  • the processor 2010 and the memory 2030 may be integrated in a baseband chip, and the transceiver 2020 may be integrated in a radio frequency chip.
  • the processor 2010, the memory 2030 and the transceiver 2020 may also be devices integrated on the same chip. This application does not limit this.
  • the apparatus 2000 is a component configured in a terminal device, such as a circuit, a chip, a chip system, and the like.
  • the apparatus 2000 is a component configured in a network device, such as a circuit, a chip, a chip system, and the like.
  • the transceiver 2020 may also be a communication interface, such as an input/output interface, a circuit, and the like.
  • the transceiver 2020, the processor 2010 and the memory 2020 can be integrated in the same chip, such as integrated in a baseband chip.
  • FIG. 10 is a schematic structural diagram of a terminal device 3000 provided by an embodiment of the present application.
  • the terminal device 3000 can be applied to the system shown in FIG. 1 to perform the functions of the terminal device in the foregoing method embodiments.
  • the terminal device 3000 includes a processor 3010 and a transceiver 3020 .
  • the terminal device 3000 further includes a memory 3030 .
  • the processor 3010, the transceiver 3020 and the memory 3030 can communicate with each other through an internal connection path to transmit control and/or data signals.
  • the computer program is invoked and executed to control the transceiver 3020 to send and receive signals.
  • the terminal device 3000 may further include an antenna 3040 for sending the uplink data or uplink control signaling output by the transceiver 3020 through wireless signals.
  • the above-mentioned processor 3010 and the memory 3030 can be combined into a processing device, and the processor 3010 is configured to execute the program codes stored in the memory 3030 to realize the above-mentioned functions.
  • the memory 3030 may also be integrated in the processor 3010 or independent of the processor 3010 .
  • the processor 3010 may correspond to the processing unit 1100 in FIG. 8 or the processor 2010 in FIG. 9 .
  • the transceiver 3020 described above may correspond to the transceiver unit 1200 in FIG. 8 or the transceiver 2020 in FIG. 9 .
  • the transceiver 3020 may include a receiver (or called receiver, receiving circuit) and a transmitter (or called transmitter, transmitting circuit). Among them, the receiver is used for receiving signals, and the transmitter is used for transmitting signals.
  • the terminal device 3000 shown in FIG. 7 can implement various processes involving the terminal device in the method embodiment shown in FIG. 3 or FIG. 6 .
  • the operations and/or functions of each module in the terminal device 3000 are respectively to implement the corresponding processes in the foregoing method embodiments.
  • the above-mentioned processor 3010 may be used to perform the actions described in the foregoing method embodiments that are implemented inside the terminal device, and the transceiver 3020 may be used to perform the operations described in the foregoing method embodiments that the terminal device sends to or receives from the network device. action.
  • the transceiver 3020 may be used to perform the operations described in the foregoing method embodiments that the terminal device sends to or receives from the network device. action.
  • the above-mentioned terminal device 3000 may further include a power supply 3050 for providing power to various devices or circuits in the terminal device.
  • the terminal device 3000 may further include one or more of an input unit 3060, a display unit 3070, an audio circuit 3080, a camera 3090, a sensor 3100, etc., the audio circuit Speakers 3082, microphones 3084, etc. may also be included.
  • FIG. 11 is a schematic structural diagram of an access network device provided by an embodiment of the present application, which may be, for example, a schematic structural diagram of a base station.
  • the base station 4000 can be applied to the system shown in FIG. 1 , and performs the functions of the access network device in the foregoing method embodiments.
  • the base station 4000 may include one or more radio frequency units, such as a remote radio unit (RRU) 4100 and one or more baseband units (BBUs) (also referred to as distributed units (DUs). )) 4200.
  • RRU 4100 may be called a transceiver unit, which may correspond to the transceiver unit 1200 in FIG. 5 or the transceiver 2020 in FIG. 6 .
  • the RRU 4100 may also be referred to as a transceiver, a transceiver circuit, or a transceiver, etc., which may include at least one antenna 4101 and a radio frequency unit 4102.
  • the RRU 4100 may include a receiving unit and a sending unit, the receiving unit may correspond to a receiver (or called a receiver, a receiving circuit), and the sending unit may correspond to a transmitter (or called a transmitter, a sending circuit).
  • the RRU 4100 part is mainly used for the transceiver of radio frequency signals and the conversion of radio frequency signals and baseband signals, for example, for sending instruction information to terminal equipment.
  • the part of the BBU 4200 is mainly used to perform baseband processing and control the base station.
  • the RRU 4100 and the BBU 4200 may be physically set together, or may be physically separated, that is, a distributed base station.
  • the BBU 4200 is the control center of the base station, and can also be called a processing unit, which can correspond to the processing unit 1100 in FIG. 8 or the processor 2010 in FIG. 9, and is mainly used to complete baseband processing functions, such as channel coding, multiplexing , modulation, spread spectrum, etc.
  • the BBU processing unit
  • the BBU may be used to control the base station to perform the operation procedure of the network device in the foregoing method embodiments, for example, to generate the foregoing indication information and the like.
  • the BBU 4200 may be composed of one or more boards, and the multiple boards may jointly support a wireless access network (such as an LTE network) of a single access standard, or may respectively support a wireless access network of different access standards.
  • Wireless access network (such as LTE network, 5G network or other network).
  • the BBU 4200 also includes a memory 4201 and a processor 4202.
  • the memory 4201 is used to store necessary instructions and data.
  • the processor 4202 is configured to control the base station to perform necessary actions, for example, to control the base station to execute the operation flow of the network device in the foregoing method embodiments.
  • the memory 4201 and the processor 4202 may serve one or more single boards. That is to say, the memory and processor can be provided separately on each single board. It can also be that multiple boards share the same memory and processor. In addition, necessary circuits may also be provided on each single board.
  • the base station 4000 shown in FIG. 11 can implement each process involving the access network device in the method embodiment shown in FIG. 2 , FIG. 3 or FIG. 4 .
  • the operations and/or functions of each module in the base station 4000 are respectively to implement the corresponding processes in the foregoing method embodiments.
  • the above-mentioned BBU 4200 may be used to perform the actions implemented by the access network equipment described in the foregoing method embodiments, and the RRU 4100 may be used to perform the access network equipment described in the foregoing method embodiments. received action.
  • the RRU 4100 may be used to perform the access network equipment described in the foregoing method embodiments. received action.
  • the base station 4000 shown in FIG. 11 is only a possible form of the access network device, and should not constitute any limitation to the present application.
  • the method provided in this application may be applicable to other forms of access network equipment.
  • it includes AAU, may also include CU and/or DU, or includes BBU and adaptive radio unit (ARU), or BBU; may also be customer terminal equipment (customer premises equipment, CPE), may also be
  • AAU adaptive radio unit
  • BBU adaptive radio unit
  • CPE customer premises equipment
  • the CU and/or DU may be used to perform the actions implemented by the access network device described in the foregoing method embodiments, and the AAU may be used to execute the access network device described in the foregoing method embodiments to send or Action received from the end device.
  • the AAU may be used to execute the access network device described in the foregoing method embodiments to send or Action received from the end device.
  • the present application further provides a processing apparatus, including at least one processor, where the at least one processor is configured to execute a computer program stored in a memory, so that the processing apparatus executes the terminal device or access point in any of the foregoing method embodiments The method performed by the network device.
  • the embodiment of the present application also provides a processing apparatus, which includes a processor and a communication interface.
  • the communication interface is coupled with the processor.
  • the communication interface is used to input and/or output information.
  • the information includes at least one of instructions and data.
  • the processor is configured to execute a computer program, so that the processing apparatus executes the method executed by the terminal device or the access network device in any of the foregoing method embodiments.
  • Embodiments of the present application further provide a processing apparatus, including a processor and a memory.
  • the memory is used to store a computer program
  • the processor is used to call and run the computer program from the memory, so that the processing apparatus executes the execution of the terminal device or the access network device in any of the foregoing method embodiments. method.
  • the above-mentioned processing device may be one or more chips.
  • the processing device may be a field programmable gate array (FPGA), an application specific integrated circuit (ASIC), a system on chip (SoC), or a It is a central processing unit (CPU), a network processor (NP), a digital signal processing circuit (DSP), or a microcontroller (microcontroller unit). , MCU), it can also be a programmable logic device (PLD) or other integrated chips.
  • FPGA field programmable gate array
  • ASIC application specific integrated circuit
  • SoC system on chip
  • MCU microcontroller unit
  • MCU programmable logic device
  • PLD programmable logic device
  • each step of the above-mentioned method can be completed by a hardware integrated logic circuit in a processor or an instruction in the form of software.
  • the steps of the methods disclosed in conjunction with the embodiments of the present application may be directly embodied as executed by a hardware processor, or executed by a combination of hardware and software modules in the processor.
  • the software modules may be located in random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, registers and other storage media mature in the art.
  • the storage medium is located in the memory, and the processor reads the information in the memory, and completes the steps of the above method in combination with its hardware. To avoid repetition, detailed description is omitted here.
  • the processor in this embodiment of the present application may be an integrated circuit chip, which has a signal processing capability.
  • each step of the above method embodiments may be completed by a hardware integrated logic circuit in a processor or an instruction in the form of software.
  • the aforementioned processors may be general purpose processors, digital signal processors (DSPs), application specific integrated circuits (ASICs), field programmable gate arrays (FPGAs) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components .
  • DSPs digital signal processors
  • ASICs application specific integrated circuits
  • FPGAs field programmable gate arrays
  • the methods, steps, and logic block diagrams disclosed in the embodiments of this application can be implemented or executed.
  • a general purpose processor may be a microprocessor or the processor may be any conventional processor or the like.
  • the steps of the method disclosed in conjunction with the embodiments of the present application may be directly embodied as executed by a hardware decoding processor, or executed by a combination of hardware and software modules in the decoding processor.
  • the software modules may be located in random access memory, flash memory, read-only memory, programmable read-only memory or electrically erasable programmable memory, registers and other storage media mature in the art.
  • the storage medium is located in the memory, and the processor reads the information in the memory, and completes the steps of the above method in combination with its hardware.
  • the memory in this embodiment of the present application may be a volatile memory or a non-volatile memory, or may include both volatile and non-volatile memory.
  • the non-volatile memory may be read-only memory (ROM), programmable read-only memory (PROM), erasable programmable read-only memory (EPROM), electrically programmable Erase programmable read-only memory (electrically EPROM, EEPROM) or flash memory.
  • Volatile memory may be random access memory (RAM), which acts as an external cache.
  • RAM random access memory
  • DRAM dynamic random access memory
  • SDRAM synchronous DRAM
  • SDRAM double data rate synchronous dynamic random access memory
  • ESDRAM enhanced synchronous dynamic random access memory
  • SLDRAM synchronous link dynamic random access memory
  • direct rambus RAM direct rambus RAM
  • the present application also provides a computer program product, the computer program product includes: computer program code, when the computer program code is run on a computer, the computer is made to execute the embodiment shown in FIG. 3 .
  • the present application further provides a computer-readable storage medium, where the computer-readable storage medium stores program codes, and when the program codes are run on a computer, the computer executes the program shown in FIG. 3 .
  • the present application further provides a system, which includes the aforementioned one or more terminal devices and one or more network devices.
  • the network equipment in each of the above apparatus embodiments completely corresponds to the terminal equipment and the network equipment or terminal equipment in the method embodiments, and corresponding steps are performed by corresponding modules or units.
  • a processing unit processor
  • processor For functions of specific units, reference may be made to corresponding method embodiments.
  • the number of processors may be one or more.
  • the terminal device may be used as an example of a receiving device
  • the network device may be used as an example of a sending device.
  • the sending device and the receiving device may both be terminal devices or the like. This application does not limit the specific types of the sending device and the receiving device.
  • a component may be, but is not limited to, a process running on a processor, a processor, an object, an executable, a thread of execution, a program, and/or a computer.
  • an application running on a computing device and the computing device may be components.
  • One or more components may reside within a process and/or thread of execution, and a component may be localized on one computer and/or distributed between 2 or more computers.
  • these components can execute from various computer readable media having various data structures stored thereon.
  • a component may, for example, be based on a signal having one or more data packets (eg, data from two components interacting with another component between a local system, a distributed system, and/or a network, such as the Internet interacting with other systems via signals) Communicate through local and/or remote processes.
  • data packets eg, data from two components interacting with another component between a local system, a distributed system, and/or a network, such as the Internet interacting with other systems via signals
  • the disclosed system, apparatus and method may be implemented in other manners.
  • the apparatus embodiments described above are only illustrative.
  • the division of the units is only a logical function division. In actual implementation, there may be other division methods.
  • multiple units or components may be combined or Can be integrated into another system, or some features can be ignored, or not implemented.
  • the shown or discussed mutual coupling or direct coupling or communication connection may be through some interfaces, indirect coupling or communication connection of devices or units, and may be in electrical, mechanical or other forms.
  • the units described as separate components may or may not be physically separated, and components displayed as units may or may not be physical units, that is, may be located in one place, or may be distributed to multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution in this embodiment.
  • each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically alone, or two or more units may be integrated into one unit.
  • the functions, if implemented in the form of software functional units and sold or used as independent products, may be stored in a computer-readable storage medium.
  • the technical solution of the present application can be embodied in the form of a software product in essence, or the part that contributes to the prior art or the part of the technical solution.
  • the computer software product is stored in a storage medium, including Several instructions are used to cause a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application.
  • the aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, ROM), random access memory (Random Access Memory, RAM), magnetic disk or optical disk and other media that can store program codes .

Landscapes

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

Abstract

本申请提供了一种传输数据的方法和装置，该方法包括：接入网设备对信道状态信息参考信号CSI-RS进行第一编码，以生成第一信息，该CSI-RS用于获取接入网设备和该终端设备之间信道对应的信道状态信息CSI；该接入网设备通过N个射频链路向终端设备发送该第一信息，该N为正整数；该接入网设备从该终端设备接收第三信息，该第三信息由终端设备对所述第二信息进行第二编码生成，该第二信息由该终端设备基于该第一信息和该CSI-RS确定；该接入网设备对该第三信息进行第一解码，以生成该CSI，该第一解码包括对应该第一编码的解码。从而，可以降低获取CSI过程中对资源的占用。

Description

一种传输数据的方法和装置
本申请要求于2020年12月8日提交中国专利局、申请号为202011444244.0、申请名称为“一种用于信道信息反馈的复合编解码结构”的中国专利申请的优先权，2021年1月6日提交中国专利局、申请号为202110012794.3、申请名称为“一种传输数据的方法和装置”的中国专利申请的优先权，以及2021年2月25日提交中国专利局、申请号为202110214243.5、申请名称为“一种传输数据的方法和装置”的中国专利申请的优先权，其全部内容通过引用结合在本申请中。
技术领域
本申请涉及无线通信的领域，并且更具体地，涉及一种传输数据的方法和装置。
背景技术
在多天线无线通信系统中，终端设备需要将估计的信道状态信息(channel state information，CSI)反馈给接入网设备，以便接入网设备利用CSI进行多天线预编码，经过预编码之后的无线信号，可以对抗信道失真，提升信道容量。
随着天线数量和带宽的增加，信道状态信息参考信号(CSI-reference signal，CSI-RS)和CSI在传输的过程中需要占用大量宝贵的空口资源。因此，亟需一种获得CSI的方法，减少对资源的占用。
发明内容
本申请提供一种传输数据的方法和装置，能够减少对资源的占用。
第一方面，提供了一种传输数据的方法，该方法可以由接入网设备或接入网设备中的芯片执行，该方法包括：接入网设备对信道状态信息参考信号CSI-RS进行第一编码，以生成第一信息，该CSI-RS用于获取接入网设备和该终端设备之间信道对应的信道状态信息CSI；该接入网设备通过N个射频链路向终端设备发送该第一信息，该N为正整数；该接入网设备从该终端设备接收第三信息，第三信息由终端设备对该第二信息进行第二编码生成，该第二信息由该终端设备基于该第一信息和该CSI-RS确定；该接入网设备对该第三信息进行第一解码，以生成该CSI，该第一解码包括对应该第一编码的解码。
例如，接入网设备通过N个射频链路使用M个资源向终端设备发送经过第一编码的CSI-RS后，终端设备对承载经过第一编码的CSI-RS进行观测，从而确定隐含CSI的第二信息，其中M为正整数。
接入网设备向终端设备发送用于获取CSI的CSI-RS，为了节省宝贵的资源，可以将CSI-RS进行编码后发送给终端设备，即向终端设备发送第一信息，终端设备接收到经过N个射频链路传输的第一信息后，可以根据该第一信息和CSI-RS生成第二信息，其中该CSI-RS可以是预先配置于终端设备中，即，终端设备接收到经过信道传输的第一信息进 行观测，从而获取隐含包括CSI的第二信息。终端设备对第二信息进行第二编码，以生成第三信息。接入网设备接收到第三信息对其解码后获得CSI。
从而，在本申请中，接入网设备将经过编码后的CSI-RS通过N个射频链路发送给终端设备，终端设备对该经过编码后的CSI-RS无需进行解码操作，而是再进行一次编码发给接入网设备由接入网设备进行解码，以获得CSI，可以降低获取CSI过程中对资源的占用。
结合第一方面，在第一方面的某些实现方式中，该方法还包括：该接入网设备对该第三信息进行第二解码，该第二解码与该第二编码对应；该接入网设备对该第三信息进行第一解码，包括：该接入网设备对经过该第二解码的第三信息进行该第一解码。
接入网设备接收到经过两次编码后的信息，可以先针对第二编码进行第二解码，再对经过第二解码后的第三信息针对第一编码进行第一解码，即采用两级解码的结构对第三信息进行解码，以获得CSI。
结合第一方面，在第一方面的某些实现方式中，该第一解码还包括对应该第二编码的解码。
需要说明的是，接入网接收到的第三信息是经过两次编码后的信息，但是对应两次编码的解码结构可以融合为一个结构，即，可以对第三信息进行第一解码，该第一解码既包括对应第一编码的解码，又包括对应第二编码的解码，通过一次解码获得所需要的CSI。
从而，在本申请中，接入网设备可以采用一个融合结构对经过两次编码的信息进行解码，提高了解码的准确度。
结合第一方面，在第一方面的某些实现方式中，该第二编码为基于第一神经网络的编码，该第一神经网络的参数与该信道的采样点数量F和该第一信息占用的资源的数量M有关，该M和F为正整数，M<N，该资源包括以下至少一种：时域资源、频域资源或码域资源。
结合第一方面，在第一方面的某些实现方式中，该第一神经网络的参数为该第一神经网络的输入矩阵的维度。
从而，在本申请中，终端设备可以基于神经网络的方式进行再一次编码后将CSI反馈给网络设备，进一步节省了CSI反馈的资源开销。
结合第一方面，在第一方面的某些实现方式中，该第一编码为基于压缩感知方式的编码，该第一编码使用第一矩阵，该第一矩阵的维度与该M和该N有关。
可选地，该第一矩阵的维度为M×N。
可选地，该第一矩阵为多个矩阵相乘的形式，多个矩阵相乘后的维度为M×N。
从而，在本申请中，针对N个射频链路，接入网设备可以采用压缩感知的方式仅预留M组资源用以发送CSI-RS，减少了发送CSI-RS资源的占用。
结合第一方面，在第一方面的某些实现方式中，该第一编码为基于第二神经网络的编码，该第二神经网络包括全连接线性层，该全连接线性层的参数与该M和该N有关。
例如，全连接线性层可以为至少一个矩阵相乘的形式，形成的矩阵的维度为M×N。
从而，在本申请中，针对N个射频链路，接入网设备可以包括全连接线性层的第二神经网络的方式仅预留M组资源用以发送CSI-RS，减少了发送CSI-RS资源的占用。
结合第一方面，在第一方面的某些实现方式中，该第一解码为基于第三神经网络的解 码，该第三神经网络的参数与该N，该M和该F有关。
可选地，该第三神经网络仅用于对应第一编码的解码，该第三神经网络的参数与N，M和F有关。
可选地，该第三神经网络既用于对应第一编码的解码，也用于对第二编码的解码，该第三神经网络的参数与N，M和F有关。
结合第一方面，在第一方面的某些实现方式中，该第一神经网络的训练方式为串联该第一矩阵、该第一神经网络和该第三神经网络统一训练获得。
结合第一方面，在第一方面的某些实现方式中，该第一神经网络的训练方式为串联该第二神经网络、该第一神经网络和该第三神经网络统一训练获得。
需要说明的是，此时的第三神经网络为对应第一解码和第二解码的融合解码结构。
结合第一方面，在第一方面的某些实现方式中，该第二解码为基于第四神经网络方式的解码，该第一神经网络的训练方式为串联该第第一神经网络和第四神经网络的训练方式，或者为串联第一矩阵、第一神经网络、第四神经网络和第三神经网络的训练方式，或者为串联第二神经网络、第一神经网络、第四神经网络和第三神经网络的训练方式。
从而，在本申请中，接入网设备将经过编码后的CSI-RS通过N个射频链路发送给终端设备，终端设备对该经过编码后的CSI-RS无需进行解码操作，而是再进行一次编码后发给接入网设备进行由接入网设备进行解码，以获得CSI，可以降低获得CSI过程中对资源的占用。
第二方面，提供了一种传输数据的方法，该方法可以由终端设备或终端设备中的芯片执行，该方法包括：终端设备接收接入网设备通过N个射频链路向该终端设备发送的第一信息，该第一信息由该接入网设备对信道状态信息参考信号CSI-RS进行第一编码生成，该CSI-RS用于获取信道信息CSI；该终端设备基于该第一信息和该CSI-RS确定第二信息；所述终端设备对所述第二信息进行第二编码，以生成所述第三信息；该终端设备向该接入网设备发送第三信息，该第三信息用于该接入网设备对该第三信息进行第一解码，以生成该CSI，该第一解码包括对应该第一编码的解码。
接入网设备向终端设备发送用于获取CSI的CSI-RS，为了节省宝贵的资源，可以将CSI-RS进行编码后发送给终端设备，即向终端设备发送第一信息，终端设备接收到经过N个射频链路传输的第一信息后，可以根据该第一信息和CSI-RS生成第二信息，其中该CSI-RS可以是预先配置于终端设备中，即，终端设备接收到经过信道传输的第一信息进行观测，从而获取隐含包括CSI的第二信息。终端设备对第二信息进行第二编码，以生成第三信息。接入网设备接收到第三信息对其解码后获得CSI。
从而，在本申请中，接入网设备将经过编码后的CSI-RS通过N个射频链路发送给终端设备，终端设备对该经过编码后的CSI-RS无需进行解码操作，而是再进行一次编码发给接入网设备由接入网设备进行解码，以获得CSI，可以降低获取CSI过程中对资源的占用。
结合第二方面，在第二方面的某些实现方式中，该第三信息用于该接入网设备对经过第二解码的该第三信息进行该第一解码，以生成该CSI，该第二解码与该第二编码对应。
结合第二方面，在第二方面的某些实现方式中，该第一解码还包括对应该第二编码的解码。
结合第二方面，在第二方面的某些实现方式中，该第二编码为基于第一神经网络的编码，该第一神经网络的参数与该信道的采样点数量F和该第一信息占用的资源的数量M有关，该M和F为正整数，M<N，该资源包括以下至少一种：时域资源、频域资源或码域资源。
结合第二方面，在第二方面的某些实现方式中，该第一神经网络的参数为该第一神经网络的输入矩阵的维度。
结合第二方面，在第二方面的某些实现方式中，该第一编码为基于压缩感知方式的编码，该第一编码使用第一矩阵，该第一矩阵的维度与该M和该N有关。
结合第二方面，在第二方面的某些实现方式中，该第一编码为基于第二神经网络的编码，该第二神经网络包括全连接线性层，该全连接线性层的参数与该M和该N有关。
结合第二方面，在第二方面的某些实现方式中，该第一解码为基于第三神经网络的解码，该第三神经网络的参数与该N，该M和该F有关。
第三方面，提供了一种传输数据的装置，该装置包括：处理单元，用于对信道状态信息参考信号CSI-RS进行第一编码，以生成第一信息，该CSI-RS用于获取接入网设备和该终端设备之间信道对应的信道状态信息CSI；收发单元，用于通过N个射频链路向终端设备发送该第一信息，该N为正整数；该该第三信息由该终端设备对该第二信息进行第二编码生成，该第二信息由该终端设备基于该第一信息和该CSI-RS确定；该处理单元还用于对该第三信息进行第一解码，以生成该CSI，该第一解码包括对应该第一编码的解码。
从而，在本申请中，接入网设备将经过编码后的CSI-RS通过N个射频链路发送给终端设备，终端设备对该经过编码后的CSI-RS无需进行解码操作，而是再进行一次编码发给接入网设备由接入网设备进行解码，以获得CSI，可以降低获取CSI过程中对资源的占用。
结合第三方面，在第三方面的某些实现方式中，该处理单元还用于对该第三信息进行第二解码，该第二解码与该第二编码对应；该处理单元还用于对该第三信息进行第一解码，具体用于对经过该第二解码的第三信息进行该第一解码。
结合第三方面，在第三方面的某些实现方式中，该第一解码还包括对应该第二编码的解码。
结合第三方面，在第三方面的某些实现方式中，该第二编码为基于第一神经网络的编码，该第一神经网络的参数与该信道的采样点数量F和该第一信息占用的资源的数量M有关，该M和F为正整数，M<N，该资源包括以下至少一种：时域资源、频域资源或码域资源。
结合第三方面，在第三方面的某些实现方式中，该第一神经网络的参数为该第一神经网络的输入矩阵的维度。
结合第二方面，在第二方面的某些实现方式中，该第一编码为基于压缩感知方式的编码，该第一编码使用第一矩阵，该第一矩阵的维度与该M和该N有关。
结合第三方面，在第三方面的某些实现方式中，该第一编码为基于第二神经网络的编码，该第二神经网络包括全连接线性层，该全连接线性层的参数与该M和该N有关。
结合第三方面，在第三方面的某些实现方式中，该第一解码为基于第三神经网络的解码，该第三神经网络的参数与该N，该M和该F有关。
第四方面，提供了一种传输数据的装置，该装置包括：收发单元，用于接收接入网设备通过N个射频链路发送的第一信息，该第一信息由接入网设备对信道状态信息参考信号CSI-RS进行第一编码生成，该CSI-RS用于获取信道信息CSI；处理单元，用于基于该第一信息和该CSI-RS确定第二信息；该处理单元还用于对该第二信息进行第二编码；该收发单元还用于向该接入网设备发送第三信息，该第三信息用于该接入网设备对该第三信息进行第一解码，以生成该CSI，该第一解码包括对应该第一编码的解码。
从而，在本申请中，接入网设备将经过编码后的CSI-RS通过N个射频链路发送给终端设备，终端设备对该经过编码后的CSI-RS无需进行解码操作，而是再进行一次编码发给接入网设备由接入网设备进行解码，以获得CSI，可以降低获取CSI过程中对资源的占用。
结合第四方面，在第四方面的某些实现方式中，该第三信息用于该接入网设备对经过第二解码的该第三信息进行该第一解码，以生成该CSI，该第二解码与该第二编码对应。
结合第四方面，在第四方面的某些实现方式中，该第一解码还包括对应该第二编码的解码。
结合第四方面，在第四方面的某些实现方式中，该第二编码为基于第一神经网络的编码，该第一神经网络的参数与该信道的采样点数量F和该第一信息占用的资源的数量M有关，该M和F为正整数，M<N，该资源包括以下至少一种：时域资源、频域资源或码域资源。
结合第四方面，在第四方面的某些实现方式中，该第一神经网络的参数为该第一神经网络的输入矩阵的维度。
结合第二方面，在第二方面的某些实现方式中，该第一编码为基于压缩感知方式的编码，该第一编码使用第一矩阵，该第一矩阵的维度与该M和该N有关。
结合第四方面，在第四方面的某些实现方式中，该第一编码为基于第二神经网络的编码，该第二神经网络包括全连接线性层，该全连接线性层的参数与该M和该N有关。
结合第四方面，在第四方面的某些实现方式中，该第一解码为基于第三神经网络的解码，该第三神经网络的参数与该N，该M和该F有关。
第五方面，提供了一种传输数据的方法，该方法包括：对第一信息进行降维处理，以生成第二信息；基于第一神经网络对该第二信息进行编码，该第二信息的维度与该第一神经网络能够处理的信息的维度对应。
本申请中，对信息先进行降维处理，获得维度较低的第二信息，然后基于第一神经网络对该第二信息进行编码，保证了编码的质量的同时大大降低编码的计算量。
结合第五方面，在第五方面的某些实现方式中，该第一神经网络是基于第一训练数据训练获得，该第一训练数据是经过降维处理的数据。
结合第五方面，在第五方面的某些实现方式中，该第一训练数据为第二信息。
该第一神经网络可以通过经过降维处理的数据进行训练，从而可以处理与第二信息的维度相对应的数据，从而提高编码的性能。
结合第五方面，在第五方面的某些实现方式中，该第一神经网络的训练方式为串联该降维处理和该第一神经网络统一训练。
在本申请中，可以将降维处理和第一神经网络进行级联统一进行训练，采用该种训练 方式，使神经网络对训练数据的依赖性降低，对训练集之外的数据也具有较佳的性能，从而获得一定的泛化能力。
结合第五方面，在第五方面的某些实现方式中对该第一信息进行降维处理，包括：基于压缩感知的方式对该第一信息进行降维处理，该压缩感知方式使用第一矩阵，该第一矩阵的尺寸与该第一神经网络能够处理的信息的维度对应。
压缩感知方式可以最终表现为一层矩阵，采用压缩感知的方式可以对信道信息达到降维的效果，从而，第一神经网络对已经降维的信道信息进行编码，可以降低总体编码的计算量。并且，结合压缩编码和神经网络编码的这种复合编码方式，相较于单纯依靠压缩编码的方式，信道信息的质量也可显著提高。
结合第五方面，在第五方面的某些实现方式中对该第一信息进行降维处理，对该第一信息进行降维处理，以生成第二信息，包括：基于该第一神经网络的第1至N层对该第一信息进行降维处理，该第一神经网络的第1至N层为第一全连接线性层，该N为整数，N≥1；基于该第一神经网络对该第二信息进行编码，包括：基于该第一神经网络的第N+1至M层对该第二信息进行编码，该第一全连接线性层的尺寸与该第一神经网络的第N+1至M层能够处理的信息的维度对应，该M为整数，M≥N+1。
在本申请中，对第一信息进行降维处理生成第二信息，与基于第一神经网络对第二信息进行编码，可以表现为一个神经网络的编码结构。该神经网络的第1至N层为全连接线性层，即，通过全连接线性层对信道信息先进行降维处理，然后再进入后面的层对降维后的信道信息进行编码。本申请与目前的基于神经网络的编码方式不同，目前的神经网络通常先对数据进行填充，在后面的层才会再进行降维，本申请则是先进行降维，大大减少了编码的计算量。
结合第五方面，在第五方面的某些实现方式中，对该第一信息进行降维处理，对该第一信息进行降维处理，以生成该第二信息，包括：基于第二神经网络对该第一信息进行降维处理，以生成第二信息，该第二神经网络包括第二全连接线性层，该第二全连接线性层与该第一神经网络能够处理的信息的维度对应。
全连接线性层在数学上可以表现为非稀疏矩阵相乘，因此，可以通过全连接线性层对信道信息进行降维处理。
结合第五方面，在第五方面的某些实现方式中，第一信息为信道状态信息。
对该第一信息进行降维处理，以生成第二信息，即对信道状态信息进行降维处理，以生成降维后的信道状态信息，基于第一神经网络对降维后的信道状态信息进行编码，以生成待发送经过编码的信道状态信息。
第六方面，提供了一种传输数据的装置，该装置包括：收发单元，用于接收该第一信息；处理单元，用于对第一信息进行降维处理，以生成第二信息；该处理单元还用于基于第一神经网络对该第二信息进行编码，该第二信息的维度与该第一神经网络能够处理的信息的维度对应。
该装置对信道信息先进行降维处理，获得维度较低的第二信息，然后基于第一神经网络对该第二信息进行编码，可以大大降低编码的计算量，同时也保证了编码的质量。
结合第六方面，在第六方面的某些实现方式中，该第一神经网络是基于第一训练数据训练获得，该第一训练数据是经过降维处理的数据。
结合第六方面，在第六方面的某些实现方式中，该第一训练数据为第二信息。
结合第六方面，在第六方面的某些实现方式中，该第一神经网络的训练方式为串联该降维处理和该第一神经网络统一训练。
结合第六方面，在第六方面的某些实现方式中，该处理单元具体用于：基于压缩感知的方式对该第一信息进行降维处理，该压缩感知方式使用第一矩阵，该第一矩阵的尺寸与该第一神经网络能够处理的信息的维度对应。
结合第六方面，在第六方面的某些实现方式中，该处理单元具体用于：基于该第一神经网络的第1至N层对该第一信息进行降维处理，该第一神经网络的第1至N层为第一全连接线性层，该N为整数，N≥1；基于该第一神经网络的第N+1至M层对该第二信息进行编码，该第一全连接线性层的尺寸与该第一神经网络的第N+1至M层能够处理的信息的维度对应，该M为整数，M≥N+1。
结合第六方面，在第六方面的某些实现方式中，该处理单元具体用于：基于第二神经网络对该第一信息进行降维处理，以生成第二信息，该第二神经网络包括第二全连接线性层，该第二全连接线性层与该第一神经网络能够处理的信息的维度对应。
结合第六方面，在第六方面的某些实现方式中，第一信息为信道状态信息。
第七方面，提供了一种传输数据的方法，该方法包括：接收第三信息；基于第三神经网络对该第三信息进行解码，以生成第四信息，该第三信息的维度与该第三神经网络能够处理的信息的维度对应；对第四信息进行恢复处理，该恢复处理与降维处理对应。
结合第七方面，在第七方面的某些实现方式中，该第三神经网络的训练方式为串联该降维处理，第一神经网络，第三神经网络和恢复处理统一训练。
结合第七方面，在第七方面的某些实现方式中，基于压缩感知的方式对该第四信息进行恢复处理。
结合第七方面，在第七方面的某些实现方式中，该第三信息为编码后的信道状态信息。
第八方面，提供了一种传输数据的装置，其特征在于，该装置包括收发单元，用于接收该第三信息；处理单元，用于基于第三神经网络对该第三信息进行解码，该第三信息的维度与该第一神经网络能够处理的信息的维度对应。
结合第八方面，在第八方面的某些实现方式中，该第三神经网络的训练方式为串联该降维处理，第一神经网络，第三神经网络和恢复处理统一训练。
结合第八方面，在第八方面的某些实现方式中，基于压缩感知的方式对该第四信息进行恢复处理。
结合第八方面，在第八方面的某些实现方式中，该第三信息为编码后的信道状态信息。
第九方面，提供了一种通信装置，该装置可以包括处理单元、发送单元和接收单元。可选的，发送单元和接收单元还可以为收发单元。
当该装置是接入网设备时，该处理单元可以是处理器，该发送单元和接收单元可以是收发器；该装置还可以包括存储单元，该存储单元可以是存储器；该存储单元用于存储指令，该处理单元执行该存储单元所存储的指令，以使该接入网设备执行第一方面、第二方面、第五方面或第七方面的任一方法。当该装置是接入网设备内的芯片时，该处理单元可以是处理器，该发送单元和接收单元可以是输入/输出接口、管脚或电路等；该处理单元执行存储单元所存储的指令，以使该芯片执行第一方面、第二方面、第五方面或第七方面 的方法。该存储单元用于存储指令，该存储单元可以是该芯片内的存储单元(例如，寄存器、缓存等)，也可以是该接入网设备内的位于该芯片外部的存储单元(例如，只读存储器、随机存取存储器等)。
当该装置是终端设备时，该处理单元可以是处理器，该发送单元和接收单元可以是收发器；该装置还可以包括存储单元，该存储单元可以是存储器；该存储单元用于存储指令，该处理单元执行该存储单元所存储的指令，以使该终端设备执行第一方面、第二方面、第五方面或第七方面的任一方法。当该装置是终端设备内的芯片时，该处理单元可以是处理器，该发送单元和接收单元可以是输入/输出接口、管脚或电路等；该处理单元执行存储单元所存储的指令，以使该芯片执行第一方面、第二方面、第五方面或第七方面的方法。该存储单元用于存储指令，该存储单元可以是该芯片内的存储单元(例如，寄存器、缓存等)，也可以是该终端设备内的位于该芯片外部的存储单元(例如，只读存储器、随机存取存储器等)。
第十方面，提供了一种通信装置，包括处理器和接口电路，接口电路用于接收来自该通信装置之外的其它通信装置的信号并传输至该处理器或将来自该处理器的信号发送给该通信装置之外的其它通信装置，该处理器通过逻辑电路或执行代码指令用于实现前述第一方面至第五方面的任意方法。
第十一方面，提供了一种计算机可读存储介质，该计算机可读存储介质中存储有计算机程序或指令，当该计算机程序或指令被执行时，实现前述第一方面、第二方面、第五方面或第七方面中的任意方法。
第十二方面，提供了一种包含指令的计算机程序产品，当该指令被运行时，实现前述第一方面、第二方面、第五方面或第七方面的任意方法。
第十三方面，提供了一种计算机程序，该计算机程序包括代码或指令，当该代码或指令被运行时，实现前述第一方面、第二方面、第五方面或第七方面的任意可能的实现方式中的方法。
第十四方面，提供一种芯片系统，该芯片系统包括处理器，还可以包括存储器，用于实现前述第一方面、第二方面、第五方面或第七方面描述的至少一种方法。该芯片系统可以由芯片构成，也可以包含芯片和其他分立器件。
第十五方面，提供一种通信系统，该系统包括第九方面至第十四方面任一该的装置(如接入网设备)。
第十六方面，提供一种通信系统，该系统包括第九方面至第十四方面任一该的装置(如终端设备)。
附图说明
图1是适用本申请传输数据方法的一例示意图；
图2是基于压缩感知的传输CSI的一例结构示意图；
图3是本申请实施例的传输CSI的一例示意性流程图；
图4是本申请实施例的传输CSI的一例结构示意图；
图5是本申请实施例的传输CSI的另一例结构示意图；
图6是本申请实施例的获取CSI的一例示意性流程图；
图7是获取CSI的一例结构示意图；
图8至图11是本申请实施例提供的可能的装置的结构示意图。
具体实施方式
下面将结合附图，对本申请中的技术方案进行描述。
本申请提供的技术方案可以应用于各种通信系统，例如：长期演进(Long Term Evolution，LTE)系统、LTE频分双工(frequency division duplex，FDD)系统、LTE时分双工(time division duplex，TDD)、通用移动通信系统(universal mobile telecommunication system，UMTS)、全球互联微波接入(worldwide interoperability for microwave access，WiMAX)通信系统、第五代(5th Generation，5G)移动通信系统或新无线接入技术(new radio access technology，NR)。其中，5G移动通信系统可以包括非独立组网(non-standalone，NSA)和/或独立组网(standalone，SA)。
本申请提供的技术方案还可以应用于机器类通信(machine type communication，MTC)、机器间通信长期演进技术(Long Term Evolution-machine，LTE-M)、设备到设备(devicetodevice，D2D)网络、机器到机器(machine to machine，M2M)网络、物联网(internet of things，IoT)网络或者其他网络。其中，IoT网络例如可以包括车联网。其中，车联网系统中的通信方式统称为车到其他设备(vehicle to X，V2X，X可以代表任何事物)，例如，该V2X可以包括：车辆到车辆(vehicle to vehicle，V2V)通信，车辆与基础设施(vehicle to infrastructure，V2I)通信、车辆与行人之间的通信(vehicle to pedestrian，V2P)或车辆与网络(vehicle to network，V2N)通信等。
本申请提供的技术方案还可以应用于未来的通信系统，如第六代移动通信系统等。本申请对此不作限定。
本申请实施例中，接入网设备可以是任意一种具有无线收发功能的设备。该设备包括但不限于：演进型节点B(evolved Node B，eNB)、无线网络控制器(radio network controller，RNC)、节点B(Node B，NB)、基站控制器(base station controller，BSC)、基站收发台(base transceiver station，BTS)、家庭基站(例如，home evolved NodeB，或home Node B，HNB)、基带单元(baseband unit，BBU)，无线保真(wireless fidelity，WiFi)系统中的接入点(access point，AP)、无线中继节点、无线回传节点、传输点(transmission point，TP)或者发送接收点(transmission and reception point，TRP)等，还可以为5G，如，NR，系统中的gNB，或，传输点(TRP或TP)，5G系统中的基站的一个或一组(包括多个天线面板)天线面板，或者，还可以为构成gNB或传输点的网络节点，如基带单元(BBU)，或，分布式单元(distributed unit，DU)等。
在一些部署中，gNB可以包括集中式单元(centralized unit，CU)和DU。gNB还可以包括有源天线单元(active antenna unit，AAU)。CU实现gNB的部分功能，DU实现gNB的部分功能，比如，CU负责处理非实时协议和服务，实现无线资源控制(radio resource control，RRC)，分组数据汇聚层协议(packet data convergence protocol，PDCP)层的功能。DU负责处理物理层协议和实时服务，实现无线链路控制(radio link control，RLC)层、介质接入控制(medium access control，MAC)层和物理(physical，PHY)层的功能。AAU实现部分物理层处理功能、射频处理及有源天线的相关功能。由于RRC层的信息最终会变成PHY层的信息，或者，由PHY层的信息转变而来，因而，在这种架构下，高层 信令，如RRC层信令，也可以认为是由DU发送的，或者，由DU+AAU发送的。可以理解的是，接入网设备可以为包括CU节点、DU节点、AAU节点中一项或多项的设备。此外，可以将CU划分为接入网(radio access network，RAN)中的接入网设备，也可以将CU划分为核心网(core network，CN)中的接入网设备，本申请对此不做限定。
接入网设备为小区提供服务，终端设备通过接入网设备分配的传输资源(例如，频域资源，或者说，频谱资源)与小区进行通信，该小区可以属于宏基站(例如，宏eNB或宏gNB等)，也可以属于小小区对应的基站，这里的小小区可以包括：城市小区、微小区、微微小区、毫微微小区等，这些小小区具有覆盖范围小、发射功率低的特点，适用于提供高速率的数据传输服务。
在本申请实施例中，终端设备也可以称为用户设备(user equipment，UE)、接入终端、用户单元、用户站、移动站、移动台、远方站、远程终端、移动设备、用户终端、终端、无线通信设备、用户代理或用户装置。
终端设备可以是一种向用户提供语音/数据连通性的设备，例如，具有无线连接功能的手持式设备、车载设备等。目前，一些终端的举例可以为：手机、平板电脑、带无线收发功能的电脑(如笔记本电脑、掌上电脑等)、移动互联网设备(mobile internet device，MID)、虚拟现实(virtual reality，VR)设备、增强现实(augmented reality，AR)设备、工业控制中的无线终端、无人驾驶中的无线终端、远程医疗中的无线终端、智能电网中的无线终端、运输安全中的无线终端、智慧城市中的无线终端、智慧家庭中的无线终端、蜂窝电话、无绳电话、会话启动协议(session initiation protocol，SIP)电话、无线本地环路(wireless local loop，WLL)站、个人数字助理(personal digital assistant，PDA)、具有无线通信功能的手持设备、计算设备或连接到无线调制解调器的其它处理设备、车载设备、可穿戴设备，5G网络中的终端设备或者未来演进的公用陆地移动通信网络(public land mobile network，PLMN)中的终端设备等。
其中，可穿戴设备也可以称为穿戴式智能设备，是应用穿戴式技术对日常穿戴进行智能化设计、开发出可以穿戴的设备的总称，如眼镜、手套、手表、服饰及鞋等。可穿戴设备即直接穿在身上，或是整合到用户的衣服或配件的一种便携式设备。可穿戴设备不仅仅是一种硬件设备，更是通过软件支持以及数据交互、云端交互来实现强大的功能。广义穿戴式智能设备包括功能全、尺寸大、可不依赖智能手机实现完整或者部分的功能，例如：智能手表或智能眼镜等，以及只专注于某一类应用功能，需要和其它设备如智能手机配合使用，如各类进行体征监测的智能手环、智能首饰等。
此外，终端设备还可以是物联网(internet of things，IoT)系统中的终端设备。IoT是未来信息技术发展的重要组成部分，其主要技术特点是将物品通过通信技术与网络连接，从而实现人机互连，物物互连的智能化网络。IoT技术可以通过例如窄带(narrow band，NB)技术，做到海量连接，深度覆盖，终端省电。
此外，终端设备还可以包括智能打印机、火车探测器、加油站等传感器，主要功能包括收集数据(部分终端设备)、接收接入网设备的控制信息与下行数据，并发送电磁波，向接入网设备传输上行数据。
为便于理解本申请实施例，首先结合图1详细说明适用于本申请实施例提供的信道测量方法的通信系统。图1示出了适用于本申请实施例提供的方法的通信系统100的示意图。 如图所示，该通信系统100可以包括至少一个接入网设备，如图1中所示的接入网设备101；该通信系统100还可以包括至少一个终端设备，如图1中所示的终端设备102至107。其中，该终端设备102至107可以是移动的或固定的。接入网设备101和终端设备102至107中的一个或多个均可以通过无线链路通信。每个接入网设备可以为特定的地理区域提供通信覆盖，并且可以与位于该覆盖区域内的终端设备通信。例如，接入网设备可以向终端设备发送配置信息，终端设备可以基于该配置信息向接入网设备发送上行数据；又例如，接入网设备可以向终端设备发送下行数据。因此，图1中的接入网设备101和终端设备102至107构成一个通信系统。
可选地，终端设备之间可以直接通信。例如可以利用D2D技术等实现终端设备之间的直接通信。如图中所示，终端设备105与106之间、终端设备105与107之间，可以利用D2D技术直接通信。终端设备106和终端设备107可以单独或同时与终端设备105通信。
终端设备105至107也可以分别与接入网设备101通信。例如可以直接与接入网设备101通信，如图1中的终端设备105和106可以直接与接入网设备101通信；也可以间接地与接入网设备101通信，如图1中的终端设备107经由终端设备105与接入网设备101通信。
应理解，图1示例性地示出了一个接入网设备和多个终端设备，以及各通信设备之间的通信链路。可选地，该通信系统100可以包括多个接入网设备，并且每个接入网设备的覆盖范围内可以包括其它数量的终端设备，例如更多或更少的终端设备。本申请对此不做限定。
上述各个通信设备，如图1中的接入网设备101和终端设备102至107，可以配置多个天线。该多个天线可以包括至少一个用于发送信号的发射天线和至少一个用于接收信号的接收天线。另外，各通信设备还附加地包括发射机链和接收机链，本领域普通技术人员可以理解，它们均可包括与信号发送和接收相关的多个部件(例如处理器、调制器、复用器、解调器、解复用器或天线等)。因此，接入网设备与终端设备之间可通过多天线技术通信。
可选地，该无线通信系统100还可以包括网络控制器、移动管理实体等其他网络实体，本申请实施例不限于此。
为便于理解本申请实施例，首先对下文涉及到的几个术语做简单说明。
1、天线端口：简称端口。一个天线端口可以是一个物理天线，也可以是多个物理天线的加权组合。
在本申请实施例中，天线端口可包括发射天线端口和接收天线端口。
发射天线端口可以理解为被接收端所识别的发射天线，或者，在空间上可以区分的发射天线。同一发射天线端口传输的信号所经历的信道环境相同。接收端可以据此进行信道估计从而对传输信号进行解调。在本申请实施例中，发射天线端口可以是独立的收发单元，也可以是参考信号端口。一个参考信号端口可与一个参考信号对应。参考信号端口例如可以包括但不限于，信道状态信息参考信号(channel state information reference signal，CSI-RS)端口、解调参考信号(demodulation reference signal，DMRS)端口等等。本申请对此不作限定。
接收天线端口可以理解为被接收端所识别的接收天线，或者，在空间上可以区分的接收天线。接收天线端口和发射天线端口例如可用于后续确定信道矩阵。本申请对此也不作限定。
2、空域向量：也可以称为角度向量、波束向量等。空域向量中的各个元素可用于表示各个发射天线端口的权重。基于空域向量中各个元素所表示的各个发射天线端口的权重，将各个发射天线端口的发射能量做线性叠加，可以在空间某一方向上形成能量较强的区域。
空域向量可以是长度为T的向量。其中，T可以表示发射天线端口数，T≥1且为整数。空域向量例如可以是长度为T的列向量或行向量。本申请对此不作限定。在下文中为方便理解和说明，假设该接收空域向量为长度为T的列向量。
对于一个长度为T的空域向量来说，它包含了T个空域权值(或简称，权值)，该T个权值可用于对T个发射天线端口进行加权，以使得该T个发射天线端口所发射出来的参考信号具有一定的空间指向性，从而实现波束赋形。因此，基于不同的空域向量对信号进行预编码，就相当于基于不同的空域向量对发射天线端口进行波束赋形，以使得所发射出来的信号具有不同的空间指向性。
可选地，空域向量是离散傅里叶变换(discrete fourier transform，DFT)向量。DFT向量可以是指DFT矩阵中的列向量。
可选地，空域向量是DFT向量的共轭转置向量。DFT共轭转置向量可以是指DFT矩阵的共轭转置矩阵中的列向量。
可选地，空域向量是过采样DFT向量。过采样DFT向量可以是指过采样DFT矩阵中的向量。
可选地，空域向量是过采样DFT向量的共轭转置向量。
在一种可能的设计中，该空域向量例如可以是第三代合作伙伴(3rd Generation Partnership Project，3GPP)技术规范(technical specification，TS)38.214版本15(release15，R15)或R16中类型II(type II)码本中定义的二维(2dimensions，2D)-DFT向量v l,m。也就是说，空域向量可以是2D-DFT向量或过采样2D-DFT向量。
应理解，上文对空域向量的具体形式的举例仅为示例，不应对本申请构成任何限定。
3、频域向量：也可以称为时延向量。频域向量可用于表示信道在频域的变化规律的向量。每个频域向量可以表示一种变化规律。由于信号在经过无线信道传输时，从发射天线可以经过多个路径到达接收天线。多径时延导致频率选择性衰落，就是频域信道的变化。因此，可以通过不同的频域向量来表示不同传输路径上时延导致的信道在频域上的变化规律。
频域向量可以是长度N的向量。其中，N表示频域单元数。N≥1，且为整数。频域向量例如可以是长度为N的列向量或行向量。本申请对此不作限定。在下文中为方便理解和说明，假设该频域向量为长度为N的列向量。
可选地，频域向量是DFT向量或DFT向量的共轭转置向量。
可选地，频域向量是过采样DFT向量或过采样DFT向量的共轭转置向量。
应理解，上文对频域向量的具体形式的举例仅为示例，不应对本申请构成任何限定。
本领域的技术人员可以理解，DFT向量是DFT矩阵中的向量。而DFT矩阵是正交基向量集合，故DFT矩阵中的任意两个向量之间是可以两两相互正交的。
过采样DFT矩阵可以是对DFT矩阵进行过采样得到的。过采样DFT矩阵中的向量可以分为多个子集，每个子集中的DFT相邻之间也是可以两两相互正交的，不同子集之间可以是非正交的。
4、频域单元：可用于表示不同的频域资源粒度。频域单元例如可以包括但不限于，子带、资源块(resource block，RB)、子载波、资源块组(resource block group，RBG)或预编码资源块组(precoding resource block group，PRG)等。
与频域单元对应的信道可用于确定该频域单元对应的预编码，以用于后续数据传输。对于终端设备来说，与频域单元对应的信道可以通过对该频域单元上的参考信号进行测量得到。对于网络设备来说，与频域单元对应的信道可以基于终端设备反馈的该频域单元对应的信道信息确定，也可以基于终端设备反馈的该频域单元附近的频域单元对应的信道信息确定。本申请对此不作限定。
在本申请实施例中，与频域单元对应的信道可简称为，该频域单元的信道。
5、压缩感知：也被称为压缩采样或稀疏采样，是一种寻找欠定线性系统的稀疏解的技术，用于获取和重构稀疏或可压缩的信号。对于信道信息反馈应用，在压缩感知的编码端，首先通过傅里叶变换(对应压缩感知中的字典矩阵)把信道信息转换到稀疏化的时延域信道信息，之后通过观测矩阵相乘把稀疏化的时延域信道信息转换成低维的表示，从而实现从高维到低维的压缩(或编码)。
在压缩感知的解码端，传统上通常采用迭代的方法恢复信道信息，常用的解码方法包括正交匹配追踪(orthogonal matching pursuit，OMP)、近似消息传递(approximate message passing，AMP)等。值得关注的是，近年来，随着数据驱动的人工智能技术的发展，一种新的压缩感知解码方案是把传统的迭代算法展开，构成一种传统方法指导的解码网络，并利用数据训练获得该解码网络中的参数，例如一种典型算法为学习的近似消息传递(LAMP)。
6、神经网络：一种模仿动物神经网络行为特征，进行分布式并行信息处理的算法数学模型。神经网络可以是由神经单元组成的，神经单元可以是指以x s和截距1为输入的运算单元，该运算单元的输出可以为：
Figure PCTCN2021135367-appb-000001
其中，s＝1、2、……n，n为大于1的自然数，W s为x s的权重，b为神经单元的偏置。f为神经单元的激活函数，用于将非线性特性引入神经网络中，来将神经单元中的输入信号转换为输出信号。该激活函数的输出信号可以作为下一层卷积层的输入，激活函数可以是sigmoid、Tanh、Relu或LeakyRelu函数等。神经网络是将多个上述单一的神经单元联结在一起形成的网络，即一个神经单元的输出可以是另一个神经单元的输入。每个神经单元的输入可以与前一层的局部接受域相连，来提取局部接受域的特征，局部接受域可以是由若干个神经单元组成的区域。
作为示例而非限定，在本实施例中，参考信号可以用于信道测量，进而用于解调，例如，本发明实施例中的参考信号可以包括解调参考信号(demodulation reference signal，DMRS)。
应理解，参考信号作为一种用于信道测量或信道估计的参考信号，仅为示例性说明，不应对本发明实施例构成任何限定，本申请并不排除在现有或未来的协议中采用其他的名 称代替参考信号以实现其相同或相似功能的可能。
目前，接入网设备可以通过向终端设备发送CSI-RS获取CSI，图2示出了一例接入网设备发送CSI-RS获取CSI的结构示意图。
随着天线数量和宽带的增加，CSI同步增加。因此，在CSI反馈时必须进行压缩和解压缩，也可以叫做编码和解码，从而可以节省宝贵的空口传输资源。目前有三种常见的CSI编码和解码技术：基于码本的CSI反馈、基于压缩感知的编码和解码、基于人工智能的编码和解码。
在基于码本进行CSI反馈的方案中，接收端估计获得CSI矩阵后，对CSI矩阵进行傅里叶变换，即将CSI矩阵转化为稀疏化的时延信息，然后针对稀疏化的时延信息，选择部分强波束、量化，并通过预先配置的码本，选择合适的指示符反馈给发送方。基于码本进行CSI反馈的方案通过波束选择、量化实现CSI的编码，对次要信息直接丢弃，导致大量有用信息的损失，CSI的恢复精度不佳，且对信道容量造成消极影响。
参加图2，图2示出了一例基于压缩感知的反馈CSI的结构示意图，基于压缩感知的反馈CSI的方案中，在压缩感知的编码端，首先经过压缩感知编码模块，例如先与字典矩阵相乘，将信道信息转换为稀疏化的时延信道信息，之后通过与观测矩阵相乘把稀疏化的时延信息转换为低维的表示，从而实现编码。
基于压缩感知的解码的方案中，采用压缩感知解码模块，例如通常采用迭代的方法恢复信道信息，常用的解码方法包括正交匹配追踪(orthogonal matching pursuit，OMP)、近似消息传递(approximate message passing，AMP)等。需要说明的是，随着数据驱动的人工智能技术的发展，一种新的压缩感知解码方案是把传统的迭代算法展开，构成一种传统方法指导的解码网络，并利用数据训练获得该解码网络中的参数，例如学习的近似消息传递(LAMP)算法。
基于压缩感知的反馈CSI方案，针对可编码的或在某个变换域是稀疏的信号，可以将高维度信号映射到低维空间中，从而以远低于奈奎斯特采样频率对信号采样，与基于码本的方案相比，具有较高质量的恢复信号。该方案可以以较低的计算量获得较好的编解码性能，且对各种信道环境都有较好的适应能力，即，具有较好的泛化能力。该方案在编码端通过傅里叶变换、矩阵相乘等线性变换把信道信息映射到低维表示，但是，线性变换对信息特征的提取能力有限，在高压缩比下，该方案编解码的能力表现欠佳，恢复出的信道信息失真较大。
基于人工智能的反馈CSI的方案中，以深度学习为代表的、数据驱动的人工智能技术具有强大的特征提取能力，在信道信息编解码上可以获得较优的结果，例如，编码端利用一层3x3卷积神经网络与一层全连接线性层网络的CsiNet模块，解码端利用多个RefineNet模块叠加，就可超过目前的基于压缩感知的信道信息编解码方案的性能。但是，基于人工智能的编解码方案的计算量很大，且需要大量训练数据，训练出来的神经网络通常更多体现训练数据的特征，对于与训练数据特征偏差较大的数据表现性能不佳。
图3示出了一例反馈CSI的方法100的示意性流程图。
S101，对CSI进行降维处理，以生成降维后的CSI。
例如，基于压缩感知的方式对CSI进行降维处理，例如，将CSI与字典矩阵相乘，字典矩阵用于实现CSI的稀疏化，字典矩阵可以是DFT或FFT变换矩阵，也可以是其它特 别设计的矩阵，本申请不作限定。
CSI与字典矩阵相乘后可以与观测矩阵相乘，观测矩阵用于实现CSI的降维，从而获得降维后的CSI。观测矩阵可以是高斯随机抽样矩阵，也可以是使用深度学习中梯度下降的方法训练数据获得的矩阵，也可以是其它特别设计的矩阵，本申请不作限定。
其中，字典矩阵和观测矩阵为非稀疏矩阵，即，矩阵中非零的个数大于50％。
例如，采用神经网络的方式对CSI进行降维处理，该神经网络包括全连接线性层，从而生成降维后的CSI。
需要说明的是，全连接线性层在数学上表现为矩阵相乘，全连接意味着矩阵为非稀疏矩阵，因此采用神经网络的方式对CSI进行降维处理，可以理解为将CSI与至少一个矩阵进行线性相乘。例如，神经网络的第一层为DFT或快速傅里叶变换(fast Fourier transform，FFT)计算矩阵，第二层为高斯随机抽样矩阵。
多个矩阵也可以进行合并，从而形成一个矩阵相乘，多个全连接线性层串联也可以等效为一个全连接线性层。
可选地，S102，对CSI进行降维处理中包括截断操作。
即，在与至少一个矩阵进行线性相乘的过程中，在与其中一个矩阵相乘后，可以进行截断操作后再与下一个矩阵进行线性相乘。
例如，CSI与字典矩阵相乘后，进行截断后与观测矩阵相乘。需要说明的是，与字典矩阵进行线性相乘后的信息具有高度稀疏性，从而可以采用直接截断的方式将信息的维度进一步降低。
例如，CSI输入神经网络的其中一层全连接线性层，进行截断后再进入下一层。
S103，采用神经网络的方式对降维后的CSI进行再一次编码。
可选地，神经网络可以是基于串联以上降维处理的方式和该神经网络统一训练获得，也可以理解为级联以上降维处理的方式和该神经网络进行统一训练。
需要说明的是，降维后的CSI的维度与神经网络输入参数有关，需要以上降维处理中使用的矩阵的参数与该神经网络输入参数的维度对应。
需要说明的是，采用神经网络对降维后的CSI进行编码，该神经网络可以包括非线性激活函数，该激活函数可以是Sigmoid、Tanh、Relu或LeakyRelu等，该神经网络还可以包括卷积、全连接、池化或batch normalization，本申请不作限定。
需要说明的是，在采用神经网络的全连接线性层对CSI进行降维处理，在采用神经网络对降维后的CSI进行编码的方案中，本申请的编码结构可以表现为一个整体的可训练的神经网络，该神经网络的第1至N层为全连接线性操作层，第N+1层至P层可以包含非线性激活函数。
针对本申请反馈CSI的方法100的解码侧，可以由两个级联的结构组成，顺序为面向神经网络的解码结构和面向降维处理的解码结构。该神经网络可以包括非线性激活函数，该激活函数可以是Sigmoid、Tanh、Relu或LeakyRelu等，该神经网络还可以包括卷积、全连接、池化或batch normalization，本申请不作限定。
当采用压缩感知方式对CSI进行降维处理时，面向降维处理的解码结构可以是传统的OMP算法、AMP算法或是数据驱动的压缩感知解码算法，即由迭代算法展开形成的解码结构，例如学习的近似消息传递LAMP算法，该结构可通过数据训练获得该机构中的参 数。此时的编解码结构可参见图4，CSI顺序经过压缩感知编码模块和人工智能编码模块，两个模块分别实现部分压缩比。解码端顺序经过人工智能解码模块和压缩感知解码模块，首先通过人工智能的方法恢复部分数据编码，然后再通过压缩感知的方法恢复数据。
当采用包括全连接线性层的神经网络对CSI进行降维处理时，面向降维处理的解码结构也可以是基于神经网络的解码结构，即，此时的解码结构也可以看做为一个整体可训练的神经网络。
需要说明的是，以上神经网络都是通过训练数据训练获得，在一种可能的实现方式中，可以将降维处理的结构，神经网络，面向神经网络的解码结构和面向降维处理的解码结构串联起来统一进行端对端的训练。
为了详细说明，图5示出了一例具体的反馈CSI的结构示意图，该结构示意图仅做示例，并不对本申请构成任何限定。
以下以接入网设备32根天线，1024个载波为例，终端设备1根接收天线为例，终端设备估计出的时域CSI为1024×32×2的三维矩阵为例，对图4所示的反馈CSI结构的工作过程进行举例说明，以下说明对本申请不够成任何限定。
时域CSI的1024×32×2的三维矩阵经过压缩感知编码模块，其中，压缩感知编码模块包括DFT模块，观测矩阵模块，可选地，在DFT模块和观测模块之间还可以包括矩阵截断模块。
时域CSI进入DFT模块进行DFT转换，变为具有稀疏性的三维矩阵，即，此时维度不变，仍然为1024×32×2。
可选地，由于经过DFT转换的三维矩阵具有稀疏性，可以采用截断的方式将该三维矩阵截断为32×32×2。
对截断后的CSI矩阵维度重整为，2048×1的向量，然后与观测矩阵相乘，即，得到了压缩感知压缩之后的CSI编码向量，即本申请中第二信息的一例。观测矩阵可以通过高斯随机采样获得，以观测矩阵尺寸为2048×512为例，则压缩感知编码之后的CSI编码向量尺寸为512×1。
压缩感知编码之后的CSI编码向量进入AI编码模块，此处以“(全连接+LeakyRELU)×2+全连接”结构为例，实现CSI的第二级编码，形成反馈向量。
在解码端，首先为面向AI编码模块的解码模块，此处以“(全连接+LeakyRELU)×5+全连接”结构为例，实现CSI的第一级解码，得到和解码端对等的512×1的向量。然后，进入面向压缩编码感知模块的解码模块，即进入AMP算法模块，得到2048×1的向量，然后经过维度重整为32×32×2矩阵，再经过对应编码端截断操作的补零操作，得到1024×32×2的矩阵，最后经过IDFT得到恢复的时域CSI矩阵。
根据本实施例可知，本申请所要求保护的方案中，终端设备向接入网设备反馈CSI时可以节省极为宝贵的空口资源，并降低了整体的计算量。
在上述实施例中，对CSI进行编码后反馈给接入网设备可以节省资源，本申请提出一种对CSI-RS进行编码的实现方式，可以在节省资源的同时减少终端设备的计算功耗。
图6示出了一例在接入网设备侧对CSI-RS进行编码的方法200的示意性流程图。
S201，接入网设备对CSI-RS进行第一编码，以生成第一信息。
该CSI-RS用于获取接入网设备和终端设备之间信道对应的CSI。
在一种可能的实现方式中，接入网设备采用压缩感知的方式对CSI-RS进行第一编码。
例如，针对N个射频链路(radio frequency chain，RFC)，接入网设备仅预留M组资源，其中，N和M都为正整数，M可以远远小于N，经过压缩感知方式编码的CSI-RS，可以理解为通过压缩感知的方式将CSI-RS映射到射频链路上，即，可以通过压缩感知的感知矩阵A(例如，矩阵的维度为M×N)将M个CSI-RS映射到N个射频链路上。
也就是说，如果接入网设备没有对CSI-RS进行编码，接入网设备通过N个射频链路需要使用N个资源传输CSI-RS，终端设备是对一个N×F×2的CSI-RS矩阵进行观测获得CSI矩阵，而在本申请中，可以通过一个维度为M×N(M<N)的感知矩阵A对CSI-RS进行编码，则相当于CSI-RS矩阵前乘了一个感知矩阵A，终端设备是对一个M×F×2的CSI-RS矩阵进行观测，接入网设备通过N个射频链路仅需要使用M个资源传输CSI-RS，其中F为信道的采样点数量，2则表示CSI矩阵元素的实部和虚部，以上描述仅作示例，本申请通过采用压缩感知对CSI-RS编码的方式降低传输CSI-RS时对资源的占用，参数F也可以是其它参数的组合，本申请不作任何限定。
其中，感知矩阵A可以是高斯随机抽样矩阵，也可以是使用深度学习中梯度下降的方法训练数据获得的矩阵，也可以是其它特别设计的矩阵，本申请不作限定。
需要说明的是，感知矩阵A也可以表现为多个矩阵相乘的形式，本申请不作限定。
在另一种可能的实现方式中，接入网设备采用神经网络的方式对CSI-RS进行第一编码，该神经网络包括全连接线性层，从而生成第一信息。
需要说明的是，全连接线性层在数学上表现为矩阵相乘，全连接意味着矩阵为非稀疏矩阵，因此采用神经网络的方式对CSI-RS进行第一编码，可以理解为将CSI-RS与至少一个矩阵进行线性相乘。例如，一些实施例中，神经网络的第一层为DFT或快速傅里叶变换计算矩阵，第二层为高斯随机抽样矩阵，即将CSI-RS与两个矩阵进行线性相乘。
多个矩阵也可以进行合并，从而形成一个相乘得到的矩阵，多个全连接线性层串联也可以等效为一个全连接线性层。
S202，接入网设备通过N个射频链路向终端设备发送第一信息。
对应地，终端设备接收接入网设备通过N个射频链路向该终端设备发送的第一信息。
需要说明的是，第一信息经过N个射频链路传输后，终端设备所接收到的信息已经隐含地包括了CSI。
也就是说，终端设备接收到是经过N个射频链路使用M个资源发送的经过第一编码的CSI-RS，观测到的是一个感知矩阵和CSI矩阵相乘的矩阵Y＝A×H，即CSI矩阵H隐含的包括了在Y矩阵中。
S203，终端设备基于第一信息和CSI-RS确定第二信息。
在一种可能的实现方式中，接入网设备通过N个射频链路使用M个资源向终端设备发送经过第一编码的CSI-RS后，终端设备对承载经过第一编码的CSI-RS进行观测，从而确定隐含CSI的第二信息。
例如，终端设备接收到经过N个射频链路使用M个资源传输的第一信息后，可以根据该第一信息和CSI-RS生成第二信息，其中该CSI-RS可以是预先配置于终端设备中，即，终端设备接收到经过信道传输的压缩后的CSI-RS和已知的CSI-RS可以确定隐含CSI的第二信息。
需要说明的是，终端设备向接入网设备反馈CSI，可以理解为向接入网设备反馈对第一信息的观测结果，即对经过N个射频链路使用M个资源的CSI-RS的观测结果，而在接入网设备侧通过解码获取CSI。
S204，终端设备对第二信息进行第二编码，以生成第三信息。
可选地，该第二编码可以为基于第一神经网络的编码，即，终端设备采用神经网络的方式即第一神经网络对第二信息进行编码，以生成第三信息。
该第一神经网络的参数与信道的采样点数量F和第一信息占用的资源的数量M有关，其中，F和M均为正整数，该资源包括以下资源中的至少一种：时域资源、频域资源或码域资源。
例如，第二信息可以理解为包含CSI矩阵H的矩阵Y，若采用维度为M×N的感知矩阵A进行第一编码，则Y＝A×H。
其中，CSI矩阵H的维度可以为N×F×2，N为接入网设备的射频链路数量，F为信道的采样点数量，M为经过压缩的第一信息所占用的资源的数量，2则表示CSI矩阵元素的实部和虚部，因此，可知矩阵Y的维度为M×F×2，也就是说第一神经网络的输入矩阵的维度为M×F×2，该第一神经网络的参数与信道的采样点数量F和第一信息占用的资源的数量M有关。
需要说明的是，当第一编码为基于包括全连接线性层的神经网络的编码时，全连接线性层也可以表现为矩阵的形式，所以第一神经网络的输入矩阵的维度与上述类似，在此不再赘述。
S205，终端设备向接入网设备发送第三信息，对应地，接入网设备从该终端设备接收第三信息。
在一种可能的实现方式中，若终端设备对第二信息进行第二编码以生成第三信息，S206，接入网设备对第三信息进行第二解码。
该第二解码对应于终端设备对第二信息的第二编码。
例如，终端设备将维度为M×F×2的矩阵Y进行第二编码，第二解码可以将第三信息恢复成矩阵Y。
例如，该第二解码可以为基于神经网络的解码方式，该神经网络可以至少包含一层非线性激活函数操作，例如，sigmoid、Tanh、Relu或LeakyRelu函数等，该神经网络还可包括其它操作，例如，卷积、全连接、池化、batch normalization等。
S207，终端设备对第三信息进行第一解码。
在一种可能的实现方式中，接入网设备对CSI-RS进行第一编码以生成第一信息，终端设备根据第一信息和CSI-RS确定第二信息，终端设备向接入网设备发送第三信息，其中第二信息和第三信息相同，则该第一解码为对应第一编码的解码，例如，对应压缩感知方式编码的解码，或者对应包括全连接线性层的第二神经网络方式编码的解码。
其中，对应压缩感知方式编码的解码，可以是传统的OMP或AMP算法，也可以是由数据驱动的压缩感知解码算法，即由传统的迭代算法展开形成解码结构，该解码结构中的参数可以通过数据训练的方式获得，例如LAMP算法。
对应第二神经网络方式编码的解码，可以是基于神经网络的解码方式，该神经网络可以至少包含一层非线性激活函数操作，例如，sigmoid、Tanh、Relu或LeakyRelu函数等， 该神经网络还可包括其它操作，例如，卷积、全连接、池化、batch normalization等。
在另一种可能的实现方式中，接入网设备对CSI-RS进行第一编码以生成第一信息，终端设备根据第一信息和CSI-RS确定第二信息，终端设备对第二信息进行第二编码以生成第三信息，接入网设备对第三信息进行第二解码，则该第一解码也仅对应第一编码的解码，且接入网设备对第三信息进行第一解码，可以理解为对经过第二解码的第三信息进行第一解码，第一解码的描述与上述描述类似，在此不再赘述。
在另一种可能的实现方式中，接入网设备对CSI-RS进行第一编码以生成第一信息，终端设备根据第一信息和CSI-RS确定第二信息，终端设备对第二信息进行第二编码以生成第三信息，而接入网设备并未对第三信息进行对应第二编码的解码，第一解码可以理解为对应第一编码的解码和对应第二编码的解码的深度融合，表现为基于深度神经网络的解码结构，而不是两种解码结构级联的形式。
需要说明的是，因为神经网络可以用于对基于压缩感知的方式编码的数据进行解码，所以若第一编码为基于压缩感知的方式的编码，第一解码也可以呈现为融合式的解码结构。
从而，在本申请中，接入网设备将经过编码后的CSI-RS通过N个射频链路发送给终端设备，终端设备对该经过编码后的CSI-RS无需进行解码操作，而是再进行一次编码发给接入网设备由接入网设备进行解码，以获得CSI，可以降低获取CSI过程中对资源的占用。
为了详细说明，图7示出了一例获得CSI的结构示意图。接入网设备侧的第一编码模块对CSI-RS进行编码，并使用M个资源通过N个射频链路RFC将CSI-RS发送给终端设备，终端设备对通过N个射频链路发送的信息进行信道估计，从而确定Y矩阵，Y矩阵隐含的包括CSI矩阵H。终端设备还包括第二编码模块，用于对Y矩阵进行第二编码，并将经过第二编码的Y矩阵发送给接入网设备。在一种可能的方式中，接入网设备包括第一解码模块，该第一解码模块用以进行针对第一编码模块的解码，或者，该第一解码模块用以进行针对第一编码模块和第二编码模块的解码，在另一种可能的方式中，接入网设备包括第一解码模块和第二解码模块，第一解码模块用于进行针对第一编码模块的解码，第二解码模块用于进行针对第二编码模块的解码，经过第二编码的Y矩阵先进入第二解码模块，以恢复Y矩阵，再进入第一解码模块，以获得CSI矩阵H。
另外，需要说明的是，本申请中的N个、M个等也都可以理解为N组、M组，本申请不作限定。
图8是本申请实施例提供的传输数据的装置的示意性框图。如图8所示，该装置1000可以包括处理单元1100和收发单元1200。
可选地，该装置1000可对应于上文方法实施例中的编码装置，例如，可以为编码器，或者配置于编码器中的部件(如电路、芯片或芯片系统等)。
应理解，该装置1000可对应于根据本申请实施例的方法中的编码设备，该装置1000可以包括用于执行图3中的编码设备执行的方法的单元。并且，该装置1000中的各单元和上述其他操作和/或功能分别为了实现图3中的相应流程。
其中，当该装置1000用于执行图3中的方法时，处理单元1100可用于执行图3中的步骤对CSI进行降维处理，还可用于执行图3中的基于神经网络对降维后的CSI进行编码，收发单元1200可用于接收CSI。应理解，各单元执行上述相应步骤的具体过程在上述方 法实施例中已经详细说明，为了简洁，在此不再赘述。
应理解，该通信装置1000可对应于根据本申请实施例的终端设备，该通信装置1000可以包括用于执行图6中的方法中终端设备执行的方法的单元。并且，该通信装置1000中的各单元和上述其他操作和/或功能分别为了实现图6中的相应流程。
其中，当该通信装置1000用于执行图3中的方法400时，处理单元1100可用于执行图6中的步骤基于第一信息和CSI-RS确定第二信息，收发单元1200可用于执行图6中的步骤接收接入网设备通过N个射频链路发送的第一信息，该第一信息由对信道状态信息参考信号CSI-RS进行第一编码生成，该CSI-RS用于获取信道信息CSI。处理单元1100还可用于执行图6中的对第二信息进行第二编码，以生成第三信息。收发单元1200还可用于执行图6中的向该接入网设备发送第三信息，该第三信息用于该接入网设备对该第三信息进行第一解码，以生成该CSI，该第一解码包括对应该第一编码的解码。应理解，各单元执行上述相应步骤的具体过程在上述方法实施例中已经详细说明，为了简洁，在此不再赘述。
还应理解，该装置1000为终端设备时，该装置1000中的收发单元1200可以通过收发器实现，例如可对应于图9中示出的装置2000中的收发器2020或图10中示出的终端设备3000中的收发器3020，该装置1000中的处理单元1100可通过至少一个处理器实现，例如可对应于图9中示出的装置2000中的处理器2010或图10中示出的终端设备3000中的处理器3010。
还应理解，该装置1000为配置于终端设备中的芯片或芯片系统时，该装置1000中的收发单元1200可以通过输入/输出接口、电路等实现，该装置1000中的处理单元1100可以通过该芯片或芯片系统上集成的处理器、微处理器或集成电路等实现。
可选地，该装置1000可对应于上文方法实施例中的解码设备，例如，可以为解码器，或者配置于解码器中的部件(如电路、芯片或芯片系统等)。
应理解，该装置1000可对应于根据本申请实施例中的解码设备，该装置1000可以包括用于执行解码设备执行的方法的单元。并且，该装置1000中的各单元和上述其他操作和/或功能分别为了实现方法400的相应操作。
其中，当该装置1000用于执行图3中的方法时，处理单元1100可用于执行图3中的方法的基于神经网络对编码后的CSI进行解码，收发单元1200可用于执行图3中的方法的接收编码后的CSI。应理解，各单元执行上述相应步骤的具体过程在上述方法实施例中已经详细说明，为了简洁，在此不再赘述。
还应理解，该装置1000为解码设备时，该装置1000中的收发单元1200可以通过收发器实现，例如可对应于图9中示出的装置2000中的收发器2020或图11中示出的接入网设备4000中的RRU 4100，该装置1000中的处理单元1100可通过至少一个处理器实现，例如可对应于图9中示出的装置2000中的处理器2010或图11中示出的基站4000中的处理单元4200或处理器4202。
还应理解，该装置1000为配置于网络设备中的芯片或芯片系统时，该装置1000中的收发单元1200可以通过输入/输出接口、电路等实现，该装置1000中的处理单元1100可以通过该芯片或芯片系统上集成的处理器、微处理器或集成电路等实现。
应理解，该通信装置1000可对应于根据本申请实施例的图6中的接入网设备，该通 信装置1000可以包括用于执行图6中的方法中接入网设备执行的方法的单元。并且，该通信装置1000中的各单元和上述其他操作和/或功能分别为了实现图6中的方法的相应流程。
其中，当该通信装置1000用于执行图6中的方法时，处理单元1100可用于执行图6中的步骤对信道状态信息参考信号CSI-RS进行第一编码，以生成第一信息，该CSI-RS用于获取接入网设备和该终端设备之间信道对应的信道状态信息CSI，收发单元1200可用于图6中的步骤通过N个射频链路向终端设备发送该第一信息，该N为正整数，收发单元1200还用于图6中的步骤从该终端设备接收第三信息，该第三信息由终端设备对该第二信息进行第二编码生成，该第二信息由该终端设备基于该第一信息和该CSI-RS确定，处理单元1100还可用于执行图6中的步骤对该第三信息进行第一解码，以生成该CSI，该第一解码包括对应该第一编码的解码。应理解，各单元执行上述相应步骤的具体过程在上述方法实施例中已经详细说明，为了简洁，在此不再赘述。
还应理解，该通信装置1000为接入网设备时，该通信装置1000中的收发单元1200可以通过收发器实现，例如可对应于图9中示出的通信装置2000中的收发器2020或图11中示出的基站4000中的RRU 4100，该通信装置1000中的处理单元1100可通过至少一个处理器实现，例如可对应于图9中示出的通信装置2000中的处理器2010或图11中示出的基站4000中的处理单元4200或处理器4202。
还应理解，该通信装置1000为配置于网络设备中的芯片或芯片系统时，该通信装置1000中的收发单元1200可以通过输入/输出接口、电路等实现，该通信装置1000中的处理单元1100可以通过该芯片或芯片系统上集成的处理器、微处理器或集成电路等实现。
图9是本申请实施例提供的装置2000的另一示意性框图。如图9所示，该装置2000包括处理器2010、收发器2020和存储器2030。其中，处理器2010、收发器2020和存储器2030通过内部连接通路互相通信，该存储器2030用于存储指令，该处理器2010用于执行该存储器2030存储的指令，以控制该收发器2020发送信号和/或接收信号。
应理解，该装置2000可以对应于上述方法实施例中的终端设备，并且可以用于执行上述方法实施例中网络设备或终端设备执行的各个步骤和/或流程。可选地，该存储器2030可以包括只读存储器和随机存取存储器，并向处理器提供指令和数据。存储器的一部分还可以包括非易失性随机存取存储器。存储器2030可以是一个单独的器件，也可以集成在处理器2010中。该处理器2010可以用于执行存储器2030中存储的指令，并且当该处理器2010执行存储器中存储的指令时，该处理器2010用于执行上述与网络设备或终端设备对应的方法实施例的各个步骤和/或流程。
可选地，该装置2000是前文实施例中的终端设备。
可选地，该装置2000是前文实施例中的网络设备。
其中，收发器2020可以包括发射机和接收机。收发器2020还可以进一步包括天线，天线的数量可以为一个或多个。该处理器2010和存储器2030与收发器2020可以是集成在不同芯片上的器件。如，处理器2010和存储器2030可以集成在基带芯片中，收发器2020可以集成在射频芯片中。该处理器2010和存储器2030与收发器2020也可以是集成在同一个芯片上的器件。本申请对此不作限定。
可选地，该装置2000是配置在终端设备中的部件，如电路、芯片、芯片系统等。
可选地，该装置2000是配置在网络设备中的部件，如电路、芯片、芯片系统等。
其中，收发器2020也可以是通信接口，如输入/输出接口、电路等。该收发器2020与处理器2010和存储器2020都可以集成在同一个芯片中，如集成在基带芯片中。
图10是本申请实施例提供的终端设备3000的结构示意图。该终端设备3000可应用于如图1所示的系统中，执行上述方法实施例中终端设备的功能。如图所示，该终端设备3000包括处理器3010和收发器3020。可选地，该终端设备3000还包括存储器3030。其中，处理器3010、收发器3020和存储器3030之间可以通过内部连接通路互相通信，传递控制和/或数据信号，该存储器3030用于存储计算机程序，该处理器3010用于从该存储器3030中调用并运行该计算机程序，以控制该收发器3020收发信号。可选地，终端设备3000还可以包括天线3040，用于将收发器3020输出的上行数据或上行控制信令通过无线信号发送出去。
上述处理器3010可以和存储器3030可以合成一个处理装置，处理器3010用于执行存储器3030中存储的程序代码来实现上述功能。具体实现时，该存储器3030也可以集成在处理器3010中，或者独立于处理器3010。该处理器3010可以与图8中的处理单元1100或图9中的处理器2010对应。
上述收发器3020可以与图8中的收发单元1200或图9中的收发器2020对应。收发器3020可以包括接收器(或称接收机、接收电路)和发射器(或称发射机、发射电路)。其中，接收器用于接收信号，发射器用于发射信号。
应理解，图7所示的终端设备3000能够实现图3或图6所示方法实施例中涉及终端设备的各个过程。终端设备3000中的各个模块的操作和/或功能，分别为了实现上述方法实施例中的相应流程。具体可参见上述方法实施例中的描述，为避免重复，此处适当省略详细描述。
上述处理器3010可以用于执行前面方法实施例中描述的由终端设备内部实现的动作，而收发器3020可以用于执行前面方法实施例中描述的终端设备向网络设备发送或从网络设备接收的动作。具体请见前面方法实施例中的描述，此处不再赘述。
可选地，上述终端设备3000还可以包括电源3050，用于给终端设备中的各种器件或电路提供电源。
除此之外，为了使得终端设备的功能更加完善，该终端设备3000还可以包括输入单元3060、显示单元3070、音频电路3080、摄像头3090和传感器3100等中的一个或多个，所述音频电路还可以包括扬声器3082、麦克风3084等。
图11是本申请实施例提供的接入网设备的结构示意图，例如可以为基站的结构示意图。该基站4000可应用于如图1所示的系统中，执行上述方法实施例中接入网设备的功能。如图所示，该基站4000可以包括一个或多个射频单元，如远端射频单元(remote radio unit，RRU)4100和一个或多个基带单元(BBU)(也可称为分布式单元(DU))4200。所述RRU 4100可以称为收发单元，可以与图5中的收发单元1200或图6中的收发器2020对应。可选地，该RRU 4100还可以称为收发机、收发电路、或者收发器等等，其可以包括至少一个天线4101和射频单元4102。可选地，RRU 4100可以包括接收单元和发送单元，接收单元可以对应于接收器(或称接收机、接收电路)，发送单元可以对应于发射器(或称发射机、发射电路)。所述RRU 4100部分主要用于射频信号的收发以及射频信号 与基带信号的转换，例如用于向终端设备发送指示信息。所述BBU 4200部分主要用于进行基带处理，对基站进行控制等。所述RRU 4100与BBU 4200可以是物理上设置在一起，也可以物理上分离设置的，即分布式基站。
所述BBU 4200为基站的控制中心，也可以称为处理单元，可以与图8中的处理单元1100或图9中的处理器2010对应，主要用于完成基带处理功能，如信道编码，复用，调制，扩频等等。例如所述BBU(处理单元)可以用于控制基站执行上述方法实施例中关于网络设备的操作流程，例如，生成上述指示信息等。
在一个示例中，所述BBU 4200可以由一个或多个单板构成，多个单板可以共同支持单一接入制式的无线接入网(如LTE网)，也可以分别支持不同接入制式的无线接入网(如LTE网，5G网或其他网)。所述BBU 4200还包括存储器4201和处理器4202。所述存储器4201用以存储必要的指令和数据。所述处理器4202用于控制基站进行必要的动作，例如用于控制基站执行上述方法实施例中关于网络设备的操作流程。所述存储器4201和处理器4202可以服务于一个或多个单板。也就是说，可以每个单板上单独设置存储器和处理器。也可以是多个单板共用相同的存储器和处理器。此外每个单板上还可以设置有必要的电路。
应理解，图11所示的基站4000能够实现图2，图3或图4所示方法实施例中涉及接入网设备的各个过程。基站4000中的各个模块的操作和/或功能，分别为了实现上述方法实施例中的相应流程。具体可参见上述方法实施例中的描述，为避免重复，此处适当省略详细描述。
上述BBU 4200可以用于执行前面方法实施例中描述的由接入网设备内部实现的动作，而RRU 4100可以用于执行前面方法实施例中描述的接入网设备向终端设备发送或从终端设备接收的动作。具体请见前面方法实施例中的描述，此处不再赘述。
应理解，图11所示出的基站4000仅为接入网设备的一种可能的形态，而不应对本申请构成任何限定。本申请所提供的方法可适用于其他形态的接入网设备。例如，包括AAU，还可以包括CU和/或DU，或者包括BBU和自适应无线单元(adaptive radio unit，ARU)，或BBU；也可以为客户终端设备(customer premises equipment，CPE)，还可以为其它形态，本申请对于接入网设备的具体形态不做限定。
其中，CU和/或DU可以用于执行前面方法实施例中描述的由接入网设备内部实现的动作，而AAU可以用于执行前面方法实施例中描述的接入网设备向终端设备发送或从终端设备接收的动作。具体请见前面方法实施例中的描述，此处不再赘述。
本申请还提供了一种处理装置，包括至少一个处理器，所述至少一个处理器用于执行存储器中存储的计算机程序，以使得所述处理装置执行上述任一方法实施例中终端设备或接入网设备所执行的方法。
本申请实施例还提供了一种处理装置，包括处理器和通信接口。所述通信接口与所述处理器耦合。所述通信接口用于输入和/或输出信息。所述信息包括指令和数据中的至少一项。所述处理器用于执行计算机程序，以使得所述处理装置执行上述任一方法实施例中终端设备或接入网设备所执行的方法。
本申请实施例还提供了一种处理装置，包括处理器和存储器。所述存储器用于存储计算机程序，所述处理器用于从所述存储器调用并运行所述计算机程序，以使得所述处理装 置执行上述任一方法实施例中终端设备或接入网设备所执行的方法。
应理解，上述处理装置可以是一个或多个芯片。例如，该处理装置可以是现场可编程门阵列(field programmable gate array，FPGA)，可以是专用集成芯片(application specific integrated circuit，ASIC)，还可以是系统芯片(system on chip，SoC)，还可以是中央处理器(central processor unit，CPU)，还可以是网络处理器(network processor，NP)，还可以是数字信号处理电路(digital signal processor，DSP)，还可以是微控制器(micro controller unit，MCU)，还可以是可编程控制器(programmable logic device，PLD)或其他集成芯片。
在实现过程中，上述方法的各步骤可以通过处理器中的硬件的集成逻辑电路或者软件形式的指令完成。结合本申请实施例所公开的方法的步骤可以直接体现为硬件处理器执行完成，或者用处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器，处理器读取存储器中的信息，结合其硬件完成上述方法的步骤。为避免重复，这里不再详细描述。
应注意，本申请实施例中的处理器可以是一种集成电路芯片，具有信号的处理能力。在实现过程中，上述方法实施例的各步骤可以通过处理器中的硬件的集成逻辑电路或者软件形式的指令完成。上述的处理器可以是通用处理器、数字信号处理器(DSP)、专用集成电路(ASIC)、现场可编程门阵列(FPGA)或者其他可编程逻辑器件、分立门或者晶体管逻辑器件、分立硬件组件。可以实现或者执行本申请实施例中的公开的各方法、步骤及逻辑框图。通用处理器可以是微处理器或者该处理器也可以是任何常规的处理器等。结合本申请实施例所公开的方法的步骤可以直接体现为硬件译码处理器执行完成，或者用译码处理器中的硬件及软件模块组合执行完成。软件模块可以位于随机存储器，闪存、只读存储器，可编程只读存储器或者电可擦写可编程存储器、寄存器等本领域成熟的存储介质中。该存储介质位于存储器，处理器读取存储器中的信息，结合其硬件完成上述方法的步骤。
可以理解，本申请实施例中的存储器可以是易失性存储器或非易失性存储器，或可包括易失性和非易失性存储器两者。其中，非易失性存储器可以是只读存储器(read-only memory，ROM)、可编程只读存储器(programmable ROM，PROM)、可擦除可编程只读存储器(erasable PROM，EPROM)、电可擦除可编程只读存储器(electrically EPROM，EEPROM)或闪存。易失性存储器可以是随机存取存储器(random access memory，RAM)，其用作外部高速缓存。通过示例性但不是限制性说明，许多形式的RAM可用，例如静态随机存取存储器(static RAM，SRAM)、动态随机存取存储器(dynamic RAM，DRAM)、同步动态随机存取存储器(synchronous DRAM，SDRAM)、双倍数据速率同步动态随机存取存储器(double data rate SDRAM，DDR SDRAM)、增强型同步动态随机存取存储器(enhanced SDRAM，ESDRAM)、同步连接动态随机存取存储器(synchlink DRAM，SLDRAM)和直接内存总线随机存取存储器(direct rambus RAM，DR RAM)。应注意，本文描述的系统和方法的存储器旨在包括但不限于这些和任意其它适合类型的存储器。
根据本申请实施例提供的方法，本申请还提供一种计算机程序产品，该计算机程序产品包括：计算机程序代码，当该计算机程序代码在计算机上运行时，使得该计算机执行图 3所示实施例中的终端设备执行的方法或网络设备执行的方法。
根据本申请实施例提供的方法，本申请还提供一种计算机可读存储介质，该计算机可读存储介质存储有程序代码，当该程序代码在计算机上运行时，使得该计算机执行图3所示实施例中的终端设备执行的方法或网络设备执行的方法。
根据本申请实施例提供的方法，本申请还提供一种系统，其包括前述的一个或多个终端设备以及一个或多个网络设备。
上述各个装置实施例中网络设备与终端设备和方法实施例中的网络设备或终端设备完全对应，由相应的模块或单元执行相应的步骤，例如通信单元(收发器)执行方法实施例中接收或发送的步骤，除发送、接收外的其它步骤可以由处理单元(处理器)执行。具体单元的功能可以参考相应的方法实施例。其中，处理器可以为一个或多个。
上述实施例中，终端设备可以作为接收设备的一例，网络设备可以作为发送设备的一例。但这不应对本申请构成任何限定。例如，发送设备和接收设备也可以均为终端设备等。本申请对于发送设备和接收设备的具体类型不作限定。
在本说明书中使用的术语“部件”、“模块”、“系统”等用于表示计算机相关的实体、硬件、固件、硬件和软件的组合、软件、或执行中的软件。例如，部件可以是但不限于，在处理器上运行的进程、处理器、对象、可执行文件、执行线程、程序和/或计算机。通过图示，在计算设备上运行的应用和计算设备都可以是部件。一个或多个部件可驻留在进程和/或执行线程中，部件可位于一个计算机上和/或分布在2个或更多个计算机之间。此外，这些部件可从在上面存储有各种数据结构的各种计算机可读介质执行。部件可例如根据具有一个或多个数据分组(例如来自与本地系统、分布式系统和/或网络间的另一部件交互的二个部件的数据，例如通过信号与其它系统交互的互联网)的信号通过本地和/或远程进程来通信。
本领域普通技术人员可以意识到，结合本文中所公开的实施例描述的各示例的单元及算法步骤，能够以电子硬件、或者计算机软件和电子硬件的结合来实现。这些功能究竟以硬件还是软件方式来执行，取决于技术方案的特定应用和设计约束条件。专业技术人员可以对每个特定的应用来使用不同方法来实现所描述的功能，但是这种实现不应认为超出本申请的范围。
所属领域的技术人员可以清楚地了解到，为描述的方便和简洁，上述描述的系统、装置和单元的具体工作过程，可以参考前述方法实施例中的对应过程，在此不再赘述。
在本申请所提供的几个实施例中，应该理解到，所揭露的系统、装置和方法，可以通过其它的方式实现。例如，以上所描述的装置实施例仅仅是示意性的，例如，所述单元的划分，仅仅为一种逻辑功能划分，实际实现时可以有另外的划分方式，例如多个单元或组件可以结合或者可以集成到另一个系统，或一些特征可以忽略，或不执行。另一点，所显示或讨论的相互之间的耦合或直接耦合或通信连接可以是通过一些接口，装置或单元的间接耦合或通信连接，可以是电性，机械或其它的形式。
所述作为分离部件说明的单元可以是或者也可以不是物理上分开的，作为单元显示的部件可以是或者也可以不是物理单元，即可以位于一个地方，或者也可以分布到多个网络单元上。可以根据实际的需要选择其中的部分或者全部单元来实现本实施例方案的目的。
另外，在本申请各个实施例中的各功能单元可以集成在一个处理单元中，也可以是各 个单元单独物理存在，也可以两个或两个以上单元集成在一个单元中。
所述功能如果以软件功能单元的形式实现并作为独立的产品销售或使用时，可以存储在一个计算机可读取存储介质中。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分或者该技术方案的部分可以以软件产品的形式体现出来，该计算机软件产品存储在一个存储介质中，包括若干指令用以使得一台计算机设备(可以是个人计算机，服务器，或者网络设备等)执行本申请各个实施例所述方法的全部或部分步骤。而前述的存储介质包括：U盘、移动硬盘、只读存储器(Read-Only Memory，ROM)、随机存取存储器(Random Access Memory，RAM)、磁碟或者光盘等各种可以存储程序代码的介质。
以上所述，仅为本申请的具体实施方式，但本申请的保护范围并不局限于此，任何熟悉本技术领域的技术人员在本申请揭露的技术范围内，可轻易想到变化或替换，都应涵盖在本申请的保护范围之内。因此，本申请的保护范围应以所述权利要求的保护范围为准。

Claims (28)

  1. 一种传输数据的方法，其特征在于，所述方法包括：
    接入网设备对信道状态信息参考信号CSI-RS进行第一编码，以生成第一信息，所述CSI-RS用于获取所述接入网设备和终端设备之间信道对应的信道状态信息CSI；
    所述接入网设备通过N个射频链路向所述终端设备发送所述第一信息，所述N为正整数；
    所述接入网设备从所述终端设备接收第三信息，所述第三信息由所述终端设备对第二信息进行第二编码生成，其中，所述第二信息由所述终端设备基于所述第一信息和所述CSI-RS确定；
    所述接入网设备对所述第三信息进行第一解码，以生成所述CSI，所述第一解码包括对应所述第一编码的解码。
  2. 根据权利要求1所述的方法，其特征在于，所述方法还包括：
    所述接入网设备对所述第三信息进行第二解码，所述第二解码与所述第二编码对应；
    所述接入网设备对所述第三信息进行第一解码，包括：
    所述接入网设备对经过所述第二解码的所述第三信息进行所述第一解码。
  3. 根据权利要求1所述的方法，其特征在于，所述第一解码还包括对应所述第二编码的解码。
  4. 根据权利要求2或3所述的方法，其特征在于，所述第二编码为基于第一神经网络的编码，所述第一神经网络的参数与所述信道的采样点数量F和所述第一信息占用的资源的数量M有关，所述M和所述F为正整数，M<N，所述资源包括以下至少一种：时域资源、频域资源或码域资源。
  5. 根据权利要求4所述的方法，其特征在于，所述第一神经网络的参数为所述第一神经网络的输入矩阵的维度。
  6. 根据权利要求1至5中任一项所述的方法，其特征在于，所述第一编码为基于压缩感知方式的编码，所述第一编码使用第一矩阵，所述第一矩阵的维度与所述M和所述N有关。
  7. 根据权利要求1至5中任一项所述的方法，其特征在于，所述第一编码为基于第二神经网络的编码，所述第二神经网络包括全连接线性层，所述全连接线性层的参数与所述M和所述N有关。
  8. 根据权利要求1至7中任一项所述的方法，其特征在于，所述第一解码为基于第三神经网络的解码，所述第三神经网络的参数与所述N，所述M和所述F有关。
  9. 一种传输数据的方法，其特征在于，所述方法包括：
    终端设备接收接入网设备通过N个射频链路向所述终端设备发送的第一信息，所述第一信息由所述接入网设备对信道状态信息参考信号CSI-RS进行第一编码生成，所述CSI-RS用于获取信道状态信息CSI；
    所述终端设备基于所述第一信息和所述CSI-RS确定第二信息；
    所述终端设备对所述第二信息进行第二编码，以生成第三信息；
    所述终端设备向所述接入网设备发送所述第三信息，所述第三信息用于所述接入网设备对所述第三信息进行第一解码，以生成所述CSI，所述第一解码包括对应所述第一编码的解码。
  10. 根据权利要求9所述的方法，其特征在于，所述第三信息用于所述接入网设备对经过第二解码的所述第三信息进行所述第一解码，以生成所述CSI，所述第二解码与所述第二编码对应。
  11. 根据权利要求9所述的方法，其特征在于，所述第一解码还包括对应所述第二编码的解码。
  12. 根据权利要求10或11所述的方法，其特征在于，所述第二编码为基于第一神经网络的编码，所述第一神经网络的参数与所述信道的采样点数量F和所述第一信息占用的资源的数量M有关，所述M和所述F为正整数，M<N，所述资源包括以下至少一种：时域资源、频域资源或码域资源。
  13. 根据权利要求12所述的方法，其特征在于，所述第一神经网络的参数为所述第一神经网络的输入矩阵的维度。
  14. 根据权利要求9至13中任一项所述的方法，其特征在于，所述第一编码为基于压缩感知方式的编码，所述第一编码使用第一矩阵，所述第一矩阵的维度与所述M和所述N有关。
  15. 根据权利要求9至13中任一项所述的方法，其特征在于，所述第一编码为基于第二神经网络的编码，所述第二神经网络包括全连接线性层，所述全连接线性层的参数与所述M和所述N有关。
  16. 根据权利要求9至15中任一项所述的方法，其特征在于，所述第一解码为基于第三神经网络的解码，所述第三神经网络的参数与所述N，所述M和所述F有关。
  17. 一种传输数据的装置，其特征在于，所述装置包括：
    收发单元，用于接收接入网设备通过N个射频链路发送的第一信息，所述第一信息由接入网设备对信道状态信息参考信号CSI-RS进行第一编码生成，所述CSI-RS用于获取信道信息CSI；
    处理单元，用于基于所述第一信息和所述CSI-RS确定第二信息；
    所述处理单元，还用于对所述第二信息进行第二编码，以生成第三信息；
    所述收发单元，还用于向所述接入网设备发送所述第三信息，所述第三信息用于所述接入网设备对所述第三信息进行第一解码，以生成所述CSI，所述第一解码包括对应所述第一编码的解码。
  18. 根据权利要求17所述的装置，其特征在于，所述第三信息用于所述接入网设备对经过第二解码的所述第三信息进行所述第一解码，以生成所述CSI，所述第二解码与所述第二编码对应。
  19. 根据权利要求17所述的装置，其特征在于，所述第一解码还包括对应所述第二编码的解码。
  20. 根据权利要求17至19中任一项所述的装置，其特征在于，所述第二编码为基于第一神经网络的编码，所述第一神经网络的参数与所述信道的采样点数量F和所述第一信息占用的资源的数量M有关，所述M和所述F为正整数，M<N，所述资源包括以下至少 一种：时域资源、频域资源或码域资源。
  21. 根据权利要求20所述的装置，其特征在于，所述第一神经网络的参数为所述第一神经网络的输入矩阵的维度。
  22. 根据权利要求17至21中任一项所述的装置，其特征在于，所述第一编码为基于压缩感知方式的编码，所述第一编码使用第一矩阵，所述第一矩阵的维度与所述M和所述N有关。
  23. 根据权利要求17至21中任一项所述的装置，其特征在于，所述第一编码为基于第二神经网络的编码，所述第二神经网络包括全连接线性层，所述全连接线性层的参数与所述M和所述N有关。
  24. 根据权利要求17至23中任一项所述的装置，其特征在于，所述第一解码为基于第三神经网络的解码，所述第三神经网络的参数与所述N，所述M和所述F有关。
  25. 一种处理装置，其特征在于，包括至少一个处理器，所述至少一个处理器用于执行存储器中存储的计算机程序，以使得所述装置实现如权利要求1至8中任一项所述的方法、或权利要求9至16中任一项所述的方法。
  26. 一种计算机可读存储介质，其特征在于，包括计算机程序，当所述计算机程序在计算机上运行时，使得所述计算机执行如权利要求1至8中任一项所述的方法、或权利要求9至16中任一项所述的方法。
  27. 一种计算机程序产品，其特征在于，当所述计算机程序产品在计算机上运行时，使得所述计算机执行如权利要求1至8中任一项所述的方法、或权利要求9至16中任一项所述的方法。
  28. 一种芯片系统，其特征在于，包括：处理器，用于从存储器中调用并运行计算机程序，使得安装有所述芯片系统的通信设备执行如权利要求1至8中任意一项所述的方法、或权利要求9至16中任一项所述的方法。
PCT/CN2021/135367 2020-12-08 2021-12-03 一种传输数据的方法和装置 Ceased WO2022121797A1 (zh)

Priority Applications (2)

Application Number Priority Date Filing Date Title
EP21902504.6A EP4243492A4 (en) 2020-12-08 2021-12-03 DATA TRANSMISSION METHOD AND DEVICE
US18/265,495 US20240030980A1 (en) 2020-12-08 2021-12-03 Data transmission method and apparatus

Applications Claiming Priority (6)

Application Number Priority Date Filing Date Title
CN202011444244.0 2020-12-08
CN202011444244 2020-12-08
CN202110012794.3 2021-01-06
CN202110012794 2021-01-06
CN202110214243.5A CN114614955B (zh) 2020-12-08 2021-02-25 一种传输数据的方法和装置
CN202110214243.5 2021-02-25

Publications (1)

Publication Number Publication Date
WO2022121797A1 true WO2022121797A1 (zh) 2022-06-16

Family

ID=81857856

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2021/135367 Ceased WO2022121797A1 (zh) 2020-12-08 2021-12-03 一种传输数据的方法和装置

Country Status (4)

Country Link
US (1) US20240030980A1 (zh)
EP (1) EP4243492A4 (zh)
CN (1) CN114614955B (zh)
WO (1) WO2022121797A1 (zh)

Cited By (2)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115604158A (zh) * 2022-12-15 2023-01-13 中国人民解放军国防科技大学（Cn） 一种智能设备识别方法、装置、设备及可读存储介质
WO2025139843A1 (zh) * 2023-12-29 2025-07-03 华为技术有限公司 通信方法和通信装置

Families Citing this family (6)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20220284267A1 (en) * 2021-03-05 2022-09-08 Qualcomm Incorporated Architectures for temporal processing associated with wireless transmission of encoded data
WO2024026793A1 (zh) * 2022-08-04 2024-02-08 北京小米移动软件有限公司 数据传输方法、装置、设备、存储介质及系统
CN115913857B (zh) * 2022-09-13 2024-05-28 成都芯通软件有限公司 数据处理方法、装置、射频单元、基站和存储介质
WO2024065804A1 (zh) * 2022-09-30 2024-04-04 华为技术有限公司 数据压缩传输方法、装置、设备以及存储介质
WO2025231804A1 (en) * 2024-05-10 2025-11-13 Qualcomm Incorporated Techniques for channel state feedback
CN120958866A (zh) * 2024-05-30 2025-11-14 北京小米移动软件有限公司 数据处理方法、第一设备、第二设备以及系统

Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109417412A (zh) * 2017-06-16 2019-03-01 华为技术有限公司 一种信道状态信息发送、接收方法及设备
CN109787664A (zh) * 2017-11-15 2019-05-21 索尼公司 用于无线通信系统的电子设备、方法、装置和存储介质
WO2020180221A1 (en) * 2019-03-06 2020-09-10 Telefonaktiebolaget Lm Ericsson (Publ) Compression and decompression of downlink channel estimates

Family Cites Families (11)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN108111197A (zh) * 2016-11-25 2018-06-01 索尼公司 电子设备和通信方法
CN108282321B (zh) * 2017-01-06 2022-03-29 华为技术有限公司 一种信息指示的方法、网络设备和终端设备
CN109150265B (zh) * 2017-06-15 2021-08-31 华为技术有限公司 发送和接收参考信号的方法、网络设备和终端设备
CN109327251A (zh) * 2017-08-01 2019-02-12 中国移动通信有限公司研究院 基于pmi的处理方法、装置、相关设备及存储介质
CN108990167B (zh) * 2018-07-11 2021-09-07 东南大学 一种机器学习辅助的大规模mimo下行用户调度方法
US12184361B2 (en) * 2019-01-11 2024-12-31 Telefonaktiebolaget Lm Ericsson (Publ) Low overhead channel state information (CSI) feedback for multi-transmission point (TRP) transmission
CN110535571A (zh) * 2019-03-28 2019-12-03 中兴通讯股份有限公司 一种信息传输方法和装置、信息确定方法和装置
CN111756426B (zh) * 2019-03-29 2023-04-07 华为技术有限公司 一种选择接收波束的方法及装置
CN111817798B (zh) * 2019-04-11 2022-04-12 华为技术有限公司 一种信道测量方法和通信装置
WO2020213964A1 (en) * 2019-04-16 2020-10-22 Samsung Electronics Co., Ltd. Method and apparatus for reporting channel state information
CN116420317A (zh) * 2020-09-24 2023-07-11 交互数字专利控股公司 用于mimo系统中的信道老化的自适应学习辅助预编码器的方法、架构、装置和系统

Patent Citations (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109417412A (zh) * 2017-06-16 2019-03-01 华为技术有限公司 一种信道状态信息发送、接收方法及设备
CN109787664A (zh) * 2017-11-15 2019-05-21 索尼公司 用于无线通信系统的电子设备、方法、装置和存储介质
WO2020180221A1 (en) * 2019-03-06 2020-09-10 Telefonaktiebolaget Lm Ericsson (Publ) Compression and decompression of downlink channel estimates

Non-Patent Citations (1)

* Cited by examiner, † Cited by third party
Title
See also references of EP4243492A4

Cited By (3)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN115604158A (zh) * 2022-12-15 2023-01-13 中国人民解放军国防科技大学（Cn） 一种智能设备识别方法、装置、设备及可读存储介质
CN115604158B (zh) * 2022-12-15 2023-03-21 中国人民解放军国防科技大学 一种智能设备识别方法、装置、设备及可读存储介质
WO2025139843A1 (zh) * 2023-12-29 2025-07-03 华为技术有限公司 通信方法和通信装置

Also Published As

Publication number Publication date
EP4243492A1 (en) 2023-09-13
EP4243492A4 (en) 2024-05-08
CN114614955B (zh) 2024-11-22
CN114614955A (zh) 2022-06-10
US20240030980A1 (en) 2024-01-25

Similar Documents

Publication Publication Date Title
WO2022121797A1 (zh) 一种传输数据的方法和装置
CN113765830B (zh) 获取信道信息的方法及通信装置
CN112751592B (zh) 上报信道状态信息的方法和通信装置
CN109873665B (zh) 数据传输的方法和设备
WO2022267633A1 (zh) 信息传输的方法和装置
CN113873539B (zh) 获取神经网络的方法和装置
WO2021083157A1 (zh) 一种预编码矩阵的处理方法和通信装置
US20230239028A1 (en) Channel measurement method and communication apparatus
WO2023092307A1 (zh) 通信方法、模型训练方法和设备
US20240372591A1 (en) Method for channel information feedback, transmitting device and receiving device
CN114389661B (zh) 信道测量的方法及通信装置
WO2022012256A1 (zh) 通信的方法及通信装置
CN117678257A (zh) 信息处理方法、模型生成方法及设备
WO2021134207A1 (zh) 一种指示和确定信道状态信息的方法及通信装置
WO2025201297A1 (zh) 一种信道状态信息csi上报方法及相关装置
WO2025232458A1 (zh) 数据压缩传输方法、装置、设备以及存储介质
WO2025252058A1 (zh) 通信方法和通信装置
CN121666704A (zh) 用于通信网络中的信道状态信息反馈增强的givens旋转矩阵参数化预处理
WO2025218595A1 (zh) 一种通信方法及装置
CN116800309A (zh) 一种分布式预编码的方法及装置
WO2026016502A1 (zh) 一种数据传输方法及通信装置
WO2025103061A1 (zh) 数据传输方法及通信装置
WO2025256568A1 (zh) 一种通信方法及相关装置
CN119923834A (zh) 无线通信的方法、终端设备和网络设备
CN121865408A (zh) 一种通信的方法和通信装置

Legal Events

Date Code Title Description
121 Ep: the epo has been informed by wipo that ep was designated in this application

Ref document number: 21902504

Country of ref document: EP

Kind code of ref document: A1

WWE Wipo information: entry into national phase

Ref document number: 18265495

Country of ref document: US

ENP Entry into the national phase

Ref document number: 2021902504

Country of ref document: EP

Effective date: 20230607

NENP Non-entry into the national phase

Ref country code: DE