WO2023221061A1 - 获取信道质量的方法、装置、存储介质和芯片 - Google Patents

获取信道质量的方法、装置、存储介质和芯片 Download PDF

Info

Publication number
WO2023221061A1
WO2023221061A1 PCT/CN2022/093972 CN2022093972W WO2023221061A1 WO 2023221061 A1 WO2023221061 A1 WO 2023221061A1 CN 2022093972 W CN2022093972 W CN 2022093972W WO 2023221061 A1 WO2023221061 A1 WO 2023221061A1
Authority
WO
WIPO (PCT)
Prior art keywords
channel matrix
channel
target
model
csi
Prior art date
Application number
PCT/CN2022/093972
Other languages
English (en)
French (fr)
Inventor
池连刚
陈栋
Original Assignee
北京小米移动软件有限公司
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 北京小米移动软件有限公司 filed Critical 北京小米移动软件有限公司
Priority to PCT/CN2022/093972 priority Critical patent/WO2023221061A1/zh
Publication of WO2023221061A1 publication Critical patent/WO2023221061A1/zh

Links

Images

Classifications

    • GPHYSICS
    • G06COMPUTING; CALCULATING OR COUNTING
    • G06NCOMPUTING ARRANGEMENTS BASED ON SPECIFIC COMPUTATIONAL MODELS
    • G06N3/00Computing arrangements based on biological models
    • G06N3/02Neural networks
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L1/00Arrangements for detecting or preventing errors in the information received
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04WWIRELESS COMMUNICATION NETWORKS
    • H04W24/00Supervisory, monitoring or testing arrangements
    • H04W24/02Arrangements for optimising operational condition

Definitions

  • the present disclosure relates to the field of communication technology, and specifically, to a method, device, storage medium and chip for obtaining channel quality.
  • mMIMO massive Multiple-Input Multiple-Output, large-scale multiple-input multiple-output
  • 5G the 5th Generation Mobile Communication Technology, fifth-generation mobile communication system
  • mMIMO massive Multiple-Input Multiple-Output, large-scale multiple-input multiple-output
  • CSI Channel State Information
  • the terminal device can report CSI information to the network device so that the network device can obtain the channel quality of the downlink channel and select an appropriate modulation and coding scheme for downlink transmission based on the channel quality to improve the performance of data transmission.
  • the number of antennas continues to increase, and the information contained in CSI becomes more and more abundant, making the resource overhead reported by CSI increasing.
  • CSI compression technology based on Discrete Fourier Transform (DFT) can be used, and the terminal device performs CSI compression before reporting to the network device.
  • DFT Discrete Fourier Transform
  • CSI compression reporting will reduce the accuracy of the channel quality obtained by network equipment.
  • the accuracy of CSI compression in different scenarios varies greatly, which will affect data transmission efficiency.
  • the present disclosure provides a method, device, storage medium and chip for obtaining channel quality.
  • a method for obtaining channel quality is provided, applied to a terminal device, and the method includes:
  • the first channel matrix is used to characterize the channel quality of the downlink channel
  • the first channel matrix is compressed according to the channel state information CSI compression model and the channel state information CSI compression parameters to obtain a compressed target channel matrix; wherein the CSI compression model includes a channel encoder, and the channel encoder Includes multiple sub-encoders; different sub-encoders correspond to different CSI compression parameters;
  • the target channel matrix is sent to the network device, so that the network device obtains the channel quality of the downlink channel according to the target channel matrix.
  • a method for obtaining channel quality is provided, applied to network equipment, and the method includes:
  • the target channel matrix is obtained by compressing the first channel matrix according to the channel state information CSI compression model and the channel state information CSI compression parameter by the terminal device, and the first channel
  • the matrix is a matrix obtained by the terminal equipment according to the pilot signal and used to characterize the channel quality of the downlink channel;
  • the target channel matrix is decompressed to obtain a third channel matrix;
  • the CSI decompression model includes a channel decoder, and the channel decoder includes Multiple sub-decoders, different sub-decoders correspond to different CSI compression parameters;
  • the channel quality of the downlink channel is determined according to the third channel matrix.
  • a device for obtaining channel quality which is applied to terminal equipment, and the device includes:
  • the first receiving module is configured to receive the pilot signal sent by the network device through the downlink channel
  • a first matrix acquisition module configured to acquire a first channel matrix according to the pilot signal; the first channel matrix is used to characterize the channel quality of the downlink channel;
  • the target matrix acquisition module is configured to compress the first channel matrix according to the channel state information CSI compression model and the channel state information CSI compression parameters to obtain a compressed target channel matrix; wherein the CSI compression model includes channel Encoder, the channel encoder includes multiple sub-encoders; different sub-encoders correspond to different CSI compression parameters;
  • the first sending module is configured to send the target channel matrix to the network device, so that the network device obtains the channel quality of the downlink channel according to the target channel matrix.
  • a device for obtaining channel quality is provided, which is applied to network equipment.
  • the device includes:
  • the second receiving module is configured to receive a target channel matrix sent by the terminal device; the target channel matrix is the first channel matrix after the terminal device compresses the first channel matrix according to the channel state information CSI compression model and the channel state information CSI compression parameter. Obtained, the first channel matrix is a matrix obtained by the terminal device according to the pilot signal and used to characterize the channel quality of the downlink channel;
  • the third matrix acquisition module is configured to decompress the target channel matrix according to the channel state information CSI decompression model and the CSI compression parameters to obtain a third channel matrix; wherein the CSI decompression model includes channel Decoder, the channel decoder includes multiple sub-decoders, and different sub-decoders correspond to different CSI compression parameters;
  • the channel quality determination module is configured to determine the channel quality of the downlink channel according to the third channel matrix.
  • an apparatus for obtaining channel quality including:
  • Memory used to store instructions executable by the processor
  • the processor is configured to execute the steps of the method for obtaining channel quality provided in the first aspect of the present disclosure.
  • an apparatus for obtaining channel quality including:
  • Memory used to store instructions executable by the processor
  • the processor is configured to execute the steps of the method for obtaining channel quality provided in the second aspect of this disclosure.
  • a computer-readable storage medium on which computer program instructions are stored.
  • the method for obtaining channel quality provided by the first aspect of the present disclosure is implemented. A step of.
  • a computer-readable storage medium on which computer program instructions are stored.
  • the method for obtaining channel quality provided in the second aspect of the present disclosure is implemented. A step of.
  • a chip including: a processor and an interface; the processor is configured to read instructions to execute the steps of the method for obtaining channel quality provided in the first aspect of the present disclosure.
  • a chip including: a processor and an interface; the processor is configured to read instructions to execute the steps of the method for obtaining channel quality provided in the second aspect of the present disclosure.
  • the terminal device receives the pilot signal sent by the network device through the downlink channel; obtains the first channel matrix according to the pilot signal; and according to the channel state information CSI compression model and channel state information CSI compression parameters: compress the first channel matrix to obtain a compressed target channel matrix; send the target channel matrix to the network device so that the network device determines the channel quality of the downlink channel based on the target channel matrix.
  • the first channel matrix is used to characterize the channel quality of the downlink channel;
  • the CSI compression model may include a channel encoder, and the channel encoder includes multiple sub-encoders; different sub-encoders correspond to different CSI compression parameters; the target channel matrix is used to Instructs network equipment to determine the channel quality of the downlink channel.
  • different CSI compression parameters can be adapted through multiple sub-encoders, so that in scenarios where the CSI compression rate changes, a more accurate target channel matrix can be adaptively obtained and sent to the network device, so that the network device can obtain a more accurate target channel matrix. Accurate channel quality, thereby improving data transmission efficiency.
  • FIG. 1 is a block diagram of a communication system according to an exemplary embodiment.
  • Figure 2 is a flowchart of a method for obtaining channel quality according to an exemplary embodiment.
  • Figure 3 is a flowchart of a method for obtaining channel quality according to an exemplary embodiment.
  • Figure 4 is a schematic structural diagram of a network model for obtaining channel quality according to an exemplary embodiment.
  • FIG. 5 is a schematic diagram of a feature converter in a CSI compression model according to an exemplary embodiment.
  • Figure 6 is a schematic diagram of a channel encoder in a CSI compression model according to an exemplary embodiment.
  • Figure 7 is a schematic diagram of a CSI decompression model according to an exemplary embodiment.
  • Figure 8 is a flowchart of a training method for a CSI compression model according to an exemplary embodiment.
  • Figure 9 is a flowchart of a training method for a CSI decompression model according to an exemplary embodiment.
  • Figure 10 is a flowchart of a method for obtaining channel quality according to an exemplary embodiment.
  • Figure 11 is a block diagram of an apparatus for obtaining channel quality according to an exemplary embodiment.
  • Figure 12 is a block diagram of an apparatus for obtaining channel quality according to an exemplary embodiment.
  • Figure 13 is a block diagram of an apparatus for obtaining channel quality according to an exemplary embodiment.
  • Figure 14 is a block diagram of an apparatus for obtaining channel quality according to an exemplary embodiment.
  • Figure 15 is a block diagram of an apparatus for obtaining channel quality according to an exemplary embodiment.
  • Figure 16 is a block diagram of an apparatus for obtaining channel quality according to an exemplary embodiment.
  • plural refers to two or more than two, and other quantifiers are similar; "at least one of the following” or similar expressions refers to these Any combination of items, including any combination of single items (items) or plural items (items).
  • at least one of a, b, or c can represent: a, b, c, a-b, a-c, b-c, or a-b-c, where a, b, c can be single or multiple ;
  • “And/or” is an association relationship that describes related objects, indicating that there can be three kinds of relationships.
  • a and/or B can mean: A alone exists, A and B exist simultaneously, and B alone exists. situation, where A and B can be singular or plural.
  • CSI compression technology based on Discrete Fourier Transform (DFT) can be used.
  • DFT Discrete Fourier Transform
  • the terminal device performs CSI compression before reporting to the network device.
  • CSI compression reporting will reduce the accuracy of the channel quality obtained by network equipment and affect data transmission efficiency.
  • the terminal device can use a pre-trained coding neural network to compress the CSI and report the compressed CSI to the network device.
  • the network device also uses the decoding neural network to decompress the compressed CSI to determine the channel quality.
  • the encoding neural network and decoding neural network have different requirements and parameters. Therefore, in a scenario where the CSI compression rate changes, if the same encoding neural network and decoding neural network are used, there will be As a result, the accuracy of CSI reporting varies greatly.
  • the CSI compression rate changes retraining the encoding neural network and decoding neural network will be inefficient and unable to adapt to scenarios where the CSI compression rate changes frequently.
  • the present disclosure provides a method, device, storage medium and chip for obtaining channel quality.
  • FIG. 1 is a schematic diagram of a communication system according to an exemplary embodiment.
  • the communication system may include a terminal device 101 and a network device 102.
  • the communication system can be used to support 4G (the 4th Generation, fourth generation) network access technology, such as Long Term Evolution (LTE) access technology, or 5G (the 5th Generation, the fifth generation) network access technology.
  • Access technologies such as New Radio Access Technology (New RAT), or other future wireless communication technologies.
  • the communication system may be a communication system using FDD (Frequency Division Duplexing, frequency division duplexing) technology, or it may be a communication system using TDD (Time Division Duplexing, time division duplexing) technology.
  • the number of network devices and terminal devices can be one or more. The number of network devices and terminal devices in the communication system shown in Figure 1 is only an example of adaptability, and this disclosure does not limit it. .
  • the network equipment in Figure 1 can be used to support terminal access, for example, it can be an evolutionary base station (eNB or eNodeB) in LTE; or a 5G network or a future evolved public land mobile network (public land mobile network, PLMN) base stations, Broadband Network Gateway (BNG), aggregation switches or non-3GPP (3rd Generation Partnership Project, third generation partner project) access equipment, etc.
  • eNB evolutionary base station
  • PLMN public land mobile network
  • BNG Broadband Network Gateway
  • aggregation switches or non-3GPP (3rd Generation Partnership Project, third generation partner project) access equipment, etc.
  • the network equipment in the embodiments of the present disclosure may include various forms of base stations, such as: macro base stations, micro base stations (also called small stations), relay stations, access points, 5G base stations or future base stations, satellites, Transmitting and receiving point (TRP), transmitting point (TP), mobile switching center and device-to-device (D2D), vehicle outreach (Vehicle-to-Everything, V2X), Embodiments of the present disclosure do not specifically limit devices that undertake base station functions in machine-to-machine (M2M) communications.
  • M2M machine-to-machine
  • devices that provide wireless communication functions for terminal devices are collectively referred to as network equipment or base stations.
  • the terminal device in Figure 1 can be an electronic device that provides voice or data connectivity.
  • it can also be called user equipment (User Equipment, UE), subscriber unit (SubscriberUnit), mobile station (Mobile Station), and station (Station). ), Terminal, etc.
  • the terminal device may include a smartphone, a smart wearable device, a smart speaker, a smart tablet, a wireless modem, a wireless local loop (Wireless Local Loop, WLL) station, a PDA (Personal Digital Assistant, personal digital assistant) ), CPE (Customer Premise Equipment, customer terminal equipment), etc.
  • devices that can access the communication system, communicate with network devices of the communication system, or communicate with other objects through the communication system can be terminal devices in the embodiments of the present disclosure, for example, smart devices Terminals and cars in transportation, household equipment in smart homes, power meter reading instruments, voltage monitoring instruments, environmental monitoring instruments in smart grids, video monitoring instruments in smart security networks, cash registers, etc.
  • the terminal device may communicate with a network device, such as the network device in FIG. 1 . Communication between multiple terminals is also possible.
  • the terminal may be statically fixed or mobile, and this disclosure does not limit this.
  • Figure 2 is a flowchart of a method for obtaining channel quality according to an exemplary embodiment. This method can be applied to terminal equipment, as shown in Figure 2. The method can include:
  • the terminal device receives the pilot signal sent by the network device through the downlink channel.
  • the network device may send a pilot signal to the terminal device through a downlink channel.
  • the terminal device can receive the pilot signal.
  • the pilot signal may include a channel state information reference signal CSI-RS (Channel State Information Reference Signal).
  • CSI-RS Channel State Information Reference Signal
  • the terminal device obtains the first channel matrix according to the pilot signal.
  • the first channel matrix may be used to characterize the channel quality of the downlink channel.
  • the terminal device may perform channel state information CSI estimation based on the received pilot signal (such as CSI-RS) to obtain a CSI estimation matrix; and then obtain a first channel matrix representing downlink channel quality based on the CSI estimation matrix.
  • the received pilot signal such as CSI-RS
  • the terminal device compresses the first channel matrix according to the channel state information CSI compression model and the channel state information CSI compression parameters to obtain a compressed target channel matrix.
  • the CSI compression model includes a channel encoder, and the channel encoder includes multiple sub-encoders; different sub-encoders correspond to different CSI compression parameters.
  • the terminal device may determine one or more target sub-encoders from multiple sub-encoders according to the CSI compression parameters, and use the target sub-encoders to compress the first channel matrix to obtain the target channel matrix.
  • the CSI compression parameter can represent the CSI compression rate
  • the value of the CSI compression rate can be any preset value, for example, it can include: 1/2, 1/4, 1/8, 1/16, 1 /32 or 1/64, etc., this disclosure does not limit this.
  • the terminal device sends the target channel matrix to the network device.
  • the network device can determine the channel quality of the downlink channel based on the target channel matrix.
  • the terminal device receives the pilot signal sent by the network device through the downlink channel; obtains the first channel matrix according to the pilot signal; and compresses the first channel matrix according to the channel state information CSI compression model and the channel state information CSI compression parameter. , obtain the compressed target channel matrix; send the target channel matrix to the network device, so that the network device determines the channel quality of the downlink channel based on the target channel matrix.
  • the first channel matrix is used to characterize the channel quality of the downlink channel;
  • the CSI compression model includes a channel encoder, and the channel encoder includes multiple sub-encoders; different sub-encoders correspond to different CSI compression parameters;
  • the target channel matrix is used to indicate Network equipment determines the channel quality of the downstream channel.
  • different CSI compression parameters can be adapted through multiple sub-encoders, so that in scenarios where the CSI compression rate changes, a more accurate target channel matrix can be adaptively obtained and sent to the network device, so that the network device can obtain a more accurate target channel matrix. Accurate channel quality, thereby improving data transmission efficiency.
  • the terminal device can obtain the above-mentioned first channel matrix in the following manner:
  • the spatial channel matrix is measured.
  • the spatial domain channel matrix is transformed into the angle delay domain channel matrix through discrete Fourier transform.
  • the above communication system can adopt the mMIMO technology based on OFDM (Orthogonal Frequency Division Multiplexing, Orthogonal Frequency Division Multiplexing technology), the number of subcarriers is N s , and the number of antennas used by the network device for mMIMO can be N t , as described above.
  • the pilot signal may include a channel state information reference signal CSI-RS.
  • the terminal device can measure (also called estimate) the air domain channel matrix H based on the received CSI-RS.
  • the size of the air domain channel matrix H can be N s ⁇ N t , and the air domain channel matrix H can represent Channel quality per subcarrier per antenna.
  • the terminal equipment can transform the spatial domain channel matrix into an angular delay domain channel matrix through discrete Fourier transform.
  • the angular delay domain channel matrix can be obtained through the following formula (1):
  • H represents the above-mentioned spatial domain channel matrix
  • the size of H can be N s ⁇ N t
  • Ha represents the transformed angle delay domain channel matrix
  • F d and F a represent the size of N s ⁇ N s , N t ⁇
  • the discrete Fourier transform matrix of N t , N s represents the number of subcarriers corresponding to mmIMO, and N t represents the number of antennas corresponding to mmIMO.
  • the first channel matrix is determined according to the main value part of the angular delay domain channel matrix.
  • the above-mentioned angular delay domain channel matrix Ha only has values in the first N c rows. After intercepting the main value part, the main value channel matrix in the angular delay domain is obtained.
  • the size of the value channel matrix may be N c ⁇ N t .
  • the first channel matrix can be obtained Where c is the real and imaginary dimension of the channel, for example, c can be 2, and the first channel matrix can be used as the input of the CSI compression model to obtain the target channel matrix.
  • the first channel matrix representing the quality of the downlink channel can be obtained according to the pilot signal in the above manner.
  • the above CSI compression parameters may be parameters received by the terminal device from the network device, or may be parameters preset by the terminal device.
  • the CSI compression parameter may be determined by the terminal device according to parameters received from the network device. For example, the terminal device may receive the first compression parameter sent by the network device; determine the CSI compression parameter according to the first compression parameter. For example, the terminal device may receive the first compression parameter through RRC (Radio Resource Control) signaling (such as broadcast signaling or proprietary signaling for the terminal device).
  • RRC Radio Resource Control
  • the network device may preset the value of the CSI compression parameter, determine the first compression parameter based on the CSI compression parameter, and send the first compression parameter to the terminal device.
  • the CSI compression parameter and the first compression parameter may be determined based on a preset corresponding relationship between the first compression parameter.
  • the first compression parameter corresponding to the value 1/2 of the CSI compression parameter may be 1
  • the first compression parameter corresponding to the value 1/4 of the CSI compression parameter may be is 2
  • the first compression parameter corresponding to the value 1/8 of the CSI compression parameter can be 3, and other values can be deduced in the same way.
  • the terminal device after receiving the first compression parameter sent by the network device, the terminal device can determine the CSI compression parameter according to the first compression parameter.
  • the network device may update the value of the CSI compression parameter and determine the new first compression parameter based on the updated CSI compression parameter; similarly, the terminal device may also receive the new first compression parameter when Next, new CSI compression parameters are determined according to the new first compression parameters.
  • the value of the CSI compression parameter may be preset by the terminal device, for example, it may be a preset parameter value of the terminal device, or it may be a parameter value set by the terminal device according to user input.
  • the terminal device may determine the second compression parameter according to the value of the preset CSI compression parameter, and send the second compression parameter to the network device, thereby agreeing to use the same CSI compression parameter. For example, the terminal device may send the second compression parameter to the network device through RRC signaling.
  • the CSI compression parameter and the second compression parameter can be determined based on the preset corresponding relationship between the second compression parameter.
  • the terminal device can determine the value of the CSI compression parameter according to its own device parameters.
  • the device parameters can include one or more of the following: protocol version of the terminal device, signal quality of the terminal device, terminal device and network The distance of the device, the amount of uplink data of the terminal device, and the amount of downlink data of the terminal device.
  • the signal quality of the terminal equipment can include RSRP (Reference Signal Receiving Power, reference signal receiving power) or SINR (Signal to Interference plus Noise Ratio, signal to interference plus noise ratio).
  • the terminal device may update the value of the CSI compression parameter and determine the new second compression parameter based on the updated CSI compression parameter. For example, when the terminal device determines that the above device parameters have changed, the terminal device may update the value of the CSI compression parameter according to the new changed device parameters. Similarly, when receiving a new second compression parameter, the network device may determine a new CSI compression parameter based on the new second compression parameter.
  • the CSI compression parameter may include a third compression parameter preset on the terminal device and the network device respectively, and the values of the third compression parameter preset on the terminal device and the network device may be the same.
  • Figure 3 is a flowchart of a method for obtaining channel quality according to an exemplary embodiment. This method can be applied to network devices, as shown in Figure 3. The method can include:
  • the network device receives the target channel matrix sent by the terminal device.
  • the target channel matrix is obtained by the terminal device after compressing the first channel matrix according to the channel state information CSI compression model and the channel state information CSI compression parameters.
  • the first channel matrix is obtained by the terminal device according to the pilot signal for characterization. Matrix of channel quality of the downlink channel.
  • the network device may send a pilot signal through a downlink channel, so that the terminal device receives the pilot signal and obtains the first channel matrix.
  • the pilot signal may include a channel state information reference signal CSI-RS.
  • the network device decompresses the target channel matrix according to the channel state information CSI decompression model and CSI compression parameters to obtain the third channel matrix.
  • the CSI decompression model may include a channel decoder.
  • the channel decoder includes multiple sub-decoders, and different sub-decoders correspond to different CSI compression parameters.
  • the network device may determine one or more target sub-decoders from multiple sub-decoders according to the CSI compression parameters, and use the target sub-decoders to compress the first channel matrix to obtain the target channel matrix.
  • the CSI compression parameter can represent the CSI compression rate
  • the value of the CSI compression rate can be any preset value, for example, it can include: 1/2, 1/4, 1/8, 1/16, 1 /32 or 1/64, etc., this disclosure does not limit this.
  • the network device determines the channel quality of the downlink channel according to the third channel matrix.
  • the target channel matrix is obtained by the terminal device after compressing the first channel matrix according to the channel state information CSI compression model and the channel state information CSI compression parameters.
  • the first channel matrix is obtained by the terminal device according to the pilot signal for characterization.
  • a matrix of channel quality of the downlink channel; the CSI decompression model may include a channel decoder.
  • the channel decoder includes multiple sub-decoders, and different sub-decoders correspond to different CSI compression parameters.
  • the network device can adapt to different CSI compression parameters through multiple sub-decoders, so that in a scenario where the CSI compression rate changes, a more accurate third channel matrix can be adaptively obtained, and based on the third channel matrix A more accurate channel quality is obtained, and the network equipment can determine the modulation and coding strategy corresponding to the downlink channel based on the channel quality, thereby improving data transmission efficiency.
  • the network device may use the sub-decoder corresponding to the CSI compression parameter as the target sub-decoder; decompress the target channel matrix through the target sub-decoder to obtain the fourth channel matrix; determine the fourth channel matrix according to the fourth channel matrix. Three channel matrix.
  • the sub-decoder may include a decompression layer, which may correspond to a compression layer of a sub-encoder in a channel encoder of the terminal device.
  • the CSI decompression model may also include a CSI reconstruction module, and the network device may input the above-mentioned fourth channel matrix into the CSI reconstruction module to obtain a third channel matrix.
  • the CSI reconstruction module may include a convolutional neural network CNN.
  • the value of the CSI compression parameter may be preset by the network device, for example, it may be a preset parameter value of the network device, or it may be a parameter value set by the network device according to user input.
  • the network device may determine the first compression parameter according to the CSI compression parameter, and send the first compression parameter to the terminal device to instruct the terminal device to determine the CSI compression parameter according to the first compression parameter.
  • the network device may send the first compression parameter through RRC (Radio Resource Control) signaling (such as broadcast signaling or proprietary signaling for the terminal device).
  • RRC Radio Resource Control
  • the CSI compression parameter and the first compression parameter may be determined based on a preset corresponding relationship between the first compression parameter.
  • the first compression parameter corresponding to the value 1/2 of the CSI compression parameter may be 1
  • the first compression parameter corresponding to the value 1/4 of the CSI compression parameter may be is 2
  • the first compression parameter corresponding to the value 1/8 of the CSI compression parameter can be 3, and other values can be deduced in the same way.
  • the terminal device after receiving the first compression parameter sent by the network device, the terminal device can determine the CSI compression parameter according to the first compression parameter.
  • the network device can set different CSI compression parameter values for different terminal devices.
  • the value of the CSI compression parameter can be determined according to the device parameters corresponding to the terminal device.
  • the device parameters can include one of the following: One or more items: the protocol version of the terminal device, the signal quality of the terminal device, the distance between the terminal device and the network device, the amount of uplink data of the terminal device, and the amount of downlink data of the terminal device.
  • the signal quality of the terminal equipment can include RSRP (Reference Signal Receiving Power, reference signal receiving power) or SINR (Signal to Interference plus Noise Ratio, signal to interference plus noise ratio).
  • the network device may update the value of the CSI compression parameter and determine the new first compression parameter based on the updated CSI compression parameter. For example, the network device can obtain the device parameters corresponding to the terminal device, and when it is determined that the device parameters have changed, the value of the CSI compression parameter can be updated.
  • the terminal device may determine a new CSI compression parameter based on the new first compression parameter.
  • the CSI compression parameters may be determined by the network device based on parameters received from the terminal device. For example, the network device may receive the second compression parameter sent by the terminal device; and determine the CSI compression parameter according to the second compression parameter. For example, the network device may receive the second compression parameter through RRC signaling.
  • the terminal device may preset the value of the CSI compression parameter, determine the second compression parameter according to the CSI compression parameter, and send the second compression parameter to the network device.
  • the CSI compression parameter and the second compression parameter can be determined based on the preset corresponding relationship between the second compression parameter. In this way, after receiving the second compression parameter sent by the terminal device, the network device can determine the CSI compression parameter according to the second compression parameter.
  • the terminal device may update the value of the CSI compression parameter and determine the new second compression parameter based on the updated CSI compression parameter; similarly, the network device may also receive the new second compression parameter when Next, the new CSI compression parameter is determined according to the new second compression parameter.
  • the CSI compression parameter may also be a third compression parameter preset on the terminal device and the network device respectively, and the values of the third compression parameter preset on the terminal device and the network device may be the same.
  • the above-mentioned CSI compression model and CSI decompression model may together form a network model for obtaining channel quality.
  • the network model for obtaining channel quality will be described in detail below with reference to the accompanying drawings.
  • Figure 4 is a schematic structural diagram of a network model for obtaining channel quality according to an exemplary embodiment.
  • the network model for obtaining channel quality may include a channel state information CSI compression model 41 and a channel state information CSI decompression model 42.
  • the CSI compression model 41 may be deployed as shown in Figure 1 On the terminal equipment in the communication system, for example, the terminal equipment can run the CSI compression model 41 through software, hardware or a combination of software and hardware; the CSI decompression model 42 can be deployed on the network equipment in the communication system shown in Figure 1 ( The CSI decompression model 42 can be run on a network device such as a base station (for example, a base station) through software, hardware, or a combination of software and hardware.
  • a base station for example, a base station
  • the CSI compression model 41 can compress the input first channel matrix according to the CSI compression parameters and then output a target channel matrix (the target channel matrix can also be called a codeword); the terminal device can send the target channel matrix to the network device, and accordingly Specifically, the network device can input the received target channel matrix into the CSI decompression model 42, and the CSI decompression model 42 can decode (may also be called decompression) the target channel matrix according to the CSI compression parameters and output a third channel matrix. , the network device can determine the channel quality of the downlink channel according to the third channel matrix.
  • the CSI compression model 41 may include a channel encoder 411, which may encode the input channel matrix according to the CSI compression parameters to obtain a target channel matrix.
  • the channel encoder 411 may include multiple sub-encoders, and different sub-encoders correspond to different CSI compression parameters.
  • the terminal device can use the sub-encoder corresponding to the CSI compression parameter as the first target sub-encoder; compress the first channel matrix through the first target sub-encoder to obtain the target channel matrix.
  • the target channel matrix can also be called a codeword
  • the channel encoder 411 may include sub-encoder 1 (the corresponding CSI compression parameter is 1/2), sub-encoder 2 (the corresponding CSI compression parameter is 1/2) 1/4), sub-encoder 3 (the corresponding CSI compression parameter is 1/8), sub-encoder 3 (the corresponding CSI compression parameter is 1/16), sub-encoder 4 (the corresponding CSI compression parameter is 1/ 32), sub-encoder 5 (the corresponding CSI compression parameter is 1/64), etc.
  • the CSI compression model may include a channel encoder 411 and a feature converter (which may also be called a feature optimizer or feature joint optimizer) 412 .
  • the input of the feature converter 412 may be the above-mentioned first channel matrix, which is used to extract key features of the first channel matrix to obtain a second channel matrix that represents the key features of the CSI.
  • the second channel matrix may be used as the channel encoder 411
  • the channel encoder 411 can compress the second channel matrix according to the CSI compression parameters to obtain the target channel matrix.
  • the terminal device can use the sub-encoder corresponding to the CSI compression parameter as the second target sub-encoder, and then compress the second channel matrix through the second target sub-encoder to obtain the target channel matrix.
  • the target channel matrix may also be called a codeword.
  • FIG. 5 is a schematic diagram of a feature converter in a CSI compression model according to an exemplary embodiment.
  • the feature converter 412 includes a feature extraction network 4121, an attention mechanism network 4122 and a feature restoration network 4123.
  • the terminal device can input the first channel matrix into the feature converter, extract key features from the first channel matrix, and obtain a second channel matrix that represents the key features of the CSI.
  • the second channel matrix can be obtained through the following steps:
  • the feature extraction network converts the first channel matrix H c into a first feature map where f represents the number of extracted first feature maps.
  • f can be any positive integer greater than 1.
  • the feature extraction network may include a two-dimensional convolution layer, and the size of the convolution kernel may be f ⁇ m ⁇ m, where f represents the number of first feature maps, and m ⁇ m represents the use of the convolution kernel.
  • the feature extraction network can use a two-dimensional normalization layer to normalize the output of the convolution layer.
  • the activation function of the feature extraction network can include Sigmoid, ReLU, LeakyReLU, PReLU or ELU.
  • the activation function can use the LeakyReLU (leaky linear rectification function) activation function, and the activation function can include the following formula (2):
  • x represents the input vector, such as the input first feature map
  • represents the preset coefficient, for example, the preset coefficient can be any value less than 1, for example, the preset coefficient can be 0.3
  • LeakyReLU(x) represents the output value of the Leaky ReLU activation function.
  • the second feature map may include key feature information in the first feature map.
  • the maximum pooling operation can be performed on multiple first feature maps through the attention mechanism network to obtain the maximum pooling feature map; the average pooling operation can be performed on the multiple first feature maps through the attention mechanism network, The average pooling feature map is obtained; then, the second feature map is determined based on the maximum pooling feature map and the average pooling feature map.
  • the feature information contained in some first feature maps is of great help to CSI reconstruction and can be called “key feature maps”.
  • the feature information contained in some first feature maps has little impact on CSI reconstruction and can be called “non-essential feature maps”.
  • the attention mechanism network can extract key feature maps from multiple feature maps, so that the code words generated by the subsequent encoder contain more key features.
  • the attention mechanism network can perform a maximum pooling operation on the first feature map F to obtain the maximum pooled feature map M ⁇
  • the max pooling operation can be performed by the following formula (3):
  • m i represents the i-th element in the maximum pooling feature map M
  • f represents the number of first feature maps
  • F 1,i represents the i-th element in the first first feature map
  • F 2,i represents the i-th element in the second first feature map
  • F 2,i represents the i-th element in the second first feature map
  • F f,i represents the i-th element in the f-th first feature map elements
  • N c ⁇ N t represents the size of the first channel matrix.
  • each element in the maximum pooling feature map M is composed of the largest elements at corresponding positions in multiple first feature maps.
  • the attention mechanism network can also perform an average pooling operation on the first feature map F to obtain an average pooled feature map.
  • the average pooling operation can be performed through the following formula (4):
  • vi represents the i-th element in the average pooled feature map V
  • f represents the number of first feature maps
  • F 1,i represents the i-th element in the first first feature map
  • F 2,i represents the i-th element in the second first feature map
  • F 2,i represents the i-th element in the second first feature map
  • F f-1,i represents the f-1 first feature map
  • the i-th element in, F f,i represents the i-th element in the f-th first feature map
  • N c ⁇ N t represents the size of the first channel matrix.
  • each element in the average pooled feature map V is composed of the mean value of all elements at the corresponding positions of multiple first feature maps.
  • the second feature map can be determined based on the maximum pooling feature map and the average pooling feature map.
  • the maximum pooling feature map and the average pooling feature map can be input into the fusion sub-network to obtain a fused fusion feature map; and the second feature map is calculated based on the fusion feature map and the first feature map.
  • the maximum pooling feature map M and the average pooling feature map V can be spliced to obtain a joint feature map. Pass the joint feature map C through the fusion network to obtain the fused feature map
  • the fusion network may adopt a two-dimensional convolution layer with a convolution kernel size of m ⁇ n ⁇ n.
  • the fusion network may also include a two-dimensional normalization layer and an activation function.
  • the activation function may include Sigmoid activation function.
  • the second feature map F′ may also be called an optimized feature map.
  • the feature information in the key feature map is highlighted in the second feature map F′, while the feature information in the unnecessary feature map is weakened, so that the key features of CSI can be represented.
  • the above-mentioned second feature map F′ can be restored to a second channel matrix through the feature restoration network
  • the feature reduction network may include a two-dimensional convolution layer, a two-dimensional normalization layer and an activation function.
  • the activation function may be a LeakyReLU activation function or other activation functions in related technologies. This disclosure There is no limit to this.
  • the feature reduction network and the feature extraction network may use the same activation function to avoid feature distortion.
  • the second channel matrix He obtained in this way highlights key feature information and weakens unnecessary feature information, thereby representing key CSI features.
  • Figure 6 is a schematic diagram of a channel encoder in a CSI compression model according to an exemplary embodiment.
  • the channel encoder 411 may include multiple sub-encoders, such as sub-encoder 1, sub-encoder 2, ..., sub-encoder T-1, sub-encoder T in the figure. Among them, T is the number of sub-encoders.
  • the channel encoder may first preprocess the input second channel matrix He , for example, it may perform dimension transformation, and the dimension size of the transformed second channel matrix may be The dimensionally transformed second channel matrix is input to the above sub-encoder for processing.
  • the CSI compression parameters may include multiple preset CSI compression rates.
  • the preset CSI compression rate ⁇ 1 corresponding to sub-encoder 1 may be 1/2
  • the preset CSI compression rate ⁇ 2 corresponding to sub-encoder 2 may be 1/4
  • the preset CSI compression rate ⁇ 2 corresponding to sub-encoder 3 may be 1/4.
  • the compression rate ⁇ 3 can be 1/8
  • the preset CSI compression rate ⁇ 4 corresponding to the sub-encoder 4 can be 1/16
  • the preset CSI compression rate ⁇ 5 corresponding to the sub-encoder 5 can be 1/32
  • the preset CSI compression rate ⁇ 6 corresponding to the processor 6 may be 1/64, and so on. It should be noted that the values of the preset CSI compression rate here are examples, and this disclosure does not limit the specific values.
  • the preset CSI compression rate ⁇ 1 corresponding to sub-encoder 1 may be 1/4
  • the preset CSI compression rate ⁇ 2 corresponding to sub-encoder 2 may be 1/16
  • the preset CSI compression rate ⁇ 2 corresponding to sub-encoder 3 may be 1/16.
  • the rate ⁇ 3 may be 1/32, and the preset CSI compression rate ⁇ 4 corresponding to the sub-encoder 4 may be 1/64.
  • each sub-encoder may include a compression layer, which may include a fully connected layer.
  • the size of the compression layer of each sub-encoder can be determined according to CSI compression parameters (such as a preset CSI compression rate) and the size of the first channel matrix. For example, the size of the first channel matrix is c ⁇ N c ⁇ N t , and the size of the first channel matrix is c ⁇ N c ⁇ N t .
  • the CSI compression rate is ⁇ 1
  • the size of the compression layer of the sub-encoder corresponding to ⁇ 1 is (c ⁇ N c ⁇ N t ) ⁇ d 1 , where d 1 can be c ⁇ N c ⁇ N t ⁇ 1 .
  • the size of the compression layer of the sub-encoder corresponding to the preset CSI compression rate may be 2048 ⁇ 512.
  • each sub-encoder may include a compression layer and a coding switch, so that in actual use, the target sub-encoder corresponding to the CSI compression parameter (such as CSI compression rate) can be determined (such as in the above embodiment).
  • first target sub-encoder or second target sub-encoder and close the encoding switch of the target sub-encoder and open the switches of other sub-encoders, so that the compression layer of the target sub-encoder can be used to compress the input
  • the second channel matrix is compressed and the undetermined channel matrix of the target sub-encoder is output, and the undetermined channel matrix can be used as the target channel matrix M.
  • the preset CSI compression rate with the largest value is used as the maximum CSI compression rate
  • the compression layer of the sub-encoder (such as sub-encoder 1 in Figure 5) corresponding to the maximum CSI compression rate is used as the maximum compression layer.
  • the output of the maximum compression layer can be used as the input of the compression layer of other sub-encoders, so that the efficiency of compression operation can be improved.
  • the preset CSI compression rate ⁇ 1 is the maximum compression rate
  • the size of the compression layer 1 of the sub-encoder 1 corresponding to ⁇ 1 can be (c ⁇ N c ⁇ N t ) ⁇ d 1 , where d 1 can be c ⁇ N c ⁇ N t ⁇ 1 .
  • the output of compression layer 1 can be used as the input of the compression layer corresponding to other sub-encoders.
  • the size of the compression layer corresponding to other sub-encoders can be d 1 ⁇ d k
  • d k can be c ⁇ N c ⁇ N t ⁇ ⁇ k .
  • k represents the number of the sub-encoder
  • ⁇ k represents the preset CSI compression rate corresponding to the k-th sub-encoder
  • c ⁇ N c ⁇ N t represents the size of the first channel matrix.
  • the size of the first channel matrix is 2 ⁇ 32 ⁇ 32, and the maximum compression rate ⁇ 1 is 1/2, then the size of compression layer 1 of sub-encoder 1 corresponding to ⁇ 1 can be 2048 ⁇ 1024; sub-encoder 2
  • the corresponding default CSI compression rate ⁇ 2 is 1/4, then the size of the compression layer 2 of the sub-encoder 2 can be 1024 ⁇ 512; the corresponding default CSI compression rate ⁇ 3 of the sub-encoder 3 is 1/16, Then the size of compression layer 3 of sub-encoder 3 can be 1024 ⁇ 256.
  • the compression layers in sub-encoder 2 to sub-encoder T in Figure 6 can further reduce the dimensionality.
  • Figure 7 is a schematic diagram of a CSI decompression model according to an exemplary embodiment.
  • the CSI decompression model 42 may include a channel decoder 421.
  • the channel decoder 421 may include multiple sub-decoders, and different sub-decoders correspond to different CSI compression parameters. Each sub-decoder can decompress the received target channel matrix through the decompression layer to obtain a third channel matrix.
  • the decompression layer may include a fully connected layer.
  • the channel decoder 421 may include multiple sub-decoders, for example, sub-decoder 1, sub-decoder 2, ..., sub-decoder T-1, sub-decoder T. Among them, T is the number of sub-decoders.
  • the CSI compression parameters may include multiple preset CSI compression rates.
  • the preset CSI compression rate ⁇ 1 corresponding to sub-decoder 1 may be 1/2
  • the preset CSI compression rate ⁇ 2 corresponding to sub-decoder 2 may be 1/4
  • the preset CSI compression rate ⁇ 2 corresponding to sub-decoder 3 may be 1/4.
  • the compression rate ⁇ 3 can be 1/8
  • the preset CSI compression rate ⁇ 4 corresponding to the sub-decoder 4 can be 1/16
  • the preset CSI compression rate ⁇ 5 corresponding to the sub-decoder 5 can be 1/32.
  • the preset CSI compression rate ⁇ 6 corresponding to the processor 6 may be 1/64, and so on. It should be noted that the values of the preset CSI compression rate here are examples, and this disclosure does not limit the specific values.
  • each sub-decoder may include a decompression layer and a decoding switch, so that in actual use, the target sub-decoder corresponding to the CSI compression parameter (such as CSI compression rate) can be determined and the target sub-decoder can be decoded
  • the decoding switch of the decoder is closed, and the switches of other sub-decoders are turned off, so that the decompression layer of the target sub-decoder can be used to decompress the input target channel matrix, and the undetermined channel matrix corresponding to the target sub-decoder can be output.
  • the undetermined channel matrix is used as a fourth channel matrix; and a third channel matrix is further determined based on the fourth channel matrix.
  • the CSI decompression model may also include a CSI reconstruction module 422.
  • the CSI reconstruction module may include Convolutional Neural Networks (CNN).
  • CNN Convolutional Neural Networks
  • the CSI reconstruction module includes two CNN networks. Each CNN network includes 5 convolutional layers. Each convolutional layer has The convolution kernel sizes are c ⁇ k ⁇ k, f 1 ⁇ l ⁇ l, f 2 ⁇ l ⁇ l, f 2 ⁇ n ⁇ n, c ⁇ m ⁇ m, (f 1 , f 2 ,k,l, m, n are all preset values, and different values can be preset according to different convolutional layers).
  • the step size of each convolutional layer is t, and the normalization layer and LeakyReLU activation function can be used.
  • the output element value of the second CNN module is mapped to the [0,1] interval through the Sigmoid activation function layer. In this way, the CSI reconstruction module can output the third channel matrix corresponding to the target channel matrix
  • the CSI reconstruction module can output different third channel matrices for different CSI compression parameters (such as CSI compression ratio).
  • the network device can decompress the received target channel matrix to obtain a third channel matrix, so as to determine the channel quality of the downlink channel based on the third channel matrix.
  • the above-mentioned CSI compression model and CSI decompression model can be obtained through offline training.
  • the above-mentioned CSI compression model and CSI decompression model can be jointly trained to obtain the parameters of the CSI compression model and the CSI decompression model, Make the CSI compression model and CSI decompression model match.
  • FIG 8 is a flowchart of a training method for a CSI compression model according to an exemplary embodiment. This training method can be applied to terminal devices. As shown in Figure 8, the training method can include:
  • the terminal device obtains the first sample channel matrix used for training.
  • the first sample channel matrix is a matrix obtained by the terminal device according to the received pilot signal and used to characterize the downlink channel quality.
  • the training set can be used as the above-mentioned first sample channel matrix.
  • the terminal device trains the first target network model according to the first sample channel matrix to obtain a CSI compression model.
  • the first target network model includes a first target compression model and a first target decompression model.
  • the network structure of the first target compression model is the same as that of the CSI compression model.
  • both may include a channel encoder. It includes multiple sub-encoders, and different sub-encoders correspond to different CSI compression parameters;
  • the first target decompression model includes a channel decoder, and the channel decoder includes multiple sub-decoders, and different sub-decoders correspond to different CSI compression parameters.
  • the model structure of the above-mentioned first target compression model may be the same as the CSI compression model shown in Figure 4.
  • the first target compression model may include a channel encoder and a feature converter.
  • the structure can be shown in Figure 5, and the structure of the channel encoder can be shown in Figure 6;
  • the model structure of the above-mentioned first target decompression model can be the same as the CSI decompression model shown in Figure 7, here the above model structure No further details will be given.
  • the first model training step can be performed cyclically until it is determined according to the first sample channel matrix and the first prediction channel matrix that the trained first target network model satisfies the first preset stop iteration condition, and the trained first target network model will be The first target compression model in the first target network model is used as the CSI compression model.
  • the first prediction channel matrix is a matrix output after the first sample channel matrix is input into the first target network model.
  • the first compression model parameters corresponding to the first target compression model may be used as model parameters of the CSI compression model.
  • the first model training step may include:
  • S81 Input the first sample channel matrix into the first target compression model, and compress the first sample channel matrix through multiple sub-encoders to obtain the first target sample channel matrix.
  • first preset stop iteration condition can be determined according to the loss function used in the training process.
  • the encoding switches corresponding to all sub-encoders can be closed, the decoding switches corresponding to all sub-decoders can also be closed, and joint optimization can be performed for the training errors under multiple CSI compression parameters.
  • the loss function can include the following formula (5):
  • ⁇ 1 represents the first preset CSI compression rate
  • ⁇ 2 represents the first preset CSI compression rate
  • ⁇ T represents the T-th preset CSI compression rate
  • T represents the number of preset CSI compression rates in the CSI compression parameters (at the same time It can also represent the number of sub-encoders, or the number of sub-decoders), represents the loss value corresponding to the preset CSI compression rate ⁇ 1
  • Loss represents the first loss value of the above-mentioned first sample channel matrix and the first prediction channel matrix.
  • the network can optimize the training errors corresponding to all preset CSI compression rates in one training process, thereby improving the network Adaptability to dynamic changes in CSI compression rate.
  • the parameters of the CSI compression model can be obtained.
  • the terminal device can also obtain the trained parameters of the first target decompression model through the above training process.
  • the trained parameters of the first target decompression model can be used as parameters of the CSI decompression model on the network device side. .
  • the terminal device after the terminal device determines the model parameters through the training method, it can obtain the first decompression model parameters corresponding to the first target decompression model in the trained first target network model; and obtain the first decompression model parameters.
  • the model parameters are sent to the network device to instruct the network device to determine the CSI decompression model according to the first decompression model parameters.
  • the CSI decompression model is used by the network device to determine the channel quality of the downlink channel according to the target channel matrix.
  • the terminal device may send the first decompression model parameters to the network device through signaling or data packets.
  • the above training method can be executed on the network device, and the terminal device can receive the second compression model parameters sent by the network device; and determine the CSI compression model according to the second compression model parameters.
  • the second compression model parameter may be used as a model parameter of the CSI compression model.
  • the terminal device may receive the second compression model parameter sent by the network device through signaling or a data message, and determine the parameters corresponding to the CSI compression model based on the second compression model parameter.
  • FIG 9 is a flowchart of a training method for a CSI decompression model according to an exemplary embodiment.
  • the training method can be applied to network devices.
  • the training method can include:
  • the network device obtains the second sample channel matrix used for training.
  • the second sample channel matrix may be a matrix obtained by the terminal device according to the received pilot signal and used to characterize the downlink channel quality.
  • the network device eg, base station
  • a single antenna can be configured on the terminal device.
  • 150,000 airspace CSI matrix samples were generated in the 5.3GHz indoor microcell scenario, and obtained as a training set containing 100,000 samples, a verification set containing 30,000 samples, and a test set containing 20,000 samples.
  • the training set therein can be used as the first sample channel matrix.
  • the network device trains the second target network model according to the second sample channel matrix to obtain the CSI decompression model.
  • the second target network model includes a second target compression model and a second target decompression model.
  • the network structure of the second target decompression model is the same as that of the CSI decompression model.
  • both include a channel decoder, and the channel
  • the decoder may include multiple sub-decoders, and different sub-decoders correspond to different CSI compression parameters;
  • the second target compression model includes a channel encoder, and the channel encoder includes multiple sub-encoders, and different sub-encoders correspond to different CSI compression parameters. parameter.
  • the model structure of the above-mentioned second target compression model may be the same as the CSI compression model shown in Figure 4.
  • the second target compression model may include a channel encoder and a feature converter.
  • the structure can be shown in Figure 5, and the structure of the channel encoder can be shown in Figure 6;
  • the model structure of the above-mentioned second target decompression model can be the same as the CSI decompression model shown in Figure 7, here the above model structure No further details will be given.
  • the second model training step can be performed cyclically until it is determined according to the second sample channel matrix and the second prediction channel matrix that the trained second target network model satisfies the second preset stop iteration condition, and the trained second target network model is The second target decompression model in the second target network model serves as the CSI decompression model; the second prediction channel matrix is a matrix output after the second sample channel matrix is input into the second target network model.
  • the second model training step may include:
  • the above-mentioned second preset stop iteration condition can also be determined based on the loss function used in the training process.
  • the encoding switches corresponding to all sub-encoders can be closed, and the decoding switches corresponding to all sub-decoders can also be closed, and joint optimization can be performed for the training errors under multiple CSI compression parameters.
  • the loss function can also include the above formula (5), which will not be described again here.
  • the network can optimize the training errors corresponding to all preset CSI compression rates in one training process, thereby improving the network Adaptability to dynamic changes in CSI compression rate.
  • the parameters of the CSI decompression model can be obtained.
  • the network device can also obtain the trained parameters of the second target compression model through the above training process, and the trained parameters of the second target compression model can be used as parameters of the CSI compression model on the terminal device side.
  • the network device may obtain the second compression model parameters corresponding to the second target compression model in the trained second target network model; and send the second compression model parameters to the terminal device to instruct the terminal device according to The second compression model parameters determine the CSI compression model.
  • the CSI compression model is used by the terminal device to obtain the target channel matrix according to the first channel matrix.
  • the network device may send the second compression model parameters to the terminal device through signaling or data packets.
  • the above training method can be executed on the terminal device, and the network device can receive the first decompression model parameter sent by the terminal device; and determine the CSI decompression model according to the first decompression model parameter.
  • the network device may receive the first decompression model parameter sent by the terminal device through signaling or a data message, and determine the parameters corresponding to the CSI decompression model based on the first decompression model parameter.
  • Figure 10 is a method for obtaining channel quality according to an exemplary embodiment. As shown in Figure 10, the method may include:
  • the network device sends a pilot signal through the downlink channel.
  • the terminal device receives the pilot signal through the downlink channel, and obtains the first channel matrix according to the pilot signal.
  • the first channel matrix is used to characterize the channel quality of the downlink channel
  • the terminal device compresses the first channel matrix according to the channel state information CSI compression model and the channel state information CSI compression parameters to obtain a compressed target channel matrix.
  • the CSI compression model includes a channel encoder, and the channel encoder includes multiple sub-encoders; different sub-encoders correspond to different CSI compression parameters.
  • the CSI compression model may include a channel encoder and a feature transformer.
  • the terminal device sends the target channel matrix to the network device.
  • the network device receives the target channel matrix sent by the terminal device, and decompresses the target channel matrix according to the channel state information CSI decompression model and CSI compression parameters to obtain the third channel matrix.
  • the CSI decompression model includes a channel decoder, and the channel decoder includes multiple sub-decoders, and different sub-decoders correspond to different CSI compression parameters;
  • the network device determines the channel quality of the downlink channel according to the third channel matrix.
  • the above-mentioned CSI compression model and the above-mentioned CSI decompression model can be any model structure provided in the above embodiments, and will not be described again here.
  • multiple sub-encoders and multiple sub-decoders can adapt to different CSI compression parameters, so that in scenarios where the CSI compression rate changes, the terminal device can adaptively obtain a more accurate target channel matrix and send it to the network equipment so that network equipment can obtain more accurate channel quality, thereby improving data transmission efficiency.
  • Figure 11 is a block diagram of a device 1100 for obtaining channel quality according to an exemplary embodiment.
  • the device can be applied to terminal equipment.
  • the device 1100 may include:
  • the first receiving module 1101 is configured to receive the pilot signal sent by the network device through the downlink channel;
  • the first matrix acquisition module 1102 is configured to acquire a first channel matrix according to the pilot signal; the first channel matrix is used to characterize the channel quality of the downlink channel;
  • the target matrix acquisition module 1103 is configured to compress the first channel matrix according to the CSI compression model and the channel state information CSI compression parameters to obtain a compressed target channel matrix; wherein the CSI compression model includes a channel encoder , the channel encoder includes multiple sub-encoders; different sub-encoders correspond to different CSI compression parameters;
  • the first sending module 1104 is configured to send the target channel matrix to the network device, so that the network device determines the channel quality of the downlink channel according to the target channel matrix.
  • the target matrix acquisition module 1103 is configured to use the sub-encoder corresponding to the CSI compression parameter as the first target sub-encoder; use the first target sub-encoder to obtain the first target sub-encoder.
  • the channel matrix is compressed to obtain the target channel matrix.
  • the CSI compression model further includes a feature converter; the target matrix acquisition module 1103 is configured to input the first channel matrix into the feature converter, and perform Key features are extracted to obtain a second channel matrix that represents key features of the CSI; and the second channel matrix is compressed according to the CSI compression parameters and the channel encoder to obtain the target channel matrix.
  • the target matrix acquisition module 1103 is configured to use the sub-encoder corresponding to the CSI compression parameter as the second target sub-encoder; use the second target sub-encoder to obtain the second target sub-encoder.
  • the channel matrix is compressed to obtain the target channel matrix.
  • the feature converter includes a feature extraction network, an attention mechanism network and a feature reduction network;
  • the target matrix acquisition module 1103 is configured to input the first channel matrix into the feature extraction network, Obtain a plurality of first feature maps; input the plurality of first feature maps into the attention mechanism network to obtain a second feature map; the second feature map includes key feature information in the first feature map;
  • the second feature map is input into the feature restoration network to obtain a second channel matrix.
  • the target matrix acquisition module 1103 is configured to perform a maximum pooling operation on a plurality of the first feature maps through the attention mechanism network to obtain a maximum pooling feature map; through the attention mechanism network The force mechanism network performs an average pooling operation on a plurality of the first feature maps to obtain an average pooled feature map; determines a second feature map based on the maximum pooled feature map and the average pooled feature map;
  • the attention mechanism network includes a fusion sub-network; the target matrix acquisition module 1103 is configured to input the maximum pooling feature map and the average pooling feature map into the fusion sub-network , obtain the fused feature map after fusion; calculate the second feature map according to the fused feature map and the first feature map.
  • the first matrix acquisition module 1102 is configured to measure a spatial channel matrix according to the pilot signal; and transform the spatial channel matrix into an angular delay domain through discrete Fourier transform Channel matrix; determine the first channel matrix according to the angular delay domain channel matrix.
  • Fig. 12 is a block diagram of a device 1100 for obtaining channel quality according to an exemplary embodiment. As shown in Fig. 12, the device 1100 may also include a first training module 1105. The first training module 1105 is configured as The CSI compression model is trained in the following way:
  • the first sample channel matrix is a matrix obtained by the terminal device according to the received pilot signal and used to characterize the downlink channel quality;
  • the first target network model includes a first target compression model and a first target decompression model.
  • the network structure of the first target compression model is the same as the network structure of the CSI compression model.
  • the first target decompression model includes a channel decoder, and the channel decoder includes multiple sub-decoders, and different sub-decoders correspond to different CSI compression parameters.
  • the first training module 1105 is configured to perform the first model training step cyclically until the trained first target network model is determined according to the first sample channel matrix and the first prediction channel matrix.
  • the first target compression model in the trained first target network model is used as the CSI compression model;
  • the first prediction channel matrix is the input of the first sample channel matrix. The matrix output after describing the first target network model;
  • the first model training step includes:
  • the first target network model When it is determined according to the first sample channel matrix and the first prediction channel matrix that the first target network model does not satisfy the first preset stop iteration condition, according to the first sample channel matrix and the first prediction channel matrix to determine a first loss value, update the parameters of the first target network model according to the first loss value, obtain the trained first target network model, and use the trained first The target network model serves as the new first target network model.
  • the first sending module 1104 is further configured to obtain the first decompression model parameters corresponding to the first target decompression model in the trained first target network model; Compression model parameters are sent to the network device to instruct the network device to determine a CSI decompression model according to the first decompression model parameter, and the CSI decompression model is used for the network device to determine according to the target channel matrix The channel quality of the downlink channel.
  • the first receiving module 1101 is further configured to receive a second compression model parameter sent by the network device; and determine a CSI compression model according to the second compression model parameter.
  • the first receiving module 1101 is further configured to receive a first compression parameter sent by a network device; and determine the CSI compression parameter according to the first compression parameter.
  • Figure 13 is a block diagram of a device 1300 for obtaining channel quality according to an exemplary embodiment.
  • the device can be applied to network equipment.
  • the device 1300 may include:
  • the second receiving module 1301 is configured to receive a target channel matrix sent by the terminal device; the target channel matrix is obtained by compressing the first channel matrix according to the CSI compression model and channel state information CSI compression parameters by the terminal device. , the first channel matrix is a matrix obtained by the terminal device according to the pilot signal and used to characterize the channel quality of the downlink channel;
  • the third matrix acquisition module 1302 is configured to decompress the target channel matrix according to the CSI decompression model and the CSI compression parameters to obtain a third channel matrix; wherein the CSI decompression model includes a channel decoder , the channel decoder includes multiple sub-decoders, and different sub-decoders correspond to different CSI compression parameters;
  • the channel quality determination module 1303 is configured to determine the channel quality of the downlink channel according to the third channel matrix.
  • the third matrix acquisition module 1302 is configured to use the sub-decoder corresponding to the CSI compression parameter as a target sub-decoder; solve the target channel matrix through the target sub-decoder. Compress to obtain a fourth channel matrix; determine the third channel matrix according to the fourth channel matrix.
  • the CSI decompression model further includes a CSI reconstruction module; the third matrix acquisition module 1302 is configured to input the fourth channel matrix into the CSI reconstruction module to obtain the third Three channel matrix.
  • Figure 14 is a block diagram of a device 1300 for obtaining channel quality according to an exemplary embodiment.
  • the device 1300 can also include a second training module 1304, and the second training module 1304 is configured to
  • the CSI decompression model is trained in the following way:
  • the second sample channel matrix is a matrix obtained by the terminal device according to the received pilot signal and used to characterize the downlink channel quality;
  • the second target network model includes a second target compression model and a second target decompression model
  • the network structure of the second target decompression model is the same as the network structure of the CSI decompression model
  • the second target decompression model The target compression model includes a channel encoder, and the channel encoder includes multiple sub-encoders, and different sub-encoders correspond to different CSI compression parameters.
  • the second training module 1304 is configured to perform the second model training step cyclically until it is determined according to the second sample channel matrix and the second prediction channel matrix that the trained second target network model satisfies
  • the second preset stop iteration condition is to use the second target decompression model in the trained second target network model as the CSI decompression model;
  • the second prediction channel matrix is the input of the second sample channel matrix. The matrix output after describing the second target network model;
  • the second model training step includes:
  • the second prediction channel matrix determines a second loss value, updates the parameters of the second target network model according to the second loss value, obtains the trained second target network model, and uses the trained second target network model as a new second target network model.
  • Figure 15 is a block diagram of a device 1300 for obtaining channel quality according to an exemplary embodiment. As shown in Figure 15, the device 1300 may also include:
  • the second sending module 1305 is configured to obtain the second compression model parameters corresponding to the second target compression model in the trained second target network model; send the second compression model parameters to the terminal device in order to indicate
  • the terminal device determines a CSI compression model according to the second compression model parameter, and the CSI compression model is used by the terminal device to obtain a target channel matrix according to the first channel matrix.
  • the second receiving module 1301 is further configured to receive the first decompression model parameter sent by the terminal device; and determine the CSI decompression model according to the first decompression model parameter.
  • the second sending module 1305 is configured to determine a first compression parameter according to the CSI compression parameter; and send the first compression parameter to the terminal device.
  • Figure 16 is a block diagram of an apparatus for obtaining channel quality according to an exemplary embodiment.
  • the device 2000 for obtaining channel quality may be a terminal device in the communication system shown in Figure 1, or may be a network device in the communication system.
  • the apparatus 2000 may include one or more of the following components: a processing component 2002, a memory 2004, and a communications component 2006.
  • Processing component 2002 may control overall operations of device 2000, such as operations associated with display, phone calls, data communications, camera operations, and recording operations.
  • the processing component 2002 may include one or more processors 2020 to execute instructions to complete all or part of the steps of the above method of obtaining channel quality.
  • processing component 2002 may include one or more modules that facilitate interaction between processing component 2002 and other components.
  • processing component 2002 may include a multimedia module to facilitate interaction between the multimedia component and processing component 2002.
  • Memory 2004 is configured to store various types of data to support operations at device 2000. Examples of such data include instructions for any application or method operating on device 2000, contact data, phonebook data, messages, pictures, videos, etc.
  • Memory 2004 may be implemented by any type of volatile or non-volatile storage device, or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EEPROM), Programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic memory, flash memory, magnetic or optical disk.
  • SRAM static random access memory
  • EEPROM electrically erasable programmable read-only memory
  • EEPROM erasable programmable read-only memory
  • EPROM Programmable read-only memory
  • PROM programmable read-only memory
  • ROM read-only memory
  • magnetic memory flash memory, magnetic or optical disk.
  • Communication component 2006 is configured to facilitate wired or wireless communication between apparatus 2000 and other devices.
  • the device 2000 can access a wireless network based on a communication standard, such as WiFi, communication technologies such as 2G, 3G, 4G, 5G or 6G, or a combination thereof.
  • the communication component 2006 receives broadcast signals or broadcast related information from an external broadcast management system via a broadcast channel.
  • the communications component 2006 also includes a near field communications (NFC) module to facilitate short-range communications.
  • NFC near field communications
  • the NFC module can be implemented based on radio frequency identification (RFID) technology, infrared data association (IrDA) technology, ultra-wideband (UWB) technology, Bluetooth (BT) technology and other technologies.
  • RFID radio frequency identification
  • IrDA infrared data association
  • UWB ultra-wideband
  • Bluetooth Bluetooth
  • apparatus 2000 may be configured by one or more application specific integrated circuits (ASICs), digital signal processors (DSPs), digital signal processing devices (DSPDs), programmable logic devices (PLDs), field programmable Gate array (FPGA), controller, microcontroller, microprocessor or other electronic components are implemented for performing the above method of obtaining channel quality.
  • ASICs application specific integrated circuits
  • DSPs digital signal processors
  • DSPDs digital signal processing devices
  • PLDs programmable logic devices
  • FPGA field programmable Gate array
  • controller microcontroller, microprocessor or other electronic components are implemented for performing the above method of obtaining channel quality.
  • the above-mentioned device 2000 can be an independent electronic device or a part of an independent electronic device.
  • the electronic device can be an integrated circuit (Integrated Circuit, IC) or a chip, where the integrated circuit can be an IC can also be a collection of multiple ICs; the chip can include but is not limited to the following types: GPU (Graphics Processing Unit, graphics processor), CPU (Central Processing Unit, central processing unit), FPGA (Field Programmable Gate Array, Programmable logic array), DSP (Digital Signal Processor, digital signal processor), ASIC (Application Specific Integrated Circuit, application specific integrated circuit), SOC (System on Chip, SoC, system on a chip or system-level chip), etc.
  • GPU Graphics Processing Unit, graphics processor
  • CPU Central Processing Unit, central processing unit
  • FPGA Field Programmable Gate Array, Programmable logic array
  • DSP Digital Signal Processor, digital signal processor
  • ASIC Application Specific Integrated Circuit
  • SOC System on Chip, SoC, system on a chip or system-level chip
  • the above integrated circuit or chip can be used to execute executable instructions (or codes) to implement the above method of obtaining channel quality.
  • the executable instructions may be stored in the integrated circuit or chip, or may be obtained from other devices or equipment.
  • the integrated circuit or chip may include a processor, a memory, and an interface for communicating with other devices.
  • the executable instructions can be stored in the processor, and when the executable instructions are executed by the processor, the above-mentioned method of obtaining channel quality is implemented; or, the integrated circuit or chip can receive the executable instructions through the interface and transmit them to the processor.
  • the processor is executed to implement the above method of obtaining channel quality.
  • the present disclosure also provides a computer-readable storage medium on which computer program instructions are stored.
  • the computer-readable storage medium may be a non-transitory computer-readable storage medium including instructions, for example, may be the above-mentioned memory 2004 including instructions, and the above-mentioned instructions may be executed by the processor 2020 of the device 2000 to complete the above-mentioned acquisition.
  • the non-transitory computer-readable storage medium may be ROM, random access memory (RAM), CD-ROM, magnetic tape, floppy disk, optical data storage device, etc.
  • a computer program product comprising a computer program executable by a programmable device, the computer program having a function for performing the above when executed by the programmable device.
  • the code part of the method to obtain the channel quality.

Landscapes

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

Abstract

本公开涉及一种获取信道质量的方法、装置、存储介质和芯片。该方法包括:终端设备接收网络设备通过下行信道发送的导频信号;根据导频信号获取第一信道矩阵;根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对第一信道矩阵进行压缩,得到压缩后的目标信道矩阵;将目标信道矩阵发送至网络设备,以便网络设备根据目标信道矩阵确定下行信道的信道质量。其中,第一信道矩阵用于表征下行信道的信道质量;CSI压缩模型包括信道编码器,信道编码器包括多个子编码器;不同的子编码器对应不同的CSI压缩参数;目标信道矩阵用于指示网络设备确定下行信道的信道质量。

Description

获取信道质量的方法、装置、存储介质和芯片 技术领域
本公开涉及通信技术领域,具体地,涉及一种获取信道质量的方法、装置、存储介质和芯片。
背景技术
作为5G(the 5th GenerationMobile Communication Technology,第五代移动通信系统)的一项关键性技术,近年来mMIMO(massive Multiple-Input Multiple-Output,大规模多输入多输出)技术已经成为通信领域广泛研究和使用的技术。通过在发射端采用集中式或分布式的方法部署大量的天线,mMIMO系统在系统稳定性,能量利用率,抗干扰能力方面都有着良好的表现。为了能够充分利用mMIMO系统的优势,在发射端需要获得准确的CSI(ChannelStateInformation,信道状态信息)。例如,终端设备可以向网络设备上报CSI信息,以便网络设备获取下行信道的信道质量,并根据信道质量为下行传输选择合适的调制和编码方案,提高数据传输的性能。但是,在mMIMO中,天线数持续增加,CSI所蕴含的信息越来越丰富,使得CSI上报的资源开销越来越大。
在相关技术中,为了减少CSI上报开销,可以采用基于离散傅里叶变换(DiscreteFouriertransform,DFT)的CSI压缩技术,终端设备进行CSI压缩后再上报网络设备。但是,CSI压缩上报会降低网络设备获取的信道质量的准确性,特别是不同场景下的CSI压缩后的准确性差异较大,会影响数据传输效率。
发明内容
为克服相关技术中存在的上述问题,本公开提供一种获取信道质量的方法、装置、存储介质和芯片。
根据本公开实施例的第一方面,提供一种获取信道质量的方法,应用于终端设备,所述方法包括:
接收网络设备通过下行信道发送的导频信号;
根据所述导频信号获取第一信道矩阵;所述第一信道矩阵用于表征所述下行信道的信道质量;
根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对所述第一信道矩阵进行压缩,得到压缩后的目标信道矩阵;其中,所述CSI压缩模型包括信道编码器,所述信道编码器包括多个子编码器;不同的子编码器对应不同的CSI压缩参数;
将所述目标信道矩阵发送至所述网络设备,以便所述网络设备根据所述目标信道矩阵获取所述下行信道的信道质量。
根据本公开实施例的第二方面,提供一种获取信道质量的方法,应用于网络设备,所述方法包括:
接收终端设备发送的目标信道矩阵;所述目标信道矩阵为所述终端设备根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对第一信道矩阵进行压缩后得到的,所述第一信道矩阵为所述终端设备根据导频信号获取的用于表征下行信道的信道质量的矩阵;
根据信道状态信息CSI解压缩模型和所述CSI压缩参数,对所述目标信道矩阵进行解压缩,得到第三信道矩阵;其中,所述CSI解压缩模型包括信道解码器,所述信道解码器 包括多个子解码器,不同的子解码器对应不同的CSI压缩参数;
根据所述第三信道矩阵确定下行信道的信道质量。
根据本公开实施例的第三方面,提供一种获取信道质量的装置,应用于终端设备,所述装置包括:
第一接收模块,被配置为接收网络设备通过下行信道发送的导频信号;
第一矩阵获取模块,被配置为根据所述导频信号获取第一信道矩阵;所述第一信道矩阵用于表征所述下行信道的信道质量;
目标矩阵获取模块,被配置为根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对所述第一信道矩阵进行压缩,得到压缩后的目标信道矩阵;其中,所述CSI压缩模型包括信道编码器,所述信道编码器包括多个子编码器;不同的子编码器对应不同的CSI压缩参数;
第一发送模块,被配置为将所述目标信道矩阵发送至所述网络设备,以便所述网络设备根据所述目标信道矩阵获取所述下行信道的信道质量。
根据本公开实施例的第四方面,提供一种获取信道质量的装置,应用于网络设备,所述装置包括:
第二接收模块,被配置为接收终端设备发送的目标信道矩阵;所述目标信道矩阵为所述终端设备根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对第一信道矩阵进行压缩后得到的,所述第一信道矩阵为所述终端设备根据导频信号获取的用于表征下行信道的信道质量的矩阵;
第三矩阵获取模块,被配置为根据信道状态信息CSI解压缩模型和所述CSI压缩参数,对所述目标信道矩阵进行解压缩,得到第三信道矩阵;其中,所述CSI解压缩模型包括信道解码器,所述信道解码器包括多个子解码器,不同的子解码器对应不同的CSI压缩参数;
信道质量确定模块,被配置为根据所述第三信道矩阵确定下行信道的信道质量。
根据本公开实施例的第五方面,提供一种获取信道质量的装置,包括:
处理器;
用于存储处理器可执行指令的存储器;
其中,所述处理器被配置为执行本公开第一方面所提供的获取信道质量的方法的步骤。
根据本公开实施例的第六方面,提供一种获取信道质量的装置,包括:
处理器;
用于存储处理器可执行指令的存储器;
其中,所述处理器被配置为执行本公开第二方面所提供的获取信道质量的方法的步骤。
根据本公开实施例的第七方面,提供一种计算机可读存储介质,其上存储有计算机程序指令,该计算机程序指令被处理器执行时实现本公开第一方面所提供的获取信道质量的方法的步骤。
根据本公开实施例的第八方面,提供一种计算机可读存储介质,其上存储有计算机程序指令,该计算机程序指令被处理器执行时实现本公开第二方面所提供的获取信道质量的方法的步骤。
根据本公开实施例的第九方面,提供一种芯片,包括:处理器和接口;所述处理器用于读取指令以执行本公开第一方面所提供的获取信道质量的方法的步骤。
根据本公开实施例的第十方面,提供一种芯片,包括:处理器和接口;所述处理器用于读取指令以执行本公开第二方面所提供的获取信道质量的方法的步骤。
本公开的实施例提供的技术方案可以包括以下有益效果:终端设备接收网络设备通过下行信道发送的导频信号;根据导频信号获取第一信道矩阵;根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对第一信道矩阵进行压缩,得到压缩后的目标信道矩 阵;将目标信道矩阵发送至网络设备,以便网络设备根据目标信道矩阵确定下行信道的信道质量。其中,第一信道矩阵用于表征下行信道的信道质量;CSI压缩模型可以包括信道编码器,信道编码器包括多个子编码器;不同的子编码器对应不同的CSI压缩参数;目标信道矩阵用于指示网络设备确定下行信道的信道质量。这样,通过多个子编码器可以适配不同的CSI压缩参数,从而可以在CSI压缩率发生变化的场景下,可以自适应地得到较为准确的目标信道矩阵并发送至网络设备,以便网络设备得到较为准确的信道质量,从而提高数据传输效率。
应当理解的是,以上的一般描述和后文的细节描述仅是示例性和解释性的,并不能限制本公开。
附图说明
此处的附图被并入说明书中并构成本说明书的一部分,示出了符合本公开的实施例,并与说明书一起用于解释本公开的原理。
图1是根据一示例性实施例示出的一种通信系统的框图。
图2是根据一示例性实施例示出的一种获取信道质量的方法的流程图。
图3是根据一示例性实施例示出的一种获取信道质量的方法的流程图。
图4是根据一示例性实施例示出的一种获取信道质量的网络模型的结构示意图。
图5是根据一示例性实施例示出的一种CSI压缩模型中的特征转换器的示意图。
图6是根据一示例性实施例示出的一种CSI压缩模型中的信道编码器的示意图。
图7是根据一示例性实施例示出的一种CSI解压缩模型的示意图。
图8是根据一示例性实施例示出的一种CSI压缩模型的训练方法的流程图。
图9是根据一示例性实施例示出的一种CSI解压缩模型的训练方法的流程图。
图10是根据一示例性实施例示出的一种获取信道质量的方法的流程图。
图11是根据一示例性实施例示出的一种获取信道质量的装置的框图。
图12是根据一示例性实施例示出的一种获取信道质量的装置的框图。
图13是根据一示例性实施例示出的一种获取信道质量的装置的框图。
图14是根据一示例性实施例示出的一种获取信道质量的装置的框图。
图15是根据一示例性实施例示出的一种获取信道质量的装置的框图。
图16是根据一示例性实施例示出的一种获取信道质量的装置的框图。
具体实施方式
这里将详细地对示例性实施例进行说明,其示例表示在附图中。下面的描述涉及附图时,除非另有表示,不同附图中的相同数字表示相同或相似的要素。以下示例性实施例中所描述的实施方式并不代表与本公开相一致的所有实施方式。相反,它们仅是与如所附权利要求书中所详述的、本公开的一些方面相一致的装置和方法的例子。
需要说明的是,本公开中所有获取信号、信息或数据的动作都是在遵照所在地国家相应的数据保护法规政策的前提下,并获得由相应装置所有者给予授权的情况下进行的。
在本公开中,使用的术语如“第一”、“第二”等是用于区别类似的对象,而不必理解为特定的顺序或先后次序。另外,在未作相反说明的情况下,在参考附图的描述中,不同附图中的同一标记表示相同的要素。
在本公开的描述中,除非另有说明,“多个”是指两个或多于两个,其它量词与之类似;“以下至少一项(个)”或其类似表达,是指的这些项中的任意组合,包括单项(个)或复数项(个)的任意组合。例如,a,b,或c中的至少一项(个),可以表示:a,b,c, a-b,a-c,b-c,或a-b-c,其中a,b,c可以是单个,也可以是多个;“和/或”是一种描述关联对象的关联关系,表示可以存在三种关系,例如,A和/或B,可以表示:单独存在A,同时存在A和B,单独存在B这三种情况,其中A,B可以是单数或者复数。
在本公开实施例中尽管在附图中以特定的顺序描述操作,但是不应将其理解为要求按照所示的特定顺序或是串行顺序来执行这些操作,或是要求执行全部所示的操作以得到期望的结果。在特定环境中,多任务和并行处理可能是有利的。
随着mMIMO技术的应用,为了减少CSI上报开销,可以采用基于离散傅里叶变换(DiscreteFouriertransform,DFT)的CSI压缩技术,终端设备进行CSI压缩后再上报网络设备。但是,CSI压缩上报会降低网络设备获取的信道质量的准确性,影响数据传输效率。
示例地,终端设备可以采用预先训练的编码神经网络对CSI进行压缩,向网络设备上报压缩后的CSI,网络设备同样使用解码神经网络对压缩后的CSI进行解压缩,以确定信道质量。但是,由于不同的CSI压缩率,对编码神经网络和解码神经网络的要求不同,参数也不同,因此,在CSI压缩率发生变化的场景下,若使用同样的编码神经网络和解码神经网络,会导致CSI上报的准确度差异较大。而当CSI压缩率发生变化时,若对编码神经网络和解码神经网络重新进行训练,则效率低下,无法适应CSI压缩率经常变化的场景。
为了解决上述问题,本公开提供了一种获取信道质量的方法、装置、存储介质和芯片。
下面首先介绍本公开实施例的实施环境。
图1是根据一示例性实施例示出的一种通信系统的示意图,如图1所示,该通信系统可以包括终端设备101和网络设备102。该通信系统可以用于支持4G(the 4th Generation,第四代)网络接入技术,例如长期演进(Long Term Evolution,LTE)接入技术,或者,5G(the 5th Generation,第五代)网络接入技术,如新型无线入技术(New Radio Access Technology,New RAT),或者,其他未来的无线通信技术。需要说明的是,该通信系统可以是采用FDD(Frequency Division Duplexing,频分双工)技术的通信系统,也可以是采用TDD(Time Division Duplexing,时分双工)技术的通信系统。另外,在该通信系统中,网络设备与终端设备的数量均可以为一个或多个,图1所示通信系统的网络设备与终端设备的数量仅为适应性举例,本公开对此不做限定。
图1中的网络设备可用于支持终端接入,例如,可以是LTE中的演进型基站(evolutional Node B,eNB或eNodeB);或者5G网络或者未来演进的公共陆地移动网络(public land mobile network,PLMN)中的基站,宽带网络业务网关(Broadband Network Gateway,BNG),汇聚交换机或非3GPP(3rd Generation Partnership Project,第三代合作伙伴项目)接入设备等。可选的,本公开实施例中的网络设备可以包括各种形式的基站,例如:宏基站、微基站(也称为小站)、中继站、接入点、5G基站或未来的基站、卫星、传输点(transmitting and receiving point,TRP)、发射点(transmitting point,TP)、移动交换中心以及设备到设备(Device-to-Device,D2D)、车辆外联(Vehicle-to-Everything,V2X)、机器到机器(machine-to-machine,M2M)通信中承担基站功能的设备等,本公开实施例对此不作具体限定。为方便描述,本公开所有实施例中,为终端设备提供无线通信功能的装置统称为网络设备或基站。
图1中的终端设备可以是一种提供语音或者数据连通性的电子设备,例如也可以称为用户设备(User Equipment,UE),用户单元(SubscriberUnit),移动台(Mobile Station),站台(Station),终端(Terminal)等。示例地,该终端设备可以包括智能手机、智能可穿戴设备、智能音箱、智能平板、无线调制解调器(modem)、无线本地环路(Wireless Local Loop,WLL)台、PDA(Personal Digital Assistant,个人数字助理)、CPE(Customer Premise Equipment,客户终端设备)等。随着无线通信技术的发展,可以接入通信系统、可以与通信系统的网络设备进行通信,或者通过通信系统与其它物体进行通信的设备都可以是本公 开实施例中的终端设备,例如,智能交通中的终端和汽车、智能家居中的家用设备、智能电网中的电力抄表仪器、电压监测仪器、环境监测仪器、智能安全网络中的视频监控仪器、收款机等。在本公开实施例中,终端设备可以与网络设备,例如图1中的网络设备进行通信。多个终端之间也可以进行通信。终端可以是静态固定的,也可以是移动的,本公开对此不作限定。
图2是根据一示例性实施例示出的一种获取信道质量的方法的流程图。该方法可以应用于终端设备,如图2所示,该方法可以包括:
S201、终端设备接收网络设备通过下行信道发送的导频信号。
示例地,在上述通信系统中,网络设备可以通过下行信道向终端设备发送导频信号。相应地,终端设备可以接收该导频信号。
在一些实施例中,该导频信号可以包括信道状态信息参考信号CSI-RS(Channel State Information Reference Signal)。
S202、终端设备根据导频信号获取第一信道矩阵。
该第一信道矩阵可以用于表征下行信道的信道质量。
示例地,终端设备可以根据接收到的导频信号(例如CSI-RS),进行信道状态信息CSI估计,得到CSI估计矩阵;然后可以根据该CSI估计矩阵得到表征下行信道质量的第一信道矩阵。
S203、终端设备根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对第一信道矩阵进行压缩,得到压缩后的目标信道矩阵。
其中,CSI压缩模型包括信道编码器,信道编码器包括多个子编码器;不同的子编码器对应不同的CSI压缩参数。
在一些实施例中,终端设备可以根据CSI压缩参数,从多个子编码器中确定一个或多个目标子编码器,通过该目标子编码器对第一信道矩阵进行压缩,得到目标信道矩阵。
示例地,该CSI压缩参数可以表征CSI压缩率,该CSI压缩率的取值可以是预设的任意数值,例如,可以包括:1/2、1/4、1/8、1/16、1/32或1/64等,本公开对此不作限定。
S204、终端设备将目标信道矩阵发送至网络设备。
通过发送该目标信道矩阵,可以使得网络设备根据该目标信道矩阵确定下行信道的信道质量。
采用上述方法,终端设备接收网络设备通过下行信道发送的导频信号;根据导频信号获取第一信道矩阵;根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对第一信道矩阵进行压缩,得到压缩后的目标信道矩阵;将目标信道矩阵发送至网络设备,以便网络设备根据目标信道矩阵确定下行信道的信道质量。其中,第一信道矩阵用于表征下行信道的信道质量;CSI压缩模型包括信道编码器,信道编码器包括多个子编码器;不同的子编码器对应不同的CSI压缩参数;目标信道矩阵用于指示网络设备确定下行信道的信道质量。这样,通过多个子编码器可以适配不同的CSI压缩参数,从而可以在CSI压缩率发生变化的场景下,可以自适应地得到较为准确的目标信道矩阵并发送至网络设备,以便网络设备得到较为准确的信道质量,从而提高数据传输效率。
在一些实施例中,终端设备可以通过以下方式获取上述第一信道矩阵:
首先,根据导频信号,测量得到空域信道矩阵。
其次,通过离散傅里叶变换,将空域信道矩阵变换为角度时延域信道矩阵。
再次,根据角度时延域信道矩阵,确定第一信道矩阵。
示例地,上述通信系统可以采用基于OFDM(Orthogonal Frequency Division Multiplexing,正交频分复用技术)的mMIMO技术,子载波数量为N s,网络设备用于mMIMO的天线数目可以为N t个,上述导频信号可以包括信道状态信息参考信号CSI-RS。以该通信 系统为例对上述步骤进行举例说明如下:
首先,终端设备可以根据接收到的CSI-RS,可以测量(也可以称为估计)得到空域信道矩阵H,该空域信道矩阵H的大小可以为N s×N t,该空域信道矩阵H可以表征每个天线每个子载波的信道质量。
其次,终端设备可以通过离散傅里叶变换,将空域信道矩阵变换为角度时延域信道矩阵。例如,可以通过以下公式(1)得到角度时延域信道矩阵:
H a=F dHF a (1)
其中,H表示上述空域信道矩阵,H的大小可以为N s×N t,H a表示变换后的角度时延域信道矩阵,F d和F a表示大小为N s×N s,N t×N t的离散傅里叶变换矩阵,N s表示mMIMO对应的子载波数量,N t表示mMIMO对应的天线数量。
最后,根据角度时延域信道矩阵的主值部分,确定第一信道矩阵。
需要说明的是,由于多径时延的影响,上述角度时延域信道矩阵H a仅在前N c行有值,截取主值部分后得到角度时延域下的主值信道矩阵,该主值信道矩阵的大小可以为N c×N t
进一步地,将主值信道矩阵的实部和虚部分开,可以得到第一信道矩阵
Figure PCTCN2022093972-appb-000001
其中c为信道的实虚部维度,例如,c可以为2,该第一信道矩阵可以作为CSI压缩模型的输入,以得到目标信道矩阵。
这样,通过上述方式可以根据导频信号获取表征下行信道质量的第一信道矩阵。
上述CSI压缩参数可以是终端设备从网络设备接收的参数,也可以是终端设备预先设置的参数。
在一些实施例中,该CSI压缩参数可以是终端设备根据从网络设备接收的参数确定的。示例地,终端设备可以接收网络设备发送的第一压缩参数;根据该第一压缩参数确定CSI压缩参数。例如,终端设备可以通过RRC(Radio Resource Control,无线资源控制)信令(例如广播信令或针对该终端设备的专有信令)接收该第一压缩参数。
在一些实施例中,网络设备可以预先设置CSI压缩参数的取值,根据该CSI压缩参数确定第一压缩参数,并将该第一压缩参数发送至终端设备。
示例地,该CSI压缩参数和第一压缩参数可以通过预先设置的第一压缩参数对应关系确定。例如,在该CSI压缩参数表征CSI压缩率的情况下,CSI压缩参数的取值1/2对应的第一压缩参数可以为1,CSI压缩参数的取值1/4对应的第一压缩参数可以为2,CSI压缩参数的取值1/8对应的第一压缩参数可以为3,其他取值可以以此类推。
这样,终端设备接收到网络设备发送的第一压缩参数后,可以根据该第一压缩参数确定CSI压缩参数。
在一些实施例中,网络设备可以更新CSI压缩参数的值,并根据更新后的CSI压缩参数确定新的第一压缩参数;同样地,终端设备也可以在接收到新的第一压缩参数的情况下,根据新的第一压缩参数确定新的CSI压缩参数。
在另一些实施例中,该CSI压缩参数的取值可以是终端设备预先设置的,例如,可以是终端设备的预设参数取值,或者,可以是终端设备根据用户输入设置的参数取值。
在一些实施例中,终端设备可以根据预先设置的CSI压缩参数的取值,确定第二压缩参数,并将该第二压缩参数发送至网络设备,从而协商一致使用相同的CSI压缩参数。例如,终端设备可以通过RRC信令,将该第二压缩参数发送至网络设备。
同样地,该CSI压缩参数和第二压缩参数可以通过预先设置的第二压缩参数对应关系确定。
在一些实施例中,终端设备可以根据自身的设备参数确定CSI压缩参数的取值,该设备参数可以包括以下一项或多项:终端设备的协议版本、终端设备的信号质量、终端设备与网络设备的距离、终端设备的上行数据量、以及、终端设备的下行数据量。其中,终端 设备的信号质量可以包括RSRP(Reference Signal Receiving Power,参考信号接收功率)或SINR(Signal to Interference plus Noise Ratio,信号与干扰加噪声比)。
在一些实施例中,终端设备可以更新CSI压缩参数的值,并根据更新后的CSI压缩参数确定新的第二压缩参数。例如,终端设备可以在确定上述设备参数发生变化的情况下,根据变化后的新的设备参数,更新CSI压缩参数的取值。同样地,网络设备也可以在接收到新的第二压缩参数的情况下,根据新的第二压缩参数确定新的CSI压缩参数。
在另外一些实施例中,该CSI压缩参数可以包括分别在终端设备和网络设备预先设置的第三压缩参数,终端设备和网络设备预先设置的第三压缩参数的值可以相同。
图3是根据一示例性实施例示出的一种获取信道质量的方法的流程图。该方法可以应用于网络设备,如图3所示,该方法可以包括:
S301、网络设备接收终端设备发送的目标信道矩阵。
其中,目标信道矩阵为终端设备根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对第一信道矩阵进行压缩后得到的,第一信道矩阵为终端设备根据导频信号获取的用于表征下行信道的信道质量的矩阵。
在一些实施例中,网络设备可以通过下行信道发送导频信号,以便终端设备接收该导频信号并获取第一信道矩阵。示例地,该导频信号可以包括信道状态信息参考信号CSI-RS。
S302、网络设备根据信道状态信息CSI解压缩模型和CSI压缩参数,对目标信道矩阵进行解压缩,得到第三信道矩阵。
其中,该CSI解压缩模型可以包括信道解码器,信道解码器包括多个子解码器,不同的子解码器对应不同的CSI压缩参数。
在一些实施例中,网络设备可以根据CSI压缩参数,从多个子解码器中确定一个或多个目标子解码器,通过该目标子解码器对第一信道矩阵进行压缩,得到目标信道矩阵。
示例地,该CSI压缩参数可以表征CSI压缩率,该CSI压缩率的取值可以是预设的任意数值,例如,可以包括:1/2、1/4、1/8、1/16、1/32或1/64等,本公开对此不作限定。
S303、网络设备根据第三信道矩阵确定下行信道的信道质量。
采用上述方法,接收终端设备发送的目标信道矩阵;根据信道状态信息CSI解压缩模型和CSI压缩参数,对目标信道矩阵进行解压缩,得到第三信道矩阵;根据第三信道矩阵确定下行信道的信道质量。其中,目标信道矩阵为终端设备根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对第一信道矩阵进行压缩后得到的,第一信道矩阵为终端设备根据导频信号获取的用于表征下行信道的信道质量的矩阵;该CSI解压缩模型可以包括信道解码器,信道解码器包括多个子解码器,不同的子解码器对应不同的CSI压缩参数。这样,网络设备通过多个子解码器可以适配不同的CSI压缩参数,从而可以在CSI压缩率发生变化的场景下,可以自适应地得到较为准确的第三信道矩阵,并根据该第三信道矩阵得到较为准确的信道质量,网络设备可以根据该信道质量确定下行信道对应的调制与编码策略,从而提高数据传输效率。
在一些实施例中,网络设备可以将CSI压缩参数对应的子解码器作为目标子解码器;通过目标子解码器对目标信道矩阵进行解压缩,得到第四信道矩阵;根据第四信道矩阵确定第三信道矩阵。
示例地,该子解码器可以包括解压缩层,该解压缩层可以与终端设备的信道编码器中的子编码器的压缩层相对应。
在一些实施例中,该CSI解压缩模型还可以包括CSI重构模块,网络设备可以将上述第四信道矩阵输入CSI重构模块,得到第三信道矩阵。
示例地,该CSI重构模块可以包括卷积神经网络CNN。
在一些实施例中,该CSI压缩参数的取值可以是网络设备预先设置的,例如,可以是 网络设备的预设参数取值,或者,可以是网络设备根据用户输入设置的参数取值。
在一些实施例中,网络设备可以根据该CSI压缩参数确定第一压缩参数,并将该第一压缩参数发送至终端设备,以指示终端设备根据该第一压缩参数确定CSI压缩参数。例如,网络设备可以通过RRC(Radio Resource Control,无线资源控制)信令(例如广播信令或针对该终端设备的专有信令)发送该第一压缩参数。
示例地,该CSI压缩参数和第一压缩参数可以通过预先设置的第一压缩参数对应关系确定。例如,在该CSI压缩参数表征CSI压缩率的情况下,CSI压缩参数的取值1/2对应的第一压缩参数可以为1,CSI压缩参数的取值1/4对应的第一压缩参数可以为2,CSI压缩参数的取值1/8对应的第一压缩参数可以为3,其他取值可以以此类推。这样,终端设备接收到网络设备发送的该第一压缩参数后,可以根据该第一压缩参数确定CSI压缩参数。
在一些实施例中,网络设备可以针对不同的终端设备设置不同的CSI压缩参数的取值,示例地,可以根据终端设备对应的设备参数确定CSI压缩参数的取值,该设备参数可以包括以下一项或多项:终端设备的协议版本、终端设备的信号质量、终端设备与网络设备的距离、终端设备的上行数据量、以及、终端设备的下行数据量。其中,终端设备的信号质量可以包括RSRP(Reference Signal Receiving Power,参考信号接收功率)或SINR(Signal to Interference plus Noise Ratio,信号与干扰加噪声比)。
在另一些实施例中,网络设备可以更新CSI压缩参数的值,并根据更新后的CSI压缩参数确定新的第一压缩参数。例如,网络设备可以获取终端设备对应的设备参数,在确定该设备参数发生变化的情况下,可以更新CSI压缩参数的取值。
同样地,终端设备也可以在接收到新的第一压缩参数的情况下,根据新的第一压缩参数确定新的CSI压缩参数。
在另外一些实施例中,该CSI压缩参数可以是网络设备根据从终端设备接收的参数确定的。示例地,网络设备可以接收终端设备发送的第二压缩参数;根据该第二压缩参数确定CSI压缩参数。例如,网络设备可以通过RRC信令接收该第二压缩参数。
示例地,终端设备可以预先设置CSI压缩参数的取值,根据该CSI压缩参数确定第二压缩参数,并将该第二压缩参数发送至网络设备。同样地,该CSI压缩参数和第二压缩参数可以通过预先设置的第二压缩参数对应关系确定。这样,网络设备接收到终端设备发送的第二压缩参数后,可以根据该第二压缩参数确定CSI压缩参数。
在一些实施例中,终端设备可以更新CSI压缩参数的值,并根据更新后的CSI压缩参数确定新的第二压缩参数;同样地,网络设备也可以在接收到新的第二压缩参数的情况下,根据新的第二压缩参数确定新的CSI压缩参数。
在另外一些实施例中,该CSI压缩参数也可以分别在终端设备和网络设备预先设置的第三压缩参数,终端设备和网络设备预先设置的第三压缩参数的值可以相同。
在一些实施例中,上述CSI压缩模型和CSI解压缩模型可以共同组成获取信道质量的网络模型。以下结合附图对该获取信道质量的网络模型进行详细说明。
图4是根据一示例性实施例示出的一种获取信道质量的网络模型的结构示意图。
根据图4所示,在一些实施例中,该获取信道质量的网络模型可以包括信道状态信息CSI压缩模型41和信道状态信息CSI解压缩模型42,该CSI压缩模型41可以部署在图1所示通信系统中的终端设备上,例如终端设备可以通过软件、硬件或软硬件结合的方式,运行该CSI压缩模型41;该CSI解压缩模型42可以部署在图1所示通信系统中的网络设备(例如基站)上,例如网络设备可以通过软件、硬件或软硬件结合的方式,运行该CSI解压缩模型42。
该CSI压缩模型41可以根据CSI压缩参数对输入的第一信道矩阵压缩后输出目标信道矩阵(该目标信道矩阵也可以称为码字);终端设备可以将该目标信道矩阵发送至网络 设备,相应地,网络设备可以将接收到的该目标信道矩阵输入CSI解压缩模型42,该CSI解压缩模型42可以根据CSI压缩参数将该目标信道矩阵解码(也可以称为解压缩)输出第三信道矩阵,网络设备可以根据该第三信道矩阵确定下行信道的信道质量。
根据图4所示,在一些实施例中,该CSI压缩模型41可以包括信道编码器411,该信道编码器411可以根据CSI压缩参数,将输入的信道矩阵进行编码,以得到目标信道矩阵。
进一步地,该信道编码器411可以包括多个子编码器,不同的子编码器对应不同的CSI压缩参数。这样,终端设备可以将CSI压缩参数对应的子编码器作为第一目标子编码器;通过第一目标子编码器对第一信道矩阵进行压缩,得到目标信道矩阵。该目标信道矩阵也可以称为码字
示例地,在该CSI压缩参数包括CSI压缩率的情况下,该信道编码器411可以包括子编码器1(对应的CSI压缩参数为1/2)、子编码器2(对应的CSI压缩参数为1/4)、子编码器3(对应的CSI压缩参数为1/8)、子编码器3(对应的CSI压缩参数为1/16)、子编码器4(对应的CSI压缩参数为1/32)、子编码器5(对应的CSI压缩参数为1/64)等。
根据图4所示,在另一些实施例中,该CSI压缩模型可以包括信道编码器411和特征转换器(也可以称为特征优化器或特征联合优化器)412。其中,该特征转换器412的输入可以是上述第一信道矩阵,用于对第一信道矩阵进行关键特征提取得到表征CSI关键特征的第二信道矩阵,该第二信道矩阵可以作为信道编码器411的输入,该信道编码器411可以根据CSI压缩参数对该第二信道矩阵进行压缩,得到目标信道矩阵。示例地,终端设备可以CSI压缩参数对应的子编码器作为第二目标子编码器,然后通过该第二目标子编码器对第二信道矩阵进行压缩,得到目标信道矩阵。该目标信道矩阵也可以称为码字。
图5是根据一示例性实施例示出的一种CSI压缩模型中的特征转换器的示意图。如图5所示,该特征转换器412包括特征提取网络4121、注意力机制网络4122和特征还原网络4123。
在一些实施例中,终端设备可以将第一信道矩阵输入特征转换器,对第一信道矩阵进行关键特征提取,得到表征CSI关键特征的第二信道矩阵。
示例地,可以通过以下步骤获取第二信道矩阵:
S51、将第一信道矩阵输入特征提取网络,得到多个第一特征图。
示例地,该特征提取网络将第一信道矩阵H c转换成第一特征图
Figure PCTCN2022093972-appb-000002
其中f表示提取的第一特征图的数量。f可以是大于1的任意正整数。
在一些实施例中,该特征提取网络可以包括二维卷积层,卷积核的大小可以为f×m×m,其中f表示第一特征图的数量,m×m代表该卷积核使用的卷积窗的长度和宽度。该特征提取网络可以采用二维归一化层对卷积层的输出进行归一化,该特征提取网络的激活函数可以包括Sigmoid、ReLU、LeakyReLU、PReLU或ELU。
示例地,该激活函数可以采用LeakyReLU(带泄露线性整流函数)激活函数,该激活函数可以包括以下公式(2):
Figure PCTCN2022093972-appb-000003
其中,x表示输入向量,例如输入的第一特征图;γ表示预设系数,示例地,该预设系数可以是小于1的任意数值,例如,该预设系数可以为0.3;LeakyReLU(x)表示该Leaky ReLU激活函数的输出值。
S52、将多个第一特征图输入注意力机制网络,得到第二特征图。
该第二特征图可以包括第一特征图中的关键特征信息。
在一些实施例中,可以通过注意力机制网络对多个第一特征图进行最大池化操作,得到最大池化特征图;通过注意力机制网络对多个第一特征图进行平均池化操作,得到平均 池化特征图;然后,根据最大池化特征图和平均池化特征图,确定第二特征图。
需要说明的是,针对特征提取网络提取的多个第一特征图,有些第一特征图里面包含的特征信息对CSI重构有着极大的帮助,可以称为“关键特征图”。而有些第一特征图里面包含的特征信息对CSI重构几乎没有影响,可以称为“非必要特征图”。该注意力机制网络可以在多个特征图中提取出关键特征图,从而使得后续编码器产生的码字中包含更多的关键特征。示例地:
该注意力机制网络可以对第一特征图F进行最大池化操作,得到最大池化特征图M∈
Figure PCTCN2022093972-appb-000004
例如,可以通过以下公式(3)进行最大池化操作:
m i=max{F 1,i,F 2,i,F 3,i,…,F f,i},i={1,2,…,N cN t} (3)
其中,m i表示最大池化特征图M中的第i个元素,f表示第一特征图的数量,F 1,i表示第1个第一特征图中的第i个元素,F 2,i表示第2个第一特征图中的第i个元素,F 2,i代表第2个第一特征图中的第i个元素,F f,i表示第f个第一特征图中的第i个元素,N c×N t表示第一信道矩阵的大小。
这样,最大池化特征图M中的每个元素均由多个第一特征图中对应位置的最大元素组成。
该注意力机制网络还可以对第一特征图F进行平均池化操作,得到平均池化特征图
Figure PCTCN2022093972-appb-000005
Figure PCTCN2022093972-appb-000006
例如,可以通过以下公式(4)进行平均池化操作:
Figure PCTCN2022093972-appb-000007
其中,v i表示平均池化特征图V中的第i个元素,f表示第一特征图的数量,F 1,i表示第1个第一特征图中的第i个元素,F 2,i表示第2个第一特征图中的第i个元素,F 2,i代表第2个第一特征图中的第i个元素,F f-1,i表示第f-1个第一特征图中的第i个元素,F f,i表示第f个第一特征图中的第i个元素,N c×N t表示第一信道矩阵的大小。
这样,平均池化特征图V中的每个元素由多个第一特征图对应位置的所有元素的均值组成。
进一步地,可以据最大池化特征图和平均池化特征图,确定第二特征图。
在一些实施例中,可以将最大池化特征图和平均池化特征图输入融合子网络,得到融合后的融合特征图;并根据融合特征图和第一特征图,计算得到第二特征图。
在一些实施例中,可以将该最大池化特征图M和该平均池化特征图V拼接得到联合特征图
Figure PCTCN2022093972-appb-000008
将联合特征图C通过融合网络得到融合特征图
Figure PCTCN2022093972-appb-000009
在一些实施例中,该融合网络可以采用二维卷积层,卷积核的大小为m×n×n,该融合网络还可以包括二维归一化层和激活函数,该激活函数可以包括Sigmoid激活函数。
然后,将融合特征图D与第一特征矩阵F相乘,可以得到第二特征图
Figure PCTCN2022093972-appb-000010
该第二特征图F′也可以称为优化特征图。
这样,在第二特征图F′中凸显了关键特征图中的特征信息,同时弱化了非必要特征图中的特征信息,从而可以表征CSI关键特征。
S53、将该第二特征图输入特征还原网络,得到第二信道矩阵。
示例地,可以通过该特征还原网络将上述第二特征图F′还原成第二信道矩阵
Figure PCTCN2022093972-appb-000011
Figure PCTCN2022093972-appb-000012
在一些实施例中,该特征还原网络可以包括二维卷积层、二维归一化层和激活函数,该激活函数可以是LeakyReLU激活函数,也可以是相关技术中的其他激活函数,本公开对此不作限定。
在一些实施例中,该特征还原网络和该特征提取网络可以采用相同的激活函数,以避免特征失真。
这样,通过该方式得到的第二信道矩阵H e凸显了关键特征信息,弱化了非必要特征信息,从而可以表征CSI关键特征。
图6是根据一示例性实施例示出的一种CSI压缩模型中的信道编码器的示意图。如图6所示,该信道编码器411可以包括多个子编码器,例如图中的子编码器1、子编码器2、……、子编码器T-1,子编码器T。其中,T为子编码器的数量。
在一些实施例中,该信道编码器可以首先将输入的第二信道矩阵H e进行预处理,例如可以进行维度变换,变换后的第二信道矩阵的维度大小可以为
Figure PCTCN2022093972-appb-000013
将维度变换后的第二信道矩阵输入上述子编码器进行处理。
在一些实施例中,CSI压缩参数可以包括多个预设CSI压缩率,该预设CSI压缩率的集合可以包括σ={σ 12,…,σ T},将集合中的预设CSI压缩率按照从大到小的顺序进行降序排列。对于集合里的每一个压缩率σ j,j={1,2,…,T},该信道编码器中均有与该预设CSI压缩率对应的子编码器,即子编码器的数目与预设CSI压缩率的数目相等。
示例地,子编码器1对应的预设CSI压缩率σ 1可以为1/2,子编码器2对应的预设CSI压缩率σ 2可以为1/4,子编码器3对应的预设CSI压缩率σ 3可以为1/8,子编码器4对应的预设CSI压缩率σ 4可以为1/16,子编码器5对应的预设CSI压缩率σ 5可以为1/32,子编码器6对应的预设CSI压缩率σ 6可以为1/64,以此类推。需要说明的是,此处的预设CSI压缩率的取值均为举例,本公开对具体的取值不作限定。
在一些实施例中,该预设CSI压缩率的集合可以包括σ={1/4,1/16,1/32,1/64},即子编码器的数目T=4,将集合中的预设CSI压缩率进行降序排列。对于集合里的每一个预设CSI压缩率,自适应编码器中均有对应的子编码器与之匹配,即子编码器的数目与动态压缩率的数目相等。例如,子编码器1对应的预设CSI压缩率σ 1可以为1/4,子编码器2对应的预设CSI压缩率σ 2可以为1/16,子编码器3对应的预设CSI压缩率σ 3可以为1/32,子编码器4对应的预设CSI压缩率σ 4可以为1/64。
在一些实施例中,每个子编码器可以包括压缩层,该压缩层可以包括全连接层。每个子编码器的压缩层的大小可以根据CSI压缩参数(例如预设CSI压缩率)和第一信道矩阵的大小确定,例如,第一信道矩阵的大小为c×N c×N t,预设CSI压缩率为σ 1,则σ 1对应的子编码器的压缩层的大小为(c×N c×N t)×d 1,其中d 1可以是c×N c×N t×σ 1
例如,第一信道矩阵的大小为2×32×32,预设CSI压缩率为1/4,则该预设CSI压缩率对应的子编码器的压缩层的大小可以为2048×512。
在另一些实施例中,每个子编码器可以包括压缩层和编码开关,这样,在实际使用时,可以确定CSI压缩参数(例如CSI压缩率)对应的目标子编码器(例如上述实施例中的第一目标子编码器或第二目标子编码器),并将该目标子编码器的编码开关闭合,其他子编码器的开关断开,从而可以使用该目标子编码器的压缩层对输入的第二信道矩阵进行压缩,输出该目标子编码器的待定信道矩阵,可以将该待定信道矩阵作为目标信道矩阵M。
在一些实施例中,将取值最大的预设CSI压缩率作为最大CSI压缩率,该最大CSI压缩率对应的子编码器(例如图5中的子编码器1)的压缩层作为最大压缩层,该最大压缩层的输出可以作为其他子编码器的压缩层的输入,这样,可以提高压缩运算效率。
示例地,预设CSI压缩率σ 1为最大压缩率,σ 1对应的子编码器1的压缩层1的大小可以为(c×N c×N t)×d 1,其中d 1可以是c×N c×N t×σ 1。可以将该压缩层1的输出作为其他子编码器对应的压缩层的输入,其他子编码器对应的压缩层的大小可以为d 1×d k,d k可以是c×N c×N t×σ k。其中k表示子编码器的编号,σ k表示第k个子编码器对应的预设CSI压缩率,c×N c×N t表示第一信道矩阵的大小。
例如,第一信道矩阵的大小为2×32×32,最大压缩率σ 1为1/2,则σ 1对应的子编码器1的压缩层1的大小可以为2048×1024;子编码器2对应的预设CSI压缩率σ 2为1/4, 则该子编码器2的压缩层2的大小可以为1024×512;子编码器3对应的预设CSI压缩率σ 3为1/16,则该子编码器3的压缩层3的大小可以为1024×256。
通过该方式,图6中的子编码器2至子编码器T中的压缩层可以起到进一步降维的作用。
图7是根据一示例性实施例示出的一种CSI解压缩模型的示意图。如图7所示,该CSI解压缩模型42可以包括信道解码器421。
在一些实施例中,信道解码器421可以由多个子解码器,不同的子解码器对应不同的CSI压缩参数。每个子解码器可以通过解压缩层对接收到的目标信道矩阵进行解压缩,从而得到第三信道矩阵。该解压缩层可以包括全连接层。
如图7所示,该信道解码器421可以包括多个子解码器,例如,子解码器1、子解码器2、……、子解码器T-1,子解码器T。其中,T为子解码器的数量。
在一些实施例中,CSI压缩参数可以包括多个预设CSI压缩率,该预设CSI压缩率的集合可以包括σ={σ 12,…,σ T},将集合中的预设CSI压缩率按照从大到小的顺序进行降序排列。对于集合里的每一个压缩率σ j,j={1,2,…,T},该信道解码器中均有与该预设CSI压缩率对应的子解码器,即子解码器的数目与预设CSI压缩率的数目相等。
示例地,子解码器1对应的预设CSI压缩率σ 1可以为1/2,子解码器2对应的预设CSI压缩率σ 2可以为1/4,子解码器3对应的预设CSI压缩率σ 3可以为1/8,子解码器4对应的预设CSI压缩率σ 4可以为1/16,子解码器5对应的预设CSI压缩率σ 5可以为1/32,子解码器6对应的预设CSI压缩率σ 6可以为1/64,以此类推。需要说明的是,此处的预设CSI压缩率的取值均为举例,本公开对具体的取值不作限定。
在一些实施例中,每个子解码器可以包括解压缩层和解码开关,这样,在实际使用时,可以确定CSI压缩参数(例如CSI压缩率)对应的目标子解码器,并将该目标子解码器的解码开关闭合,其他子解码器的开关断开,从而可以使用该目标子解码器的解压缩层对输入的目标信道矩阵进行解压缩,输出该目标子解码器对应的待定信道矩阵,可以将该待定信道矩阵作为第四信道矩阵;并进一步根据该第四信道矩阵确定第三信道矩阵。
在一些实施例中,该CSI解压缩模型还可以包括CSI重构模块422。该CSI重构模块可以包括CNN(Convolutional Neural Networks,卷积神经网络),示例地,该CSI重构模块包括两个CNN网络,每个CNN网络包括5个卷积层,每个卷积层的卷积核大小依次为c×k×k,f 1×l×l,f 2×l×l,f 2×n×n,c×m×m,(f 1,f 2,k,l,m,n均为预先设置的值,可以根据不同的卷积层预设不同的值)。每个卷积层的步长均为t,可以均采用归一化层和LeakyReLU激活函数。然后,通过Sigmoid激活函数层将第二个CNN模块的输出元素值映射到[0,1]区间。这样,该CSI重构模块可以输出目标信道矩阵对应的第三信道矩阵
Figure PCTCN2022093972-appb-000014
需要说明的是,该CSI重构模块,针对不同的CSI压缩参数(例如CSI压缩率),可以输出不同的第三信道矩阵。示例地,CSI压缩参数中的预设CSI压缩率的集合可以包括σ={σ 12,…,σ T},则对应输出的第三信道矩阵可以分别为
Figure PCTCN2022093972-appb-000015
这样,通过该CSI解压缩模型,网络设备可以对接收到的目标信道矩阵进行解压缩,得到第三信道矩阵,以便根据该第三信道矩阵确定下行信道的信道质量。
在一些实施例中,可以通过离线训练获取上述CSI压缩模型和CSI解压缩模型,例如,可以将上述CSI压缩模型和CSI解压缩模型进行联合训练,得到CSI压缩模型和CSI解压缩模型的参数,使得CSI压缩模型和CSI解压缩模型能够匹配。
对上述模型的训练可以在终端设备进行,也可以在网络设备进行,以下结合附图分别进行说明。
图8是根据一示例性实施例示出的一种CSI压缩模型的训练方法的流程图。该训练方法可以应用于终端设备。如图8所示,该训练方法可以包括:
S801、终端设备获取用于训练的第一样本信道矩阵。
其中,第一样本信道矩阵为终端设备根据接收到的导频信号获取的用于表征下行信道质量的矩阵。
在一些实施例中,可以在FDD下行链路mMIMO系统中,在网络设备(例如基站)侧以ULA(Uniform Linear Array,均匀线性阵列)方式,半波长间隔配置N t=32根天线,在终端设备配置单天线。使用COST2100信道模型,在5.3GHz室内微蜂窝场景产生150,000个空域CSI矩阵样本,并获取包含100,000个样本的训练集,包含30,000个样本的验证集,包含20,000个样本的测试集。可以将其中的训练集作为上述第一样本信道矩阵。
S802、终端设备根据第一样本信道矩阵,对第一目标网络模型进行训练,以得到CSI压缩模型。
其中,第一目标网络模型包括第一目标压缩模型和第一目标解压缩模型,第一目标压缩模型的网络结构与CSI压缩模型的网络结构相同,例如,都可以包括信道编码器,信道编码器包括多个子编码器,不同的子编码器对应不同的CSI压缩参数;第一目标解压缩模型包括信道解码器,信道解码器包括多个子解码器,不同的子解码器对应不同的CSI压缩参数。
在一些实施例中,上述第一目标压缩模型的模型结构可以与图4所示CSI压缩模型相同,示例地,该第一目标压缩模型可以包括信道编码器和特征转换器,该特征转换器的结构可以如图5所示,该信道编码器的结构可以如图6所示;上述第一目标解压缩模型的模型结构可以与图7所示的CSI解压缩模型相同,此处对上述模型结构均不再赘述。
在一些实施例中,可以循环执行第一模型训练步骤,直至根据第一样本信道矩阵和第一预测信道矩阵确定训练后的第一目标网络模型满足第一预设停止迭代条件,将训练后的第一目标网络模型中的第一目标压缩模型作为CSI压缩模型。
其中,第一预测信道矩阵为第一样本信道矩阵输入第一目标网络模型后输出的矩阵。
示例地,可以将第一目标压缩模型对应的第一压缩模型参数作为CSI压缩模型的模型参数。
其中,该第一模型训练步骤可以包括:
S81、将第一样本信道矩阵输入第一目标压缩模型,通过多个子编码器对第一样本信道矩阵进行压缩后,得到第一目标样本信道矩阵。
S82、将第一目标样本信道矩阵输入第一目标解压缩模型,通过多个子解码器对第一样本信道矩阵进行压缩后,得到第一预测信道矩阵。
S83、在根据第一样本信道矩阵和第一预测信道矩阵确定第一目标网络模型不满足第一预设停止迭代条件的情况下,根据第一样本信道矩阵和第一预测信道矩阵确定第一损失值,根据第一损失值更新第一目标网络模型的参数,得到训练后的第一目标网络模型,并将该训练后的第一目标网络模型作为新的第一目标网络模型。
需要说明的是,上述第一预设停止迭代条件可以根据训练过程中使用的损失函数确定。
在一些实施例中,可以将所有子编码器对应的编码开关闭合,将所有子解码器对应的解码开关也闭合,并可以针对多个CSI压缩参数下的训练误差进行联合优化,训练过程中使用的损失函数可以包括以下公式(5):
Figure PCTCN2022093972-appb-000016
其中,σ 1表示第一预设CSI压缩率,σ 2表示第一预设CSI压缩率,σ T表示第T预设CSI压缩率,T表示CSI压缩参数中预设CSI压缩率的数目(同时也可以表示子编码器的数目,或,子解码器的数目),
Figure PCTCN2022093972-appb-000017
表示预设CSI压缩率σ 1对应的损失值,Loss表示上述第一样本信道矩阵和第一预测信道矩阵的第一损失值。
这样,通过对多个CSI压缩参数(例如预设CSI压缩率)对应的训练误差进行联合优 化,网络可以在一次训练过程中将所有预设CSI压缩率对应的训练误差优化完成,从而提升了网络对CSI压缩率动态变化的适应能力。训练结束后,可以得到CSI压缩模型的参数。
在一些实施例中,终端设备通过上述训练过程,还可以得到第一目标解压缩模型训练后的参数,该第一目标解压缩模型训练后的参数可以作为网络设备侧的CSI解压缩模型的参数。
在一些实施例中,终端设备通过该训练方法确定模型参数后,可以获取训练后的第一目标网络模型中的第一目标解压缩模型对应的第一解压缩模型参数;并将第一解压缩模型参数发送至网络设备,以便指示网络设备根据第一解压缩模型参数确定CSI解压缩模型。其中,该CSI解压缩模型用于网络设备根据目标信道矩阵确定下行信道的信道质量的模型。
示例地,终端设备可以通过信令或数据报文,将第一解压缩模型参数发送至网络设备。
在本公开的另外一些实施例中,上述训练方法可以在网络设备执行,终端设备可以接收网络设备发送的第二压缩模型参数;并根据第二压缩模型参数,确定CSI压缩模型。
例如,可以将第二压缩模型参数作为CSI压缩模型的模型参数。
示例地,终端设备可以通过信令或数据报文,接收网络设备发送的第二压缩模型参数,并根据该第二压缩模型参数确定CSI压缩模型对应的参数。
图9是根据一示例性实施例示出的一种CSI解压缩模型的训练方法的流程图,该训练方法可以应用于网络设备。如图9所示,该训练方法可以包括:
S901、网络设备获取用于训练的第二样本信道矩阵。
该第二样本信道矩阵可以是终端设备根据接收到的导频信号获取的用于表征下行信道质量的矩阵。
在一些实施例中,可以在FDD下行链路mMIMO系统中,在网络设备(例如基站)侧以ULA方式,半波长间隔配置N t=32根天线,在终端设备配置单天线。使用COST2100信道模型,在5.3GHz室内微蜂窝场景产生150,000个空域CSI矩阵样本,并获取为含100,000个样本的训练集,含30,000个样本的验证集,含20,000个样本的测试集。可以将其中的训练集作为该第一样本信道矩阵。
S902、网络设备根据第二样本信道矩阵,对第二目标网络模型进行训练,以得到CSI解压缩模型。
其中,第二目标网络模型包括第二目标压缩模型和第二目标解压缩模型,第二目标解压缩模型的网络结构与CSI解压缩模型的网络结构相同,例如,都包括信道解码器,该信道解码器可以包括多个子解码器,不同的子解码器对应不同的CSI压缩参数;第二目标压缩模型包括信道编码器,信道编码器包括多个子编码器,不同的子编码器对应不同的CSI压缩参数。
在一些实施例中,上述第二目标压缩模型的模型结构可以与图4所示CSI压缩模型相同,示例地,该第二目标压缩模型可以包括信道编码器和特征转换器,该特征转换器的结构可以如图5所示,该信道编码器的结构可以如图6所示;上述第二目标解压缩模型的模型结构可以与图7所示的CSI解压缩模型相同,此处对上述模型结构均不再赘述。
在一些实施例中,可以循环执行第二模型训练步骤,直至根据第二样本信道矩阵和第二预测信道矩阵确定训练后的第二目标网络模型满足第二预设停止迭代条件,将训练后的第二目标网络模型中的第二目标解压缩模型作为CSI解压缩模型;第二预测信道矩阵为第二样本信道矩阵输入第二目标网络模型后输出的矩阵。
其中,该第二模型训练步骤可以包括:
S91、将第二样本信道矩阵输入第二目标压缩模型,通过多个子编码器对第二样本信道矩阵进行压缩后,得到第二目标样本信道矩阵;
S92、将第二目标样本信道矩阵输入第二目标解压缩模型,通过多个子解码器对第二 目标样本信道矩阵进行解压缩后,得到第二预测信道矩阵;
S93、在根据第二样本信道矩阵和第二预测信道矩阵确定第二目标网络模型不满足第二预设停止迭代条件的情况下,根据第二样本信道矩阵和第二预测信道矩阵确定第二损失值,根据第二损失值更新第二目标网络模型的参数,得到训练后的第二目标网络模型,并将该训练后的第二目标网络模型作为新的第二目标网络模型。
需要说明的是,上述第二预设停止迭代条件同样可以根据训练过程中使用的损失函数确定。
在一些实施例中,可以将所有子编码器对应的编码开关闭合,将所有子解码器对应的解码开关也闭合,可以针对多个CSI压缩参数下的训练误差进行联合优化,训练过程中使用的损失函数同样可以包括上述公式(5),此处不再赘述。
这样,通过对多个CSI压缩参数(例如预设CSI压缩率)对应的训练误差进行联合优化,网络可以在一次训练过程中将所有预设CSI压缩率对应的训练误差优化完成,从而提升了网络对CSI压缩率动态变化的适应能力。训练结束后,可以得到CSI解压缩模型的参数。
在一些实施例中,网络设备通过上述训练过程,还可以得到第二目标压缩模型训练后的参数,该第二目标压缩模型训练后的参数可以作为终端设备侧的CSI压缩模型的参数。
在一些实施例中,网络设备可以获取训练后的第二目标网络模型中的第二目标压缩模型对应的第二压缩模型参数;并将第二压缩模型参数发送至终端设备,以便指示终端设备根据第二压缩模型参数确定CSI压缩模型。其中,该CSI压缩模型用于终端设备根据第一信道矩阵获取目标信道矩阵。
示例地,网络设备可以通过信令或数据报文,将第二压缩模型参数发送至终端设备。
在本公开的另外一些实施例中,上述训练方法可以在终端设备执行,网络设备可以接收终端设备发送的第一解压缩模型参数;并根据第一解压缩模型参数,确定CSI解压缩模型。
示例地,网络设备可以通过信令或数据报文,接收终端设备发送的第一解压缩模型参数,并根据该第一解压缩模型参数确定CSI解压缩模型对应的参数。
图10是根据一示例性实施例示出的一种获取信道质量的方法,如图10所示,该方法可以包括:
S1001、网络设备通过下行信道发送导频信号。
S1002、终端设备通过下行信道接收导频信号,并根据导频信号获取第一信道矩阵。
第一信道矩阵用于表征下行信道的信道质量;
S1003、终端设备根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对第一信道矩阵进行压缩,得到压缩后的目标信道矩阵。
在一些实施例中,该CSI压缩模型包括信道编码器,信道编码器包括多个子编码器;不同的子编码器对应不同的CSI压缩参数。
在一些实施例中,该CSI压缩模型可以包括信道编码器和特征转换器。
S1004、终端设备将目标信道矩阵发送至网络设备。
S1005、网络设备接收终端设备发送的目标信道矩阵,并根据信道状态信息CSI解压缩模型和CSI压缩参数,对目标信道矩阵进行解压缩,得到第三信道矩阵。
其中,CSI解压缩模型包括信道解码器,信道解码器包括多个子解码器,不同的子解码器对应不同的CSI压缩参数;
S1006、网络设备根据第三信道矩阵确定下行信道的信道质量。
需要说明的是,上述CSI压缩模型和上述CSI解压缩模型可以是上述实施例中提供的任意一种模型结构,此处不再赘述。
这样,通过多个子编码器和多个子解码器可以适配不同的CSI压缩参数,从而可以在CSI压缩率发生变化的场景下,终端设备可以自适应地得到较为准确的目标信道矩阵并发送至网络设备,以便网络设备得到较为准确的信道质量,从而提高数据传输效率。
图11是根据一示例性实施例示出的一种获取信道质量的装置1100的框图,该装置可以应用于终端设备。如图11所示,该装置1100可以包括:
第一接收模块1101,被配置为接收网络设备通过下行信道发送的导频信号;
第一矩阵获取模块1102,被配置为根据所述导频信号获取第一信道矩阵;所述第一信道矩阵用于表征所述下行信道的信道质量;
目标矩阵获取模块1103,被配置为根据CSI压缩模型和信道状态信息CSI压缩参数,对所述第一信道矩阵进行压缩,得到压缩后的目标信道矩阵;其中,所述CSI压缩模型包括信道编码器,所述信道编码器包括多个子编码器;不同的子编码器对应不同的CSI压缩参数;
第一发送模块1104,被配置为将所述目标信道矩阵发送至所述网络设备,以便所述网络设备根据所述目标信道矩阵确定所述下行信道的信道质量。
在一些实施例中,所述目标矩阵获取模块1103,被配置为将所述CSI压缩参数对应的子编码器作为第一目标子编码器;通过所述第一目标子编码器对所述第一信道矩阵进行压缩,得到目标信道矩阵。
在一些实施例中,所述CSI压缩模型还包括特征转换器;所述目标矩阵获取模块1103,被配置为将所述第一信道矩阵输入所述特征转换器,对所述第一信道矩阵进行关键特征提取,得到表征CSI关键特征的第二信道矩阵;根据所述CSI压缩参数和所述信道编码器,对所述第二信道矩阵进行压缩,得到所述目标信道矩阵。
在一些实施例中,所述目标矩阵获取模块1103,被配置为将所述CSI压缩参数对应的子编码器作为第二目标子编码器;通过所述第二目标子编码器对所述第二信道矩阵进行压缩,得到目标信道矩阵。
在一些实施例中,所述特征转换器包括特征提取网络、注意力机制网络和特征还原网络;所述目标矩阵获取模块1103,被配置为将所述第一信道矩阵输入所述特征提取网络,得到多个第一特征图;将多个所述第一特征图输入所述注意力机制网络,得到第二特征图;所述第二特征图包括所述第一特征图中的关键特征信息;将所述第二特征图输入所述特征还原网络,得到第二信道矩阵。
在一些实施例中,所述目标矩阵获取模块1103,被配置为通过所述注意力机制网络对多个所述第一特征图进行最大池化操作,得到最大池化特征图;通过所述注意力机制网络对多个所述第一特征图进行平均池化操作,得到平均池化特征图;根据所述最大池化特征图和所述平均池化特征图,确定第二特征图;
在一些实施例中,所述注意力机制网络包括融合子网络;所述目标矩阵获取模块1103,被配置为将所述最大池化特征图和所述平均池化特征图输入所述融合子网络,得到融合后的融合特征图;根据所述融合特征图和所述第一特征图,计算得到第二特征图。
在一些实施例中,所述第一矩阵获取模块1102,被配置为根据所述导频信号,测量得到空域信道矩阵;通过离散傅里叶变换,将所述空域信道矩阵变换为角度时延域信道矩阵;根据所述角度时延域信道矩阵,确定所述第一信道矩阵。
图12是根据一示例性实施例示出的一种获取信道质量的装置1100的框图,如图12所示,该装置1100还可以包括第一训练模块1105,该第一训练模块1105,被配置为通过以下方式训练得到的所述CSI压缩模型:
获取用于训练的第一样本信道矩阵;所述第一样本信道矩阵为终端设备根据接收到的导频信号获取的用于表征所述下行信道质量的矩阵;
根据所述第一样本信道矩阵,对第一目标网络模型进行训练,以得到所述CSI压缩模型;
其中,所述第一目标网络模型包括第一目标压缩模型和第一目标解压缩模型,所述第一目标压缩模型的网络结构与所述CSI压缩模型的网络结构相同,所述第一目标解压缩模型包括信道解码器,所述信道解码器包括多个子解码器,不同的子解码器对应不同的CSI压缩参数。
在一些实施例中,所述第一训练模块1105,被配置为循环执行第一模型训练步骤,直至根据所述第一样本信道矩阵和第一预测信道矩阵确定训练后的第一目标网络模型满足第一预设停止迭代条件,将训练后的第一目标网络模型中的第一目标压缩模型作为所述CSI压缩模型;所述第一预测信道矩阵为所述第一样本信道矩阵输入所述第一目标网络模型后输出的矩阵;
所述第一模型训练步骤包括:
将所述第一样本信道矩阵输入所述第一目标压缩模型,通过多个子编码器对所述第一样本信道矩阵进行压缩后,得到第一目标样本信道矩阵;
将所述第一目标样本信道矩阵输入所述第一目标解压缩模型,通过多个子解码器对所述第一目标样本信道矩阵进行解压缩后,得到第一预测信道矩阵;
在根据所述第一样本信道矩阵和所述第一预测信道矩阵确定所述第一目标网络模型不满足所述第一预设停止迭代条件的情况下,根据所述第一样本信道矩阵和所述第一预测信道矩阵确定第一损失值,根据所述第一损失值更新所述第一目标网络模型的参数,得到训练后的第一目标网络模型,并将该训练后的第一目标网络模型作为新的第一目标网络模型。
在一些实施例中,所述第一发送模块1104,还被配置为获取训练后的第一目标网络模型中的第一目标解压缩模型对应的第一解压缩模型参数;将所述第一解压缩模型参数发送至所述网络设备,以便指示所述网络设备根据所述第一解压缩模型参数确定CSI解压缩模型,所述CSI解压缩模型用于所述网络设备根据所述目标信道矩阵确定所述下行信道的信道质量。
在一些实施例中,所述第一接收模块1101,还被配置为接收所述网络设备发送的第二压缩模型参数;根据所述第二压缩模型参数,确定CSI压缩模型。
在一些实施例中,所述第一接收模块1101,还被配置为接收网络设备发送的第一压缩参数;根据所述第一压缩参数确定所述CSI压缩参数。
图13是根据一示例性实施例示出的一种获取信道质量的装置1300的框图,该装置可以应用于网络设备。如图13所示,该装置1300可以包括:
第二接收模块1301,被配置为接收终端设备发送的目标信道矩阵;所述目标信道矩阵为所述终端设备根据CSI压缩模型和信道状态信息CSI压缩参数,对第一信道矩阵进行压缩后得到的,所述第一信道矩阵为所述终端设备根据导频信号获取的用于表征下行信道的信道质量的矩阵;
第三矩阵获取模块1302,被配置为根据CSI解压缩模型和所述CSI压缩参数,对所述目标信道矩阵进行解压缩,得到第三信道矩阵;其中,所述CSI解压缩模型包括信道解码器,所述信道解码器包括多个子解码器,不同的子解码器对应不同的CSI压缩参数;
信道质量确定模块1303,被配置为根据所述第三信道矩阵确定下行信道的信道质量。
在一些实施例中,所述第三矩阵获取模块1302,被配置为将所述CSI压缩参数对应的子解码器作为目标子解码器;通过所述目标子解码器对所述目标信道矩阵进行解压缩,得到第四信道矩阵;根据所述第四信道矩阵确定所述第三信道矩阵。
在一些实施例中,所述CSI解压缩模型还包括CSI重构模块;所述第三矩阵获取模块 1302,被配置为将所述第四信道矩阵输入所述CSI重构模块,得到所述第三信道矩阵。
图14是根据一示例性实施例示出的一种获取信道质量的装置1300的框图,如图14所示,该装置1300还可以包括第二训练模块1304,所述第二训练模块1304,被配置为通过以下方式训练得到的所述CSI解压缩模型:
获取用于训练的第二样本信道矩阵;所述第二样本信道矩阵为终端设备根据接收到的导频信号获取的用于表征所述下行信道质量的矩阵;
根据所述第二样本信道矩阵,对第二目标网络模型进行训练,以得到所述CSI压缩模型;
其中,所述第二目标网络模型包括第二目标压缩模型和第二目标解压缩模型,所述第二目标解压缩模型的网络结构与所述CSI解压缩模型的网络结构相同,所述第二目标压缩模型包括信道编码器,所述信道编码器包括多个子编码器,不同的子编码器对应不同的CSI压缩参数。
在一些实施例中,所述第二训练模块1304,被配置为循环执行第二模型训练步骤,直至根据所述第二样本信道矩阵和第二预测信道矩阵确定训练后的第二目标网络模型满足第二预设停止迭代条件,将训练后的第二目标网络模型中的第二目标解压缩模型作为所述CSI解压缩模型;所述第二预测信道矩阵为所述第二样本信道矩阵输入所述第二目标网络模型后输出的矩阵;
所述第二模型训练步骤包括:
将所述第二样本信道矩阵输入所述第二目标压缩模型,通过多个子编码器对所述第二样本信道矩阵进行压缩后,得到第二目标样本信道矩阵;
将所述第二目标样本信道矩阵输入所述第二目标解压缩模型,通过多个子解码器对所述第二目标样本信道矩阵进行解压缩后,得到第二预测信道矩阵;
在根据所述第二样本信道矩阵和所述第二预测信道矩阵确定所述第二目标网络模型不满足所述第二预设停止迭代条件的情况下,根据所述第二样本信道矩阵和所述第二预测信道矩阵确定第二损失值,根据所述第二损失值更新所述第二目标网络模型的参数,得到训练后的第二目标网络模型,并将训练后的第二目标网络模型作为新的第二目标网络模型。
图15是根据一示例性实施例示出的一种获取信道质量的装置1300的框图,如图15所示,该装置1300还可以包括:
第二发送模块1305,被配置为获取训练后的第二目标网络模型中的第二目标压缩模型对应的第二压缩模型参数;将所述第二压缩模型参数发送至所述终端设备,以便指示所述终端设备根据所述第二压缩模型参数确定CSI压缩模型,所述CSI压缩模型用于所述终端设备根据所述第一信道矩阵获取目标信道矩阵。
在一些实施例中,所述第二接收模块1301,还被配置为接收所述终端设备发送的第一解压缩模型参数;根据所述第一解压缩模型参数,确定CSI解压缩模型。
在一些实施例中,所述第二发送模块1305,被配置为根据所述CSI压缩参数确定第一压缩参数;将所述第一压缩参数发送至所述终端设备。
关于上述实施例中的装置,其中各个模块执行操作的具体方式已经在有关该方法的实施例中进行了详细描述,此处将不做详细阐述说明。
图16是根据一示例性实施例示出的一种获取信道质量的装置的框图。该获取信道质量的装置2000可以是图1所示通信系统中的终端设备,也可以是该通信系统中的网络设备。
参照图16,该装置2000可以包括以下一个或多个组件:处理组件2002,存储器2004,以及通信组件2006。
处理组件2002可以控制装置2000的整体操作,诸如与显示,电话呼叫,数据通信, 相机操作和记录操作相关联的操作。处理组件2002可以包括一个或多个处理器2020来执行指令,以完成上述的获取信道质量的方法的全部或部分步骤。此外,处理组件2002可以包括一个或多个模块,便于处理组件2002和其他组件之间的交互。例如,处理组件2002可以包括多媒体模块,以方便多媒体组件和处理组件2002之间的交互。
存储器2004被配置为存储各种类型的数据以支持在装置2000的操作。这些数据的示例包括用于在装置2000上操作的任何应用程序或方法的指令,联系人数据,电话簿数据,消息,图片,视频等。存储器2004可以由任何类型的易失性或非易失性存储设备或者它们的组合实现,如静态随机存取存储器(SRAM),电可擦除可编程只读存储器(EEPROM),可擦除可编程只读存储器(EPROM),可编程只读存储器(PROM),只读存储器(ROM),磁存储器,快闪存储器,磁盘或光盘。
通信组件2006被配置为便于装置2000和其他设备之间有线或无线方式的通信。装置2000可以接入基于通信标准的无线网络,如WiFi,2G、3G、4G、5G或6G等通信技术,或它们的组合。在一个示例性实施例中,通信组件2006经由广播信道接收来自外部广播管理系统的广播信号或广播相关信息。在一个示例性实施例中,所述通信组件2006还包括近场通信(NFC)模块,以促进短程通信。例如,在NFC模块可基于射频识别(RFID)技术,红外数据协会(IrDA)技术,超宽带(UWB)技术,蓝牙(BT)技术和其他技术来实现。
在示例性实施例中,装置2000可以被一个或多个应用专用集成电路(ASIC)、数字信号处理器(DSP)、数字信号处理设备(DSPD)、可编程逻辑器件(PLD)、现场可编程门阵列(FPGA)、控制器、微控制器、微处理器或其他电子元件实现,用于执行上述获取信道质量的方法。
上述装置2000可以是独立的电子设备,也可以是独立电子设备的一部分,例如在一种实施例中,该电子设备可以是集成电路(Integrated Circuit,IC)或芯片,其中该集成电路可以是一个IC,也可以是多个IC的集合;该芯片可以包括但不限于以下种类:GPU(Graphics Processing Unit,图形处理器)、CPU(Central Processing Unit,中央处理器)、FPGA(Field Programmable Gate Array,可编程逻辑阵列)、DSP(Digital Signal Processor,数字信号处理器)、ASIC(Application Specific Integrated Circuit,专用集成电路)、SOC(System on Chip,SoC,片上系统或系统级芯片)等。上述的集成电路或芯片中可以用于执行可执行指令(或代码),以实现上述获取信道质量的方法。其中该可执行指令可以存储在该集成电路或芯片中,也可以从其他的装置或设备获取,例如该集成电路或芯片中可以包括处理器、存储器,以及用于与其他的装置通信的接口。该可执行指令可以存储于该处理器中,当该可执行指令被处理器执行时实现上述获取信道质量的方法;或者,该集成电路或芯片可以通过该接口接收可执行指令并传输给该处理器执行,以实现上述获取信道质量的方法。
在示例性实施例中,本公开还提供了一种计算机可读存储介质,其上存储有计算机程序指令,该程序指令被处理器执行时实现本公开提供的获取信道质量的方法的步骤。示例地,该计算机可读存储介质可以是一种包括指令的非临时性计算机可读存储介质,例如,可以是包括指令的上述存储器2004,上述指令可由装置2000的处理器2020执行以完成上述获取信道质量的方法。例如,所述非临时性计算机可读存储介质可以是ROM、随机存取存储器(RAM)、CD-ROM、磁带、软盘和光数据存储设备等。
在另一示例性实施例中,还提供一种计算机程序产品,该计算机程序产品包含能够由可编程的装置执行的计算机程序,该计算机程序具有当由该可编程的装置执行时用于执行上述获取信道质量的方法的代码部分。
本领域技术人员在考虑说明书及实践本公开后,将容易想到本公开的其它实施方案。 本申请旨在涵盖本公开的任何变型、用途或者适应性变化,这些变型、用途或者适应性变化遵循本公开的一般性原理并包括本公开未公开的本技术领域中的公知常识或惯用技术手段。说明书和实施例仅被视为示例性的,本公开的真正范围和精神由下面的权利要求指出。
应当理解的是,本公开并不局限于上面已经描述并在附图中示出的精确结构,并且可以在不脱离其范围进行各种修改和改变。本公开的范围仅由所附的权利要求来限制。

Claims (26)

  1. 一种获取信道质量的方法,其特征在于,应用于终端设备,所述方法包括:
    接收网络设备通过下行信道发送的导频信号;
    根据所述导频信号获取第一信道矩阵;所述第一信道矩阵用于表征所述下行信道的信道质量;
    根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对所述第一信道矩阵进行压缩,得到压缩后的目标信道矩阵;其中,所述CSI压缩模型包括信道编码器,所述信道编码器包括多个子编码器;不同的子编码器对应不同的CSI压缩参数;
    将所述目标信道矩阵发送至所述网络设备,以便所述网络设备根据所述目标信道矩阵确定所述下行信道的信道质量。
  2. 根据权利要求1所述的方法,其特征在于,所述根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对所述第一信道矩阵进行压缩,得到目标信道矩阵包括:
    将所述CSI压缩参数对应的子编码器作为第一目标子编码器;
    通过所述第一目标子编码器对所述第一信道矩阵进行压缩,得到目标信道矩阵。
  3. 根据权利要求1所述的方法,其特征在于,所述CSI压缩模型还包括特征转换器;所述根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对所述第一信道矩阵进行压缩,得到压缩后的目标信道矩阵包括:
    将所述第一信道矩阵输入所述特征转换器,对所述第一信道矩阵进行关键特征提取,得到表征CSI关键特征的第二信道矩阵;
    根据所述CSI压缩参数和所述信道编码器,对所述第二信道矩阵进行压缩,得到所述目标信道矩阵。
  4. 根据权利要求3所述的方法,其特征在于,所述根据所述CSI压缩参数和所述信道编码器,对所述第二信道矩阵进行压缩,得到所述目标信道矩阵包括:
    将所述CSI压缩参数对应的子编码器作为第二目标子编码器;
    通过所述第二目标子编码器对所述第二信道矩阵进行压缩,得到目标信道矩阵。
  5. 根据权利要求3所述的方法,其特征在于,所述特征转换器包括特征提取网络、注意力机制网络和特征还原网络;所述将所述第一信道矩阵输入所述特征转换器,对第一信道矩阵进行关键特征提取,得到表征CSI关键特征的第二信道矩阵包括:
    将所述第一信道矩阵输入所述特征提取网络,得到多个第一特征图;
    将多个所述第一特征图输入所述注意力机制网络,得到第二特征图;所述第二特征图包括所述第一特征图中的关键特征信息;
    将所述第二特征图输入所述特征还原网络,得到第二信道矩阵。
  6. 根据权利要求5所述的方法,其特征在于,所述将多个所述第一特征图输入所述注意力机制网络,得到第二特征图包括:
    通过所述注意力机制网络对多个所述第一特征图进行最大池化操作,得到最大池化特征图;
    通过所述注意力机制网络对多个所述第一特征图进行平均池化操作,得到平均池化特征图;
    根据所述最大池化特征图和所述平均池化特征图,确定第二特征图。
  7. 根据权利要求6所述的方法,其特征在于,所述注意力机制网络包括融合子网络;所述根据所述最大池化特征图和所述平均池化特征图,确定第二特征图包括:
    将所述最大池化特征图和所述平均池化特征图输入所述融合子网络,得到融合后的融合特征图;
    根据所述融合特征图和所述第一特征图,确定第二特征图。
  8. 根据权利要求1所述的方法,其特征在于,所述根据所述导频信号获取第一信道矩阵包括:
    根据所述导频信号,测量得到空域信道矩阵;
    通过离散傅里叶变换,将所述空域信道矩阵变换为角度时延域信道矩阵;
    根据所述角度时延域信道矩阵,确定所述第一信道矩阵。
  9. 根据权利要求1至8中任一项所述的方法,其特征在于,所述CSI压缩模型是通过以下方式训练得到的:
    获取用于训练的第一样本信道矩阵;所述第一样本信道矩阵为终端设备根据接收到的导频信号获取的用于表征所述下行信道质量的矩阵;
    根据所述第一样本信道矩阵,对第一目标网络模型进行训练,以得到所述CSI压缩模型;
    其中,所述第一目标网络模型包括第一目标压缩模型和第一目标解压缩模型,所述第一目标压缩模型的网络结构与所述CSI压缩模型的网络结构相同,所述第一目标解压缩模型包括信道解码器,所述信道解码器包括多个子解码器,不同的子解码器对应不同的CSI压缩参数。
  10. 根据权利要求9所述的方法,其特征在于,所述根据所述第一样本信道矩阵,对第一目标网络模型进行训练,包括:
    循环执行第一模型训练步骤,直至根据所述第一样本信道矩阵和第一预测信道矩阵确定训练后的第一目标网络模型满足第一预设停止迭代条件,将训练后的第一目标网络模型中的第一目标压缩模型作为所述CSI压缩模型;所述第一预测信道矩阵为所述第一样本信道矩阵输入所述第一目标网络模型后输出的矩阵;
    所述第一模型训练步骤包括:
    将所述第一样本信道矩阵输入所述第一目标压缩模型,通过多个子编码器对所述第一样本信道矩阵进行压缩后,得到第一目标样本信道矩阵;
    将所述第一目标样本信道矩阵输入所述第一目标解压缩模型,通过多个子解码器对所述第一目标样本信道矩阵进行解压缩后,得到第一预测信道矩阵;
    在根据所述第一样本信道矩阵和所述第一预测信道矩阵确定所述第一目标网络模型不满足所述第一预设停止迭代条件的情况下,根据所述第一样本信道矩阵和所述第一预测信道矩阵确定第一损失值,根据所述第一损失值更新所述第一目标网络模型的参数,得到训练后的第一目标网络模型,并将训练后的第一目标网络模型作为新的第一目标网络模型。
  11. 根据权利要求9所述的方法,其特征在于,所述方法还包括:
    获取训练后的第一目标网络模型中的第一目标解压缩模型对应的第一解压缩模型参数;
    将所述第一解压缩模型参数发送至所述网络设备,以便指示所述网络设备根据所述第一解压缩模型参数确定信道状态信息CSI解压缩模型,所述CSI解压缩模型用于所述网络设备根据所述目标信道矩阵确定所述下行信道的信道质量。
  12. 根据权利要求1至8中任一项所述的方法,其特征在于,所述方法还包括:
    接收所述网络设备发送的第二压缩模型参数;
    根据所述第二压缩模型参数,确定CSI压缩模型。
  13. 根据权利要求1至8中任一项所述的方法,其特征在于,所述方法还包括:
    接收网络设备发送的第一压缩参数;
    根据所述第一压缩参数确定所述CSI压缩参数。
  14. 一种获取信道质量的方法,其特征在于,应用于网络设备,所述方法包括:
    接收终端设备发送的目标信道矩阵;所述目标信道矩阵为所述终端设备根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对第一信道矩阵进行压缩后得到的,所述第一信道矩阵为所述终端设备根据导频信号获取的用于表征下行信道的信道质量的矩阵;
    根据信道状态信息CSI解压缩模型和所述CSI压缩参数,对所述目标信道矩阵进行解压缩,得到第三信道矩阵;其中,所述CSI解压缩模型包括信道解码器,所述信道解码器包括多个子解码器,不同的子解码器对应不同的CSI压缩参数;
    根据所述第三信道矩阵确定下行信道的信道质量。
  15. 根据权利要求14所述的方法,其特征在于,所述根据信道状态信息CSI解压缩模型和所述CSI压缩参数,对所述目标信道矩阵进行解压缩,得到第三信道矩阵包括:
    将所述CSI压缩参数对应的子解码器作为目标子解码器;
    通过所述目标子解码器对所述目标信道矩阵进行解压缩,得到第四信道矩阵;
    根据所述第四信道矩阵确定所述第三信道矩阵。
  16. 根据权利要求15所述的方法,其特征在于,所述CSI解压缩模型还包括CSI重构模块;所述根据所述第四信道矩阵确定所述第三信道矩阵包括:
    将所述第四信道矩阵输入所述CSI重构模块,得到所述第三信道矩阵。
  17. 根据权利要求14至16中任一项所述的方法,其特征在于,所述CSI解压缩模型是通过以下方式训练得到的:
    获取用于训练的第二样本信道矩阵;所述第二样本信道矩阵为终端设备根据接收到的导频信号获取的用于表征所述下行信道质量的矩阵;
    根据所述第二样本信道矩阵,对第二目标网络模型进行训练,以得到所述CSI解压缩模型;
    其中,所述第二目标网络模型包括第二目标压缩模型和第二目标解压缩模型,所述第二目标解压缩模型的网络结构与所述CSI解压缩模型的网络结构相同,所述第二目标压缩模型包括信道编码器,所述信道编码器包括多个子编码器,不同的子编码器对应不同的CSI压缩参数。
  18. 根据权利要求17所述的方法,其特征在于,所述根据所述第二样本信道矩阵, 对第二目标网络模型进行训练包括:
    循环执行第二模型训练步骤,直至根据所述第二样本信道矩阵和第二预测信道矩阵确定训练后的第二目标网络模型满足第二预设停止迭代条件,将训练后的第二目标网络模型中的第二目标解压缩模型作为所述CSI解压缩模型;所述第二预测信道矩阵为所述第二样本信道矩阵输入所述第二目标网络模型后输出的矩阵;
    所述第二模型训练步骤包括:
    将所述第二样本信道矩阵输入所述第二目标压缩模型,通过多个子编码器对所述第二样本信道矩阵进行压缩后,得到第二目标样本信道矩阵;
    将所述第二目标样本信道矩阵输入所述第二目标解压缩模型,通过多个子解码器对所述第二目标样本信道矩阵进行解压缩后,得到第二预测信道矩阵;
    在根据所述第二样本信道矩阵和所述第二预测信道矩阵确定所述第二目标网络模型不满足所述第二预设停止迭代条件的情况下,根据所述第二样本信道矩阵和所述第二预测信道矩阵确定第二损失值,根据所述第二损失值更新所述第二目标网络模型的参数,得到训练后的第二目标网络模型,并将训练后的第二目标网络模型作为新的第二目标网络模型。
  19. 根据权利要求17所述的方法,其特征在于,所述方法还包括:
    获取训练后的第二目标网络模型中的第二目标压缩模型对应的第二压缩模型参数;
    将所述第二压缩模型参数发送至所述终端设备,以便指示所述终端设备根据所述第二压缩模型参数确定CSI压缩模型,所述CSI压缩模型用于所述终端设备根据所述第一信道矩阵获取目标信道矩阵。
  20. 根据权利要求14至16中任一项所述的方法,其特征在于,所述方法还包括:
    接收所述终端设备发送的第一解压缩模型参数;
    根据所述第一解压缩模型参数,确定CSI解压缩模型。
  21. 根据权利要求14至16中任一项所述的方法,其特征在于,所述方法还包括:
    根据所述CSI压缩参数确定第一压缩参数;
    将所述第一压缩参数发送至所述终端设备。
  22. 一种获取信道质量的装置,其特征在于,应用于终端设备,所述装置包括:
    第一接收模块,被配置为接收网络设备通过下行信道发送的导频信号;
    第一矩阵获取模块,被配置为根据所述导频信号获取第一信道矩阵;所述第一信道矩阵用于表征所述下行信道的信道质量;
    目标矩阵获取模块,被配置为根据信道状态信息CSI压缩模型和信道状态信息CSI压缩参数,对所述第一信道矩阵进行压缩,得到压缩后的目标信道矩阵;其中,所述CSI压缩模型包括信道编码器,所述信道编码器包括多个子编码器;不同的子编码器对应不同的CSI压缩参数;
    第一发送模块,被配置为将所述目标信道矩阵发送至所述网络设备,以便所述网络设备根据所述目标信道矩阵确定所述下行信道的信道质量。
  23. 一种获取信道质量的装置,其特征在于,应用于网络设备,所述装置包括:
    第二接收模块,被配置为接收终端设备发送的目标信道矩阵;所述目标信道矩阵为所述终端设备根据CSI压缩模型和信道状态信息CSI压缩参数,对第一信道矩阵进行压 缩后得到的,所述第一信道矩阵为所述终端设备根据导频信号获取的用于表征下行信道的信道质量的矩阵;
    第三矩阵获取模块,被配置为根据信道状态信息CSI解压缩模型和所述CSI压缩参数,对所述目标信道矩阵进行解压缩,得到第三信道矩阵;其中,所述CSI解压缩模型包括信道解码器,所述信道解码器包括多个子解码器,不同的子解码器对应不同的CSI压缩参数;
    信道质量确定模块,被配置为根据所述第三信道矩阵确定下行信道的信道质量。
  24. 一种获取信道质量的装置,其特征在于,所述装置包括:
    处理器;
    用于存储处理器可执行指令的存储器;
    其中,所述处理器被配置为执行权利要求1至13中任一项所述方法的步骤,或者,所述处理器被配置为执行权利要求14至21中任一项所述方法的步骤。
  25. 一种计算机可读存储介质,其上存储有计算机程序指令,其特征在于,所述计算机程序指令被处理器执行时实现权利要求1至13中任一项所述方法的步骤,或者,所述计算机程序指令被处理器执行时实现权利要求14至21中任一项所述方法的步骤。
  26. 一种芯片,其特征在于,包括处理器和接口;所述处理器用于读取指令以执行权利要求1至13中任一项所述方法的步骤,或者,所述处理器用于读取指令以执行权利要求14至21中任一项所述方法的步骤。
PCT/CN2022/093972 2022-05-19 2022-05-19 获取信道质量的方法、装置、存储介质和芯片 WO2023221061A1 (zh)

Priority Applications (1)

Application Number Priority Date Filing Date Title
PCT/CN2022/093972 WO2023221061A1 (zh) 2022-05-19 2022-05-19 获取信道质量的方法、装置、存储介质和芯片

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
PCT/CN2022/093972 WO2023221061A1 (zh) 2022-05-19 2022-05-19 获取信道质量的方法、装置、存储介质和芯片

Publications (1)

Publication Number Publication Date
WO2023221061A1 true WO2023221061A1 (zh) 2023-11-23

Family

ID=88834319

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/093972 WO2023221061A1 (zh) 2022-05-19 2022-05-19 获取信道质量的方法、装置、存储介质和芯片

Country Status (1)

Country Link
WO (1) WO2023221061A1 (zh)

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117938283A (zh) * 2024-03-21 2024-04-26 深圳市必联电子有限公司 无线路由设备的质量检测方法、装置、设备及介质

Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN110350958A (zh) * 2019-06-13 2019-10-18 东南大学 一种基于神经网络的大规模mimo的csi多倍率压缩反馈方法
CN111277360A (zh) * 2019-01-11 2020-06-12 维沃移动通信有限公司 信道状态信息csi报告的传输方法、终端及网络设备
CN112737985A (zh) * 2020-12-25 2021-04-30 东南大学 基于深度学习的大规模mimo信道联合估计和反馈方法
WO2021142605A1 (zh) * 2020-01-14 2021-07-22 华为技术有限公司 用于信道测量的方法和装置
WO2022073160A1 (zh) * 2020-10-07 2022-04-14 浙江大学 编码方法、解码方法、编码器、解码器以及存储介质

Patent Citations (5)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN111277360A (zh) * 2019-01-11 2020-06-12 维沃移动通信有限公司 信道状态信息csi报告的传输方法、终端及网络设备
CN110350958A (zh) * 2019-06-13 2019-10-18 东南大学 一种基于神经网络的大规模mimo的csi多倍率压缩反馈方法
WO2021142605A1 (zh) * 2020-01-14 2021-07-22 华为技术有限公司 用于信道测量的方法和装置
WO2022073160A1 (zh) * 2020-10-07 2022-04-14 浙江大学 编码方法、解码方法、编码器、解码器以及存储介质
CN112737985A (zh) * 2020-12-25 2021-04-30 东南大学 基于深度学习的大规模mimo信道联合估计和反馈方法

Cited By (1)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN117938283A (zh) * 2024-03-21 2024-04-26 深圳市必联电子有限公司 无线路由设备的质量检测方法、装置、设备及介质

Similar Documents

Publication Publication Date Title
US20220338189A1 (en) Method and apparatus for support of machine learning or artificial intelligence techniques for csi feedback in fdd mimo systems
WO2021217519A1 (zh) 用于调整神经网络的方法和装置
WO2021108940A1 (en) Channel state information feedback
US20230246695A1 (en) Terminal and base station of wireless communication system, and methods executed by terminal and base station
CN114616762A (zh) 用于传输信道状态信息的方法和设备
US20220060917A1 (en) Online training and augmentation of neural networks for channel state feedback
WO2023221061A1 (zh) 获取信道质量的方法、装置、存储介质和芯片
CN108282238B (zh) 一种基于网络编码技术的中继选择方法及装置
US9479945B2 (en) Determination of network parameters in mobile communication networks
JP2024509502A (ja) ニューラルネットワークアーキテクチャのための符号化技法
US20230136416A1 (en) Neural network obtaining method and apparatus
WO2022012256A1 (zh) 通信的方法及通信装置
CN114492784A (zh) 神经网络的测试方法和装置
CN115442847A (zh) 提高链路自适应
CN115250485A (zh) 模型分发的方法和装置
WO2023236143A1 (zh) 信息收发方法与装置
WO2024046215A1 (zh) 一种通信方法及装置
WO2024134636A1 (en) Data encoding
WO2024045193A1 (zh) 波束上报方法、装置、终端、网络设备和介质
Maudet et al. Practical evaluation of Wi-Fi HaLow performance
WO2023208082A1 (zh) 一种通信方法、装置、芯片及模组设备
WO2023133886A1 (zh) 信道信息反馈的方法、发端设备和收端设备
WO2024032775A1 (zh) 量化方法和装置
WO2024131900A1 (zh) 一种通信的方法和通信装置
WO2024008004A1 (zh) 一种通信方法及装置

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: 22942101

Country of ref document: EP

Kind code of ref document: A1