Channel prediction method and device, communication equipment and computer readable storage medium

Abstract

The invention discloses a channel prediction method and device, communication equipment and a computer readable storage medium, and the method comprises the steps that terminal equipment receives a reference signal sent by network equipment, and determines a first channel prediction moment number based on the received reference signal; and the terminal device sends the first channel prediction time to the network device.

Description

Technical Field

[0001] This application relates to the field of wireless communication technologies, and in particular, to a channel prediction method and apparatus, a communication device, and a computer-readable storage medium. Background Art

[0002] In a wireless communication system, a terminal device feeds back downlink channel quality to a network device through a Physical Uplink Shared Channel (PUSCH). The specific process is as follows: The terminal device receives a reference signal sent by the network device, estimates Channel State Information (CSI) based on the reference signal, and then feeds back the estimated CSI to the network device through the PUSCH. The CSI characterizes the downlink channel quality. This process is also the CSI feedback process.

[0003] However, the current CSI feedback process does not consider the mobile scenario. In the mobile scenario, the demand for channel prediction is increasing day by day. Therefore, the CSI feedback process needs to introduce channel prediction (i.e., CSI prediction) technology. As the moving speed of the terminal device varies, the number of channel prediction moments (simply referred to as the number of channel prediction moments) also changes. How to negotiate the number of channel prediction moments is crucial for the CSI feedback process. Summary of the Invention

[0004] Embodiments of this application provide a channel prediction method and apparatus, a communication device, a chip, and a computer-readable storage medium.

[0005] The channel prediction method provided by the embodiments of this application includes:

[0006] The terminal device receives a reference signal sent by the network device and determines a first number of channel prediction moments based on the received reference signal;

[0007] The terminal device sends the first number of channel prediction moments to the network device.

[0008] The channel prediction method provided by the embodiments of this application includes:

[0009] The network device sends a reference signal to the terminal device, and the reference signal is used for the terminal device to determine a first number of channel prediction moments;

[0010] The network device receives the first number of channel prediction moments sent by the terminal device.

[0011] The channel prediction apparatus provided by the embodiments of this application is applied to a terminal device, and the apparatus includes:

[0012] The first communication unit is configured to receive a reference signal sent by a network device;

[0013] The first processing unit is configured to determine the number of first channel prediction time instances based on the received reference signal;

[0014] The first communication unit is further configured to send the number of first channel prediction time instances to the network device.

[0015] The channel prediction apparatus provided by an embodiment of the present application is characterized in that it is applied to a network device, and the apparatus includes:

[0016] The second communication unit is configured to send a reference signal to a terminal device, where the reference signal is used for the terminal device to determine the number of first channel prediction time instances; and receive the number of first channel prediction time instances sent by the terminal device.

[0017] The communication device provided by an embodiment of the present application includes: a processor and a memory. The memory is used to store a computer program, and the processor is used to call and run the computer program stored in the memory to execute any one of the above channel prediction methods.

[0018] The chip provided by an embodiment of the present application includes: a processor, which is used to call and run a computer program from a memory, so that a device installed with the chip executes any one of the above methods.

[0019] The computer-readable storage medium provided by an embodiment of the present application is used to store a computer program, and the computer program enables a computer to execute any one of the above methods.

[0020] In the technical solution of the embodiment of the present application, the terminal device determines the number of first channel prediction time instances based on the received reference signal from the network device, and sends the number of first channel prediction time instances to the network device, realizing the negotiation of the number of channel prediction time instances between the terminal device and the network device, so that the terminal device and the network device have the same understanding of the number of channel prediction time instances. In this way, when CSI feedback is performed between the terminal device and the network device, the correct CSI feedback process can be performed based on the mutually understood number of channel prediction time instances. BRIEF DESCRIPTION OF THE DRAWINGS

[0021] Figure 1 is a schematic diagram of an application scenario of an embodiment of the present application;

[0022] Figure 2 is a schematic diagram of an AI-based CSI feedback framework;

[0023] Figure 3 is a flowchart of the channel prediction method provided by an embodiment of the present application Figure 1 ;

[0024] Figure 4 It is a flowchart of the channel prediction method provided by the embodiments of the present application Figure 2 ;

[0025] Figure 5 It is a flowchart of the channel prediction method provided by the embodiments of the present application Figure 3 ;

[0026] Figure 6 It is a flowchart of the channel prediction method provided by the embodiments of the present application Figure 4 ;

[0027] Figure 7 It is a flowchart of the channel prediction method provided by the embodiments of the present application Figure 5 ;

[0028] Figure 8 It is a flowchart of the channel prediction method provided by the embodiments of the present application Figure 6 ;

[0029] Figure 9 It is a schematic diagram of the integration of CSI prediction and CSI compression feedback provided by the embodiments of the present application;

[0030] Figure 10 It is a schematic diagram of the correlation extraction model provided by the embodiments of the present application;

[0031] Figure 11 It is a schematic diagram of the structural composition of the channel prediction device provided by the embodiments of the present application Figure 1 ;

[0032] Figure 12 It is a schematic diagram of the structural composition of the channel prediction device provided by the embodiments of the present application Figure 2 ;

[0033] Figure 13 It is a schematic structural diagram of a communication device provided by the embodiments of the present application;

[0034] Figure 14 It is a schematic structural diagram of the chip provided by the embodiments of the present application. Detailed implementation manners

[0035] Next, the technical solutions in the embodiments of the present application will be described with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are some, rather than all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art based on the embodiments in the present application without creative efforts shall fall within the protection scope of the present application.

[0036] Figure 1 It is a schematic diagram of an application scenario of the embodiments of the present application.

[0037] As shown Figure 1 in FIG. 1, the communication system 100 may include a terminal device 110 and a network device 120. The network device 120 may communicate with the terminal device 110 via an air interface. Multi-service transmission is supported between the terminal device 110 and the network device 120.

[0038] It should be understood that the embodiments of the present application are only exemplified by the communication system 100, but the embodiments of the present application are not limited thereto. That is to say, the technical solutions of the embodiments of the present application can be applied to various communication systems, such as: 5G communication systems (also known as New Radio (NR) communication systems), or future communication systems, etc.

[0039] In Figure 1 the communication system 100 shown in FIG. 1, the network device 120 may be an access network device that communicates with the terminal device 110. The access network device may provide communication coverage for a specific geographical area and may communicate with the terminal device 110 (such as a UE) located within the coverage area.

[0040] The network device 120 may be a base station (gNB) in an NR system or a network device in a future evolved Public Land Mobile Network (PLMN), etc.

[0041] The terminal device 110 may be any terminal device. For example, the terminal device 110 may refer to an access terminal, a user equipment (UE), a user unit, a user station, a mobile station, a mobile device, a remote station, a remote terminal, a mobile device, a user terminal, a terminal, a wireless communication device, a user agent, or a user device. The access terminal may be a cellular phone, a cordless phone, a Session Initiation Protocol (SIP) phone, an IoT device, a satellite handheld terminal, a Wireless Local Loop (WLL) station, a Personal Digital Assistant (PDA), a handheld device with wireless communication capabilities, a computing device, or other processing devices connected to a wireless modem, a vehicle-mounted device, a wearable device, a terminal device in a 5G network, or a terminal device in a future evolved network, etc.

[0042] Figure 1 Exemplarily, one base station and two terminal devices are shown. Optionally, the wireless communication system 100 may include multiple base station devices, and each base station's coverage area may include other numbers of terminal devices. The embodiments of the present application do not limit this.

[0043] It should be noted thatFigure 1 The system applicable to the present application is only schematically shown by way of example. Of course, the method shown in the embodiments of the present application can also be applicable to other systems. In addition, the terms "system" and "network" in this document are often used interchangeably. The term "and / or" in this document is merely a description of the association relationship of associated objects, indicating that three relationships can exist. For example, A and / or B can represent: A exists alone, A and B exist simultaneously, and B exists alone. In addition, the character " / " in this document generally represents an "or" relationship between the front and rear associated objects. It should also be understood that the "indication" mentioned in the embodiments of the present application can be a direct indication, an indirect indication, or a representation of an association relationship. For example, A indicates B, which can mean that A directly indicates B. For example, B can be obtained through A; it can also mean that A indirectly indicates B. For example, A indicates C, and B can be obtained through C; it can also mean that there is an association relationship between A and B. It should also be understood that the "correspondence" mentioned in the embodiments of the present application can represent a direct or indirect correspondence relationship between the two, can also represent an association relationship between the two, or can be an indication and being indicated, configuration and being configured, etc. relationships. It should also be understood that the "predefined" or "predefined rule" mentioned in the embodiments of the present application can be implemented by pre-saving corresponding codes, tables, or other means that can be used to indicate relevant information in a device (for example, including a terminal device and a network device). The present application does not limit its specific implementation manner. For example, predefined can refer to being defined in a protocol. It should also be understood that in the embodiments of the present application, the "protocol" can refer to a standard protocol in the communication field. For example, it can include the NR protocol and related protocols applied to future communication systems. The present application does not limit this.

[0044] The neural network model can be used for CSI feedback between the terminal device and the network device, which is called AI-based CSI feedback. As Figure 2 shown, the AI-based CSI feedback framework deploys an encoder on the terminal device side and a decoder on the network device side. Among them, both the encoder and the decoder can be implemented by a neural network model. In the AI-based CSI feedback process, after the terminal device estimates the CSI, it compresses the CSI through the encoder to obtain a CSI bitstream (such as P bits, where P is a positive integer), and feeds the CSI bitstream back to the network device through the PUSCH; the network device decodes (or decompresses) the received CSI bitstream through the decoder to obtain CSI', and the network device can determine the channel precoding matrix according to this CSI.

[0045] Figure 2The shown framework performs compression and feedback based on CSI at a single moment. In this case, the data volume of the CSI bitstream is fixed (i.e., the value of P is fixed).

[0046] In mobile communication, the demand for channel prediction in mobile scenarios is increasing day by day. Therefore, the CSI feedback process needs to introduce channel prediction (i.e., CSI prediction) technology. Therefore, the technical solution of the embodiment of the present application proposes an enhanced CSI feedback scheme that integrates CSI prediction, CSI compression, and CSI feedback. In this enhanced CSI feedback scheme, the number of channel prediction moments is crucial. As the moving speed of the terminal device varies, the number of channel prediction moments will also change. For example, if the moving speed of the terminal device is slow, the channel change is also slow, and the number of channel prediction moments will be larger; if the moving speed of the terminal device is fast, the channel change is also fast, and the number of channel prediction moments will be smaller. The number of channel prediction moments can also be referred to as the channel prediction duration. As the channel prediction duration varies, the data volume for compressing and feedbacking the CSI within this channel prediction duration will also be different.

[0047] To facilitate understanding of the technical solution of the embodiment of the present application, the technical solution of the present application is described in detail below through specific embodiments. The above related technologies can be arbitrarily combined with the technical solution of the embodiment of the present application, and all of them fall within the protection scope of the embodiment of the present application. The embodiment of the present application includes at least some of the following content.

[0048] It should be noted that the "moment" described in the embodiment of the present application can be but is not limited to "time slot", "sub-time slot", "sub-frame", etc. That is: the "moment" can be replaced and described as "time slot" or "sub-time slot" or "sub-frame", etc.

[0049] Figure 3 is the flowchart of the channel prediction method provided by the embodiment of the present application Figure 1 as Figure 3 shown, the channel prediction method includes:

[0050] Step 301: The terminal device receives the reference signal sent by the network device and determines the first number of channel prediction moments based on the received reference signal.

[0051] In some embodiments, the reference signal sent by the network device is a Channel State Information-Reference Signal (CSI-RS).

[0052] In some embodiments, before step 301, the terminal device receives the reference signal configuration sent by the network device, and the terminal device receives the reference signal sent by the network device based on this reference signal configuration.

[0053] In the embodiments of the present application, the terminal device receives the reference signals sent by the network device at M moments respectively, and estimates the CSI at M moments according to the received reference signals, where M is a positive integer. The terminal device determines the number of first channel prediction moments based on the CSI at M moments. The number of first channel prediction moments can also be understood as the number of channel prediction moments recommended by the terminal device.

[0054] The manner in which the terminal device determines the number of first channel prediction moments based on the CSI at M moments can be but is not limited to the following manners:

[0055] Manner 1: The terminal device calculates the channel correlation of the CSI at M moments; the terminal device determines the number of first channel prediction moments based on the channel correlation of the CSI at M moments.

[0056] Here, the terminal device can calculate the channel correlation of the CSI at M moments through the following formula (1):

[0057]

[0058] where ρ represents the channel correlation of the CSI at M moments; w i represents the CSI at moment i, w i+1 represents the CSI at moment i + 1, 0 ≤ i ≤ M - 1; T represents the transpose symbol.

[0059] There is a corresponding relationship between the channel correlation and the number of channel prediction moments. The terminal device can determine the corresponding number of channel prediction moments, that is, the number of first channel prediction moments, according to the channel correlation of the CSI at M moments and this corresponding relationship.

[0060] It should be noted that in the above corresponding relationship, the higher the channel correlation, the larger the number of channel prediction moments; conversely, the lower the channel correlation, the smaller the number of channel prediction moments.

[0061] Manner 2: The terminal device predicts the number of first channel prediction moments based on the CSI at M moments through a moment number prediction model.

[0062] Here, a moment number prediction model can be pre-trained. The CSI at M moments is input into the trained moment number prediction model, and the number of first channel prediction moments is output through this moment number prediction model.

[0063] It should be noted that different "CSI at M moments" can predict different "number of first channel prediction moments", and the number of first channel prediction moments is variable.

[0064] Step 302: The terminal device sends the number of first channel prediction moments to the network device.

[0065] Step 303: The terminal device performs channel prediction based on the received reference signal to obtain the CSI at multiple moments, where the number of moments included in the multiple moments is the first channel prediction number of moments.

[0066] In the embodiments of the present application, the terminal device receives the reference signals sent by the network device at M moments respectively, and estimates the CSI at M moments according to the received reference signals, where M is a positive integer. The terminal device performs channel prediction based on the CSI at M moments to obtain the CSI at N moments, where N is a positive integer. Here, the M moments are before the N moments, that is, the N moments are after the M moments. The value of N is determined based on the above first channel prediction number of moments.

[0067] The terminal device inputs the CSI at M moments into the channel prediction module, and outputs the CSI at N moments through the channel prediction module. In some embodiments, the channel prediction module can be implemented by using traditional channel prediction algorithms (such as Bayesian, interpolation and other prediction algorithms), or can also be implemented by using a neural network model (such as Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), etc.).

[0068] Step 304: The terminal device selects the first model based on the first channel prediction number of moments; the terminal device compresses the CSI at multiple moments through the second model, and extracts features from the compressed CSI through the first model to obtain the CSI bitstream.

[0069] In the embodiments of the present application, the second model is an encoder for compressing the CSI. The first model is a feature extraction model for extracting features from the compressed CSI, such as extracting time correlation from the compressed CSI.

[0070] The first models corresponding to different channel prediction numbers of moments are different, and the terminal device needs to select the corresponding first model according to the adopted channel prediction number of moments. Specifically, the terminal device selects the corresponding first model according to the adopted first channel prediction number of moments.

[0071] In an example, the first model is called a correlation extraction model, and there is a corresponding relationship between the correlation extraction model and the number of prediction moments. The terminal device selects the corresponding correlation extraction model according to the first channel prediction number of moments and the corresponding relationship.

[0072] In the embodiments of the present application, after the terminal device obtains the CSI at N moments through channel prediction, it compresses the CSI at N moments through a second model to obtain K-bit information, and extracts features from the K-bit information through a first model to obtain L-bit information, where K is a positive integer and L is a positive integer less than K; this L-bit information is the CSI bit stream to be sent to the network device.

[0073] Step 305: The terminal device sends the CSI bit stream to the network device.

[0074] In the technical solution of the embodiments of the present application, the terminal device determines the number of first channel prediction moments based on the received reference signal from the network device and sends the number of first channel prediction moments to the network device, realizing the negotiation of the number of channel prediction moments between the terminal device and the network device, so that the understanding of the number of channel prediction moments by the terminal device and the network device is consistent. In this way, when the terminal device and the network device perform CSI feedback, the correct CSI feedback process can be carried out based on the mutually understood number of channel prediction moments.

[0075] Figure 4 is a schematic flow chart of the channel prediction method provided by the embodiments of the present application Figure 2 as Figure 4 shown, the channel prediction method includes:

[0076] Step 401: The terminal device receives the reference signal sent by the network device and determines the number of first channel prediction moments based on the received reference signal.

[0077] This step can refer to the relevant description of step 301 in the foregoing Figure 3 section.

[0078] Step 402: The terminal device sends the number of first channel prediction moments to the network device.

[0079] Step 403: The terminal device receives the number of second channel prediction moments sent by the network device.

[0080] In some embodiments, the number of second channel prediction moments is determined based on the number of first channel prediction moments. Here, the network device can give the recommended number of channel prediction moments, that is, the number of second channel prediction moments, according to the system performance in combination with the number of first channel prediction moments. The number of second channel prediction moments may be different from the number of first channel prediction moments, or the number of second channel prediction moments may also be the same as the number of first channel prediction moments.

[0081] Step 404: The terminal device performs channel prediction based on the received reference signal to obtain the CSI at multiple moments, and the number of moments included in the multiple moments is the number of second channel prediction moments.

[0082] In the embodiments of the present application, the terminal device receives the reference signals sent by the network device at M moments respectively, and estimates the CSI at the M moments according to the received reference signals, where M is a positive integer. The terminal device performs channel prediction based on the CSI at the M moments to obtain the CSI at N moments, where N is a positive integer. Here, the M moments are before the N moments, that is, the N moments are after the M moments. The value of N is determined based on the above-mentioned number of second channel prediction moments.

[0083] The terminal device inputs the CSI at the M moments into the channel prediction module, and the CSI at the N moments is output through this channel prediction module. In some embodiments, the channel prediction module can be implemented by using traditional channel prediction algorithms (such as Bayesian, interpolation and other prediction algorithms), or can also be implemented by using a neural network model (such as RNN, LSTM, etc.).

[0084] Step 405: The terminal device selects a first model based on the number of second channel prediction moments; the terminal device compresses the CSI at multiple moments through a second model, and extracts features from the compressed CSI through the first model to obtain a CSI bit stream.

[0085] In the embodiments of the present application, the second model is an encoder for compressing the CSI. The first model is a feature extraction model for extracting features from the compressed CSI, such as extracting time correlation from the compressed CSI.

[0086] The first models corresponding to different numbers of channel prediction moments are different. The terminal device needs to select the corresponding first model according to the number of channel prediction moments adopted. Specifically, the terminal device selects the corresponding first model according to the number of second channel prediction moments adopted.

[0087] In an example, the first model is called a correlation extraction model, and there is a corresponding relationship between the correlation extraction model and the number of prediction moments. The terminal device selects the corresponding correlation extraction model according to the number of second channel prediction moments and this corresponding relationship.

[0088] In the embodiments of the present application, after the terminal device obtains the CSI at N moments through channel prediction, it compresses the CSI at the N moments through the second model to obtain K bits of information, and extracts features from the K bits of information through the first model to obtain L bits of information, where K is a positive integer and L is a positive integer less than K; this L bits of information is the CSI bit stream that needs to be sent to the network device.

[0089] Step 406: The terminal device sends the CSI bit stream to the network device.

[0090] In the technical solution of the embodiment of the present application, the terminal device determines the number of first channel prediction moments based on the received reference signal from the network device, and sends the number of first channel prediction moments to the network device. The network device sends the proposed number of second channel prediction moments to the terminal device, realizing the negotiation of the number of channel prediction moments between the terminal device and the network device, so that the understanding of the number of channel prediction moments between the terminal device and the network device is consistent. In this way, when CSI feedback is performed between the terminal device and the network device, the correct CSI feedback process can be carried out based on the mutually understood number of channel prediction moments.

[0091] Figure 5 is a schematic flow chart of the channel prediction method provided by the embodiment of the present application Figure 3 , such as Figure 5 shown, the channel prediction method includes:

[0092] Step 501: The network device sends a reference signal to the terminal device, and the reference signal is used for the terminal device to determine the number of first channel prediction moments.

[0093] In some embodiments, the reference signal sent by the network device is CSI-RS.

[0094] Step 502: The network device receives the number of first channel prediction moments sent by the terminal device.

[0095] Step 503: The network device selects a third model based on the number of first channel prediction moments; the network device receives the CSI bit stream sent by the terminal device; the network device analyzes the features of the CSI bit stream through the third model to obtain the compressed CSI; and decompresses the compressed CSI through the fourth model to obtain the CSI at multiple moments, and the number of moments included in the multiple moments is the number of first channel prediction moments.

[0096] In the embodiment of the present application, the third model is a feature analysis model for analyzing the features of the CSI bit stream, such as analyzing the time correlation of the CSI bit stream. The fourth model is a decoder for decompressing the compressed CSI.

[0097] The third models corresponding to different numbers of channel prediction moments are different, and the network device needs to select the corresponding third model according to the adopted number of channel prediction moments. Specifically, the network device selects the corresponding third model according to the adopted number of first channel prediction moments.

[0098] In an example, the third model is called a correlation analysis model, and there is a corresponding relationship between the correlation analysis model and the number of prediction moments. The network device selects the corresponding correlation analysis model according to the number of first channel prediction moments and the corresponding relationship.

[0099] In the embodiments of the present application, after receiving the CSI bitstream sent by the terminal device, the network device decompresses the temporal correlation from the L-bit information of the CSI bitstream through a third model to obtain K-bit information, and decompresses the K-bit information through a fourth model to obtain the CSI at N moments, where N is a positive integer, K is a positive integer, and L is a positive integer less than K.

[0100] It should be noted that Figure 5 the solution on the network device side in Figure 3 can be implemented in combination with the solution on the terminal device side in

[0101] Figure 6 is a schematic flow of the channel prediction method provided by the embodiments of the present application Figure 4 as Figure 6 shown, the channel prediction method includes:

[0102] Step 601: The network device sends a reference signal to the terminal device, and the reference signal is used for the terminal device to determine the number of first channel prediction moments.

[0103] In some embodiments, the reference signal sent by the network device is CSI-RS.

[0104] Step 602: The network device receives the number of first channel prediction moments sent by the terminal device.

[0105] Step 603: The network device sends the number of second channel prediction moments to the terminal device.

[0106] In some embodiments, the number of second channel prediction moments is determined based on the number of first channel prediction moments. Here, the network device can give a recommended number of channel prediction moments, that is, the number of second channel prediction moments, in combination with the number of first channel prediction moments according to the system performance. The number of second channel prediction moments may be different from the number of first channel prediction moments, or the number of second channel prediction moments may also be different from the number of first channel prediction moments.

[0107] Step 604: The network device selects a third model based on the number of second channel prediction moments; the network device receives the CSI bitstream sent by the terminal device; the network device performs feature analysis on the CSI bitstream through the third model to obtain the compressed CSI; and decompresses the compressed CSI through the fourth model to obtain the CSI at multiple moments, and the number of moments included in the multiple moments is the number of second channel prediction moments.

[0108] In the embodiments of the present application, the third model is a feature analysis model for performing feature analysis on the CSI bitstream, such as performing temporal correlation analysis on the CSI bitstream. The fourth model is a decoder for decompressing the compressed CSI.

[0109] The third models corresponding to different numbers of channel prediction moments are different. The network device needs to select the corresponding third model according to the number of channel prediction moments adopted. Specifically, the network device selects the corresponding third model according to the adopted second number of channel prediction moments.

[0110] In one example, the third model is called a correlation analysis model. There is a corresponding relationship between the correlation analysis model and the number of prediction moments. The network device selects the corresponding correlation analysis model according to the second number of channel prediction moments and this corresponding relationship.

[0111] In the embodiments of this application, after the network device receives the CSI bit stream sent by the terminal device, it decompresses the time correlation from the L-bit information of the CSI bit stream through the third model to obtain K-bit information, and decompresses this K-bit information through the fourth model to obtain the CSI at N moments. N is a positive integer, K is a positive integer, and L is a positive integer less than K.

[0112] It should be noted that Figure 6 the solution on the network device side can be combined with Figure 4 the solution on the terminal device side in

[0113] Figure 7 is a schematic flow of the channel prediction method provided by the embodiments of this application Figure 5 , UE corresponds to the above terminal device, and the base station corresponds to the above network device. As Figure 7 shown, the channel prediction method includes:

[0114] Step 701: The UE estimates the CSI at M moments, calculates the channel correlation of the CSI at M moments, and determines the recommended number of channel prediction moments as N according to the channel correlation of the CSI at M moments; N and M are positive integers.

[0115] Step 702: The UE feeds back the number of channel prediction moments N to the base station.

[0116] Step 703: The UE predicts the CSI at N moments according to the CSI at M moments.

[0117] Step 704: The UE selects a correlation extraction model according to the number of channel prediction moments N.

[0118] Step 705: The base station selects a correlation analysis model according to the number of channel prediction moments N.

[0119] Step 706: The UE compresses the CSI at N moments through an encoder to obtain K-bit information, where K is a positive integer.

[0120] Step 707: The UE extracts features from the K-bit information through the correlation extraction model to obtain L-bit information. The L-bit information is the CSI bit stream that needs to be fed back to the base station, where L is a positive integer less than K.

[0121] Step 708: The UE sends the CSI bit stream to the base station.

[0122] Step 709: The base station performs feature analysis on the CSI bit stream through the correlation analysis model to obtain the compressed CSI. The compressed CSI is K-bit information.

[0123] Step 710: The base station decompresses the compressed CSI through the decoder to obtain the CSI at N moments.

[0124] Step 711: The base station generates a corresponding precoding matrix according to the CSI at N moments.

[0125] Figure 8 is the process schematic of the channel prediction method provided by the embodiments of the present application Figure 6 , the UE corresponds to the above terminal device, and the base station corresponds to the above network device, as Figure 8 shown, the channel prediction method includes:

[0126] Step 801: The UE estimates the CSI at M moments, calculates the channel correlation of the CSI at M moments, and determines the recommended number of channel prediction moments as N' according to the channel correlation of the CSI at M moments; N' and M are positive integers.

[0127] Step 802: The UE feeds back the number of channel prediction moments N' to the base station.

[0128] Step 803: The base station determines the recommended number of channel prediction moments N according to the system performance and the number of channel prediction moments N'.

[0129] Step 804: The base station sends the number of channel prediction moments N to the UE.

[0130] Step 805: The UE predicts the CSI at N moments according to the CSI at M moments.

[0131] Step 806: The UE selects the correlation extraction model according to the number of channel prediction moments N.

[0132] Step 807: The base station selects the correlation analysis model according to the number of channel prediction moments N.

[0133] Step 808: The UE compresses the CSI at N moments through the encoder to obtain K-bit information, where K is a positive integer.

[0134] Step 809: The UE extracts features from the K-bit information through the correlation extraction model to obtain L-bit information. The L-bit information is the CSI bit stream to be fed back to the base station, where L is a positive integer less than K.

[0135] Step 810: The UE sends the CSI bit stream to the base station.

[0136] Step 811: The base station performs feature analysis on the CSI bit stream through the correlation analysis model to obtain the compressed CSI. The compressed CSI is K-bit information.

[0137] Step 812: The base station decompresses the compressed CSI through a decoder to obtain the CSI at N moments.

[0138] Step 813: The base station generates a corresponding precoding matrix according to the CSI at N moments.

[0139] It should be noted that the overhead of CSI feedback is L bits, and the size of the L bits is fixed. After receiving the L-bit information, a network device (such as a base station) cannot determine how many moments of CSI (i.e., the number of channel prediction moments) the terminal device (such as a UE) performs CSI feedback. Therefore, the terminal device (such as a UE) needs to notify the network device (such as a base station) of the channel prediction moment number corresponding to the L-bit information through a signaling. In some embodiments, the signaling sent by the terminal device (such as a UE) to the network device (such as a base station) carries an index corresponding to the channel prediction moment number, and the network device (such as a base station) can determine the channel prediction moment number according to the index. The protocol can define multiple (for example, R, where R is an integer greater than 1) candidate channel prediction moment numbers, each channel prediction moment number corresponds to an index, and the overhead of the index is determined according to the value of R.

[0140] The following combines Figure 9 to further illustrate the technical solution of the embodiment of the present application by way of example.

[0141] Figure 9 is a schematic diagram of the integration of CSI prediction and CSI compression feedback provided by the embodiment of the present application. As Figure 9 shown, the CSI feedback process includes six stages, namely: CSI prediction stage, CSI compression stage, feature extraction stage, CSI feedback stage, feature analysis stage, and CSI decompression stage. Among them, the CSI prediction stage, CSI compression stage, and feature extraction stage are implemented on the UE side; the feature analysis stage and CSI decompression stage are implemented on the base station side; the CSI feedback stage is implemented on both the UE side and the base station side. The following describes each stage.

[0142] The stages implemented on the UE side include:

[0143] 1) CSI prediction stage.

[0144] In the CSI prediction stage, the UE predicts the CSI at N time moments based on the CSI at M time moments, where M and N are positive integers and the N time moment is after the M time moment.

[0145] exist Figure 9 In the example, the CSI at M moments is input into the neural network model, and the CSI at N moments is output through the neural network model. The CSI at M moments are recorded as: CSI 1, CSI 2, ..., CSIM. The CSI at N moments are recorded as: CSI pre 1. CSI pre 2. …, CSI pre N.

[0146] Here, the channel time-varying situation can be inferred based on the CSI at M moments. If the channel time-varying situation is fast, the smaller the value of N is, that is, the fewer the channel prediction moments are; if the channel time-varying situation is slow, the larger the value of N is, that is, the more the channel prediction moments are. It can be understood that the value of N is variable or the value of N is not fixed.

[0147] 2) CSI compression stage.

[0148] In the CSI compression stage, the UE inputs the CSI at N moments in parallel to the encoder. The encoder compresses the CSI at N moments in parallel. The CSI at each moment can be compressed to obtain k bits of information. The CSI at N moments can be compressed to obtain N×k bits of information in total. The N×k bits of information can be recorded as K bits of information.

[0149] Here, by inputting the CSI at N moments in parallel into N encoders, the CSI compression delay can be effectively reduced. In addition, since N×k bits of information are obtained by compression, when CSI feedback is performed subsequently, feedback can be performed once instead of N times, thereby reducing the CSI feedback delay.

[0150] Here, the N encoders on the UE side are of the same model, and similarly, the N decoders on the base station side are also of the same model. That is to say, the technical solution of the embodiment of the present application can realize the compression of CSI at N moments by sharing a set of encoders and decoders. Compared with the solution of compressing CSI at N moments by using N sets of encoders and decoders, the technical solution of the embodiment of the present application does not need to train N sets of encoders and decoders, but only needs to train one set of encoders and decoders, thus effectively reducing the training overhead and the memory overhead caused by the training.

[0151] Here, by compressing the CSI at N moments, the CSI feedback overhead can be reduced.

[0152] 3) Feature extraction stage.

[0153] In the feature extraction stage, the UE inputs N×k-bit information into a correlation extraction model (such as a time correlation extraction model), performs correlation extraction (such as time correlation extraction) through the correlation extraction model, and obtains L-bit information. The L-bit information is the CSI bit stream that needs to be fed back to the base station.

[0154] Here, a correlation extraction model (such as a time correlation extraction model) is added on the UE side. Through this correlation extraction model, the N×k-bit information is further converted into L-bit information, where L ≤ N×k. In some cases, it is possible to further compress the N×k-bit information and reduce the CSI feedback overhead.

[0155] Here, according to the number of channel prediction time instances, corresponding correlation extraction models will be trained. That is, there will be multiple independent correlation extraction models, and different correlation extraction models correspond to different numbers of channel prediction time instances. The UE can select the corresponding correlation extraction model according to the number of channel prediction time instances used, so as to ensure that the size of the L-bit information output by the correlation extraction model is fixed, ensuring the air interface transmission performance.

[0156] In one example, the correlation extraction model is implemented by the superposition of a CNN dimensionality reduction network and a time correlation extraction network (which can be implemented by a Transformer network). As Figure 10 shown, the N×k-bit information is input into the CNN dimensionality reduction network, and the CNN dimensionality reduction network outputs n×k-bit information, where n is a positive integer less than N. Then, the n×k-bit information is input into the time correlation extraction network, and through the time correlation extraction network, processing related to multi-head attention and processing related to channel time domain correlation are performed, and then L-bit information is output.

[0157] 4) CSI feedback stage

[0158] In the CSI feedback stage, the UE feeds back the L-bit information (i.e., the CSI bit stream) to the base station.

[0159] 5) Feature parsing stage.

[0160] In the feature parsing stage, the base station inputs the L-bit information into a correlation parsing model (such as a time correlation parsing model), and performs correlation parsing (such as time correlation parsing) through the correlation parsing model to obtain N×k-bit information.

[0161] Here, a correlation parsing model (such as a time correlation parsing model) is added on the base station side. Through this correlation parsing model, the L-bit information is converted into N×k-bit information.

[0162] Here, according to the number of channel prediction moments, corresponding correlation analysis models will be trained, that is, there will be multiple independent correlation analysis models, and different correlation analysis models correspond to different numbers of channel prediction moments. The base station can select the corresponding correlation analysis model according to the number of channel prediction moments adopted.

[0163] 6) CSI decompression stage.

[0164] In the CSI decompression stage, the base station inputs N×k-bit information into the decoder in parallel, and the decoder decompresses the N×k-bit information in parallel to obtain the CSI at N moments. The CSI at N moments are respectively denoted as: CSI pre 1', CSI pre 2', …, CSI pre N'.

[0165] The models involved in the above scheme can be trained in the following manner:

[0166] 1) When the number of channel prediction moments (i.e., N) is 1, the CSI at 1 moment is input into the "encoder - correlation extraction model - correlation analysis model - decoder", and a set of "encoder - correlation extraction model - correlation analysis model - decoder" is trained through the CSI at 1 moment. At this time, the channel prediction moments corresponding to the correlation extraction model and the correlation analysis model are 1. After the training is completed, keep the encoder and decoder unchanged. This encoder and decoder can be used for the case where the number of channel prediction moments is 1 or greater than 1.

[0167] 2) When the number of channel prediction moments (i.e., N) is greater than 1, the CSI at N (N>1) moments is input into the "encoder - correlation extraction model - correlation analysis model - decoder", where the encoder and decoder are fixed (already trained through the above step 1), and the correlation extraction model and the correlation analysis model are trained through the CSI at N (N>1) moments. At this time, the channel prediction moments corresponding to the correlation extraction model and the correlation analysis model are N.

[0168] Through the above training process, the models corresponding to different numbers of channel prediction moments can be trained.

[0169] Figure 11 is a schematic structural composition of the channel prediction device provided by an embodiment of the present application Figure 1 , applied to a terminal device, such as Figure 11 shown, the channel prediction device includes:

[0170] The first communication unit 1101 is used to receive the reference signal sent by the network device;

[0171] The first processing unit 1102 is configured to determine the number of first channel prediction moments based on the received reference signal;

[0172] The first communication unit 1101 is further configured to send the number of first channel prediction moments to the network device.

[0173] In some embodiments, the first processing unit 1102 is further configured to perform channel prediction based on the received reference signal to obtain the CSI at multiple moments, and the number of moments included in the multiple moments is the number of first channel prediction moments.

[0174] In some embodiments, the first processing unit 1102 is further configured to select a first model based on the number of first channel prediction moments; compress the CSI at the multiple moments through a second model, and perform feature extraction on the compressed CSI through the first model to obtain a CSI bitstream;

[0175] The first communication unit 1101 is further configured to send the CSI bitstream to the network device.

[0176] In some embodiments, the first communication unit 1101 is further configured to receive the number of second channel prediction moments sent by the network device.

[0177] In some embodiments, the number of second channel prediction moments is determined based on the number of first channel prediction moments.

[0178] In some embodiments, the first processing unit 1102 is further configured to perform channel prediction based on the received reference signal to obtain the CSI at multiple moments, and the number of moments included in the multiple moments is the number of second channel prediction moments.

[0179] In some embodiments, the first processing unit 1102 is further configured to select a first model based on the number of second channel prediction moments; compress the CSI at the multiple moments through a second model, and perform feature extraction on the compressed CSI through the first model to obtain a CSI bitstream;

[0180] The first communication unit 1101 is further configured to send the CSI bitstream to the network device.

[0181] Those skilled in the art should understand that Figure 11 The implementation functions of the units in the illustrated channel prediction device can be understood with reference to the relevant descriptions of the foregoing method. Figure 11 The functions of the units in the illustrated channel prediction device can be implemented by a program running on a processor or by specific logic circuits.

[0182] Figure 12 This is a schematic diagram of the structural composition of the channel prediction device provided by an embodiment of the present application Figure 2 which is applied to a network device, such as Figure 12 as shown, the channel prediction device includes:

[0183] A second communication unit 1201, configured to send a reference signal to a terminal device, where the reference signal is used for the terminal device to determine the number of first channel prediction times; and receive the number of first channel prediction times sent by the terminal device.

[0184] In some embodiments, the channel prediction device further includes: a second processing unit 1202.

[0185] In some embodiments, the second processing unit 1202 is further configured to select a third model based on the number of first channel prediction times;

[0186] The second communication unit 1201 is further configured to receive a CSI bit stream sent by the terminal device;

[0187] The second processing unit 1202 is further configured to perform feature analysis on the CSI bit stream through the third model to obtain compressed CSI; and decompress the compressed CSI through a fourth model to obtain CSI at multiple times, where the number of times included in the multiple times is the number of first channel prediction times.

[0188] In some embodiments, the second communication unit 1201 is further configured to send a second channel prediction time number to the terminal device.

[0189] In some embodiments, the second channel prediction time number is determined based on the number of first channel prediction times.

[0190] In some embodiments, the second processing unit 1202 is further configured to select a third model based on the number of second channel prediction times;

[0191] The second communication unit 1201 is further configured to receive a CSI bit stream sent by the terminal device;

[0192] The second processing unit 1202 is further configured to perform feature analysis on the CSI bit stream through the third model to obtain compressed CSI; and decompress the compressed CSI through a fourth model to obtain CSI at multiple times, where the number of times included in the multiple times is the number of second channel prediction times.

[0193] Those skilled in the art should understand that Figure 12 the implementation functions of the various units in the channel prediction device shown can be understood with reference to the relevant descriptions of the foregoing method.Figure 12 The functions of the units in the illustrated channel prediction device can be implemented by a program running on a processor or by specific logic circuits.

[0194] Figure 13 It is a schematic structural diagram of a communication device 1300 provided by an embodiment of the present application. The communication device can be a terminal device or a network device. Figure 13 The illustrated communication device 1300 includes a processor 1310. The processor 1310 can call and run a computer program from a memory to implement the method in the embodiment of the present application.

[0195] Optionally, as Figure 13 shown, the communication device 1300 may further include a memory 1320. Among them, the processor 1310 can call and run a computer program from the memory 1320 to implement the method in the embodiment of the present application.

[0196] Among them, the memory 1320 can be a separate device independent of the processor 1310 or integrated in the processor 1310.

[0197] Optionally, as Figure 13 shown, the communication device 1300 may further include a transceiver 1330. The processor 1310 can control the transceiver 1330 to communicate with other devices. Specifically, it can send information or data to other devices or receive information or data sent by other devices.

[0198] Among them, the transceiver 1330 can include a transmitter and a receiver. The transceiver 1330 may further include an antenna, and the number of antennas can be one or more.

[0199] Optionally, the communication device 1300 can specifically be the network device in the embodiment of the present application, and the communication device 1300 can implement the corresponding processes implemented by the network device in the various methods of the embodiment of the present application. For the sake of brevity, it will not be elaborated here.

[0200] Optionally, the communication device 1300 can specifically be the mobile terminal / terminal device in the embodiment of the present application, and the communication device 1300 can implement the corresponding processes implemented by the mobile terminal / terminal device in the various methods of the embodiment of the present application. For the sake of brevity, it will not be elaborated here.

[0201] Figure 14 It is a schematic structural diagram of a chip according to an embodiment of the present application. Figure 14 The illustrated chip 1400 includes a processor 1410. The processor 1410 can call and run a computer program from a memory to implement the method in the embodiment of the present application.

[0202] Optionally, asFigure 14 As shown, the chip 1400 may further include a memory 1420. Among them, the processor 1410 may call and run a computer program from the memory 1420 to implement the method in the embodiments of the present application.

[0203] Among them, the memory 1420 may be a separate device independent of the processor 1410, or may be integrated in the processor 1410.

[0204] Optionally, the chip 1400 may further include an input interface 1430. Among them, the processor 1410 may control the input interface 1430 to communicate with other devices or chips. Specifically, it may obtain information or data sent by other devices or chips.

[0205] Optionally, the chip 1400 may further include an output interface 1440. Among them, the processor 1410 may control the output interface 1440 to communicate with other devices or chips. Specifically, it may output information or data to other devices or chips.

[0206] Optionally, the chip may be applied to the network device in the embodiments of the present application, and the chip may implement the corresponding processes implemented by the network device in the various methods of the embodiments of the present application. For the sake of brevity, it will not be elaborated here.

[0207] Optionally, the chip may be applied to the mobile terminal / terminal device in the embodiments of the present application, and the chip may implement the corresponding processes implemented by the mobile terminal / terminal device in the various methods of the embodiments of the present application. For the sake of brevity, it will not be elaborated here.

[0208] It should be understood that the chip mentioned in the embodiments of the present application may also be referred to as a system-on-chip, system chip, chip system, or system-on-chip, etc.

[0209] It should be understood that the processor in the embodiments of the present application may be an integrated circuit chip with signal processing capabilities. In the implementation process, each step of the above method embodiments can be completed by the integrated logic circuit in the hardware of the processor or instructions in the form of software. The above-mentioned processor may be a general-purpose processor, a digital signal processor (DSP), an application specific integrated circuit (ASIC), a field programmable gate array (FPGA) or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components. It can implement or execute the various methods, steps and logic block diagrams disclosed in the embodiments of the present application. The general-purpose processor may be a microprocessor or the processor may also be any conventional processor, etc. The steps of the method disclosed in combination with the embodiments of the present application can be directly embodied as being executed by the hardware decoding processor, or executed by a combination of the hardware and software modules in the decoding processor. The software module may be located in a mature storage medium in the art such as a random access memory, a flash memory, a read-only memory, a programmable read-only memory or an electrically erasable programmable memory, a register, etc. This storage medium is located in the memory, and the processor reads the information in the memory and combines its hardware to complete the steps of the above method.

[0210] It can be understood that the memory in the embodiments of the present application can be a volatile memory or a non-volatile memory, or can include both volatile and non-volatile memories. Among them, the non-volatile memory can be a read-only memory (ROM), a programmable read-only memory (PROM), an erasable programmable read-only memory (EPROM), an electrically erasable programmable read-only memory (EEPROM), or a flash memory. The volatile memory can be a random access memory (RAM), which is used as an external cache. By way of example but not limitation, many forms of RAM are available, such as static random access memory (SRAM), dynamic random access memory (DRAM), synchronous dynamic random access memory (SDRAM), double data rate synchronous dynamic random access memory (DDR SDRAM), enhanced synchronous dynamic random access memory (ESDRAM), synchlink dynamic random access memory (SLDRAM), and direct rambus random access memory (DR RAM). It should be noted that the memory of the systems and methods described herein is intended to include but not be limited to these and any other suitable types of memory.

[0211] It should be understood that the above memory is by way of example but not limitation. For example, the memory in the embodiments of the present application can also be a static random access memory (SRAM), a dynamic random access memory (DRAM), a synchronous dynamic random access memory (SDRAM), a double data rate SDRAM (DDR SDRAM), an enhanced SDRAM (ESDRAM), a synch link DRAM (SLDRAM), and a direct rambus random access memory (DR RAM), etc. That is to say, the memory in the embodiments of the present application is intended to include but not be limited to these and any other suitable types of memory.

[0212] The embodiments of the present application further provide a computer-readable storage medium for storing a computer program.

[0213] Optionally, the computer-readable storage medium can be applied to the network device in the embodiments of the present application, and the computer program enables the computer to execute the corresponding processes implemented by the network device in the various methods of the embodiments of the present application. For the sake of brevity, it will not be elaborated here.

[0214] Optionally, the computer-readable storage medium can be applied to the mobile terminal / terminal device in the embodiments of the present application, and the computer program enables the computer to execute the corresponding processes implemented by the mobile terminal / terminal device in the various methods of the embodiments of the present application. For the sake of brevity, it will not be elaborated here.

[0215] The embodiments of the present application further provide a computer program product, including computer program instructions.

[0216] Optionally, the computer program product can be applied to the network device in the embodiments of the present application, and the computer program instructions enable the computer to execute the corresponding processes implemented by the network device in the various methods of the embodiments of the present application. For the sake of brevity, it will not be elaborated here.

[0217] Optionally, the computer program product can be applied to the mobile terminal / terminal device in the embodiments of the present application, and the computer program instructions enable the computer to execute the corresponding processes implemented by the mobile terminal / terminal device in the various methods of the embodiments of the present application. For the sake of brevity, it will not be elaborated here.

[0218] Those of ordinary skill in the art can realize that the units and algorithm steps of the examples described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or by a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0219] Those skilled in the art can clearly understand that for the convenience and brevity of description, the specific working processes of the systems, devices, and units described above can refer to the corresponding processes in the foregoing method embodiments, and will not be elaborated here.

[0220] In several embodiments provided in the present application, it should be understood that the disclosed systems, devices, and methods can be implemented in other ways. For example, the device embodiments described above are merely illustrative. For example, the division of the units is only a logical function division. In actual implementation, there may be other division methods. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the displayed or discussed couplings or direct couplings or communication connections to each other can be through some interfaces, and the indirect couplings or communication connections of the devices or units can be in electrical, mechanical, or other forms.

[0221] The units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they can be located in one place or distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0222] In addition, in each embodiment of the present application, the functional units can be integrated into one processing unit, or each unit can exist physically alone, or two or more units can be integrated into one unit.

[0223] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in each embodiment of the present application. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical discs that can store program codes.

[0224] As described above, the above are only the specific implementation manners of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed in the present application can easily think of changes or substitutions, which should all be covered by the protection scope of the present application. Therefore, the protection scope of the present application should be subject to the protection scope of the claims.

Claims

1. A channel prediction method, characterized in that, The method includes: The terminal device receives a reference signal sent by the network device, and determines the number of first channel prediction time instants based on the received reference signal; The terminal device sends the number of first channel prediction time instants to the network device.

2. The method according to claim 1, wherein The method further includes: The terminal device performs channel prediction based on the received reference signal, and obtains channel state information (CSI) at multiple time instants, where the number of time instants included in the multiple time instants is the number of first channel prediction time instants.

3. The method according to claim 2, characterized in that, The method further includes: The terminal device selects a first model based on the number of first channel prediction time instants; The terminal device compresses the CSI at the multiple time instants through a second model, and extracts features from the compressed CSI through the first model to obtain a CSI bitstream; The terminal device sends the CSI bitstream to the network device.

4. The method according to claim 1, characterized in that The method further includes: The terminal device receives the number of second channel prediction time instants sent by the network device.

5. The method according to claim 4, wherein The number of second channel prediction time instants is determined based on the number of first channel prediction time instants.

6. The method according to claim 4, characterized in that, The method further includes: The terminal device performs channel prediction based on the received reference signal, and obtains CSI at multiple time instants, where the number of time instants included in the multiple time instants is the number of second channel prediction time instants.

7. The method according to claim 6, characterized in that, The method further includes: The terminal device selects a first model based on the number of second channel prediction time instants; The terminal device compresses the CSI at the multiple time instants through a second model, and extracts features from the compressed CSI through the first model to obtain a CSI bitstream; The terminal device sends the CSI bitstream to the network device.

8. A channel prediction method, characterized in that, The method includes: The network device sends a reference signal to the terminal device, where the reference signal is used for the terminal device to determine the number of first channel prediction time instants; The network device receives the number of first channel prediction time instants sent by the terminal device.

9. The method according to claim 8, wherein The method further includes: The network device selects a third model based on the number of first channel prediction time instants; The network device receives the CSI bitstream sent by the terminal device; The network device performs feature analysis on the CSI bitstream through the third model to obtain compressed CSI; and decompresses the compressed CSI through a fourth model to obtain CSI at multiple time instants, where the number of time instants included in the multiple time instants is the number of first channel prediction time instants.

10. The method according to claim 8, characterized in that, The method further includes: The network device sends the number of second channel prediction time instants to the terminal device.

11. The method according to claim 10, wherein, The number of second channel prediction time instants is determined based on the number of first channel prediction time instants.

12. The method according to claim 10, wherein The method further includes: The network device selects a third model based on the number of second channel prediction time instants; The network device receives the CSI bitstream sent by the terminal device; The network device performs feature analysis on the CSI bitstream through the third model to obtain compressed CSI; and decompresses the compressed CSI through a fourth model to obtain CSI at multiple time instants, where the number of time instants included in the multiple time instants is the number of second channel prediction time instants.

13. A channel prediction device, characterized in that, Applied to a terminal device, the apparatus includes: A first communication unit, configured to receive a reference signal sent by a network device; A first processing unit, configured to determine a first number of channel prediction time instants based on the received reference signal; The first communication unit is further configured to send the first number of channel prediction time instants to the network device.

14. A channel prediction device, characterized in that, Applied to a network device, the apparatus includes: A second communication unit, configured to send a reference signal to a terminal device, where the reference signal is used by the terminal device to determine a first number of channel prediction time instants; and receive the first number of channel prediction time instants sent by the terminal device.

15. A communication device, characterized in that, Including: A processor and a memory, where the memory is configured to store a computer program, and the processor is configured to call and run the computer program stored in the memory to execute the method according to any one of claims 1 to 12.

16. A chip, characterized in that, Including: a processor, configured to call and run a computer program from a memory, such that a device installed with the chip executes the method according to any one of claims 1 to 12.

17. A computer-readable storage medium, characterized in that, For storing a computer program, where the computer program causes a computer to execute the method according to any one of claims 1 to 12.

Images