WO2023088387A1 - 信道预测方法、装置、ue及系统 - Google Patents

信道预测方法、装置、ue及系统 Download PDF

Info

Publication number
WO2023088387A1
WO2023088387A1 PCT/CN2022/132642 CN2022132642W WO2023088387A1 WO 2023088387 A1 WO2023088387 A1 WO 2023088387A1 CN 2022132642 W CN2022132642 W CN 2022132642W WO 2023088387 A1 WO2023088387 A1 WO 2023088387A1
Authority
WO
WIPO (PCT)
Prior art keywords
model
prediction
coefficient
target
information
Prior art date
Application number
PCT/CN2022/132642
Other languages
English (en)
French (fr)
Inventor
李佳林
孙鹏
杨昂
Original Assignee
维沃移动通信有限公司
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 维沃移动通信有限公司 filed Critical 维沃移动通信有限公司
Publication of WO2023088387A1 publication Critical patent/WO2023088387A1/zh

Links

Images

Classifications

    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/373—Predicting channel quality or other radio frequency [RF] parameters
    • H—ELECTRICITY
    • H04—ELECTRIC COMMUNICATION TECHNIQUE
    • H04B—TRANSMISSION
    • H04B17/00—Monitoring; Testing
    • H04B17/30—Monitoring; Testing of propagation channels
    • H04B17/391—Modelling the propagation channel

Definitions

  • the embodiments of the present invention relate to the field of communication technologies, and in particular to a channel prediction method, device, UE and system.
  • specific operator models such as linear models, polynomial fitting models, etc.
  • the specific operator model is a time filter, which inputs historical channel estimates and outputs equal time intervals in the future channel at any time; and since the specific operator model only needs endogenous variables without other exogenous variables, the modeling of channel prediction using the specific operator model is relatively simple.
  • Embodiments of the present invention provide a channel prediction method, device, UE and system, which can solve the problem of high computational complexity in channel prediction using a specific operator model.
  • a channel prediction method which is applied to a user equipment UE (User Equipment, UE), and the method may include: the UE sends first information to a network side device, and the first information is used by the network side device to perform channel prediction , the first information includes the channel prediction result and the model parameters of the target-specific operator model, or includes the model parameters of the target-specific operator model; wherein, the channel prediction result is the model parameter and the historical channel estimation result of the UE based on the target-specific operator model
  • the target specific operator model is a specific operator model constructed by the UE based on historical channel estimation results.
  • a channel prediction device including: a sending module.
  • a sending module configured to send first information to the network-side device, the first information is used by the network-side device to perform channel prediction, and the first information includes a channel prediction result and model parameters of a target-specific operator model, or includes a target-specific operator model The model parameters of the sub-model; where the channel prediction result is the channel result predicted by the UE based on the model parameters of the target-specific operator model and the historical channel estimation results, and the target-specific operator model is a specific operator constructed by the UE based on the historical channel estimation results Model.
  • a channel prediction method which is applied to a network-side device.
  • the method includes: the network-side device acquires first information, and the first information includes a channel prediction result and model parameters of a target-specific operator model, or includes a target Model parameters of a specific operator model; wherein, the channel prediction result is the channel result predicted by the UE based on the model parameters of the target specific operator model and historical channel estimation results, and the target specific operator model is constructed by the UE based on the historical channel estimation results A specific operator model; the network side device performs channel prediction according to the first information.
  • a channel prediction device including: an acquisition module and a prediction module.
  • An acquisition module configured to acquire first information, where the first information includes a channel prediction result and model parameters of a target-specific operator model, or includes model parameters of a target-specific operator model; wherein, the channel prediction result is the UE based on a target-specific operator
  • a user equipment UE which includes a processor and a memory, the memory stores programs or instructions that can run on the processor, and when the programs or instructions are executed by the processor, the following steps are implemented: The steps of the method described in the first aspect.
  • a user equipment UE including a processor and a communication interface, wherein the processor is used to: build a target-specific operator model, or to build a target-specific operator model and a target-specific operator model The channel results obtained by predicting the model parameters and the historical channel estimation results are performed, and the channel prediction results are obtained; the target specific operator model is a specific operator model constructed by the UE based on the historical channel estimation results; the communication interface is used to send to the network The side device sends first information, where the first information includes the model parameters and/or channel prediction results.
  • a network-side device in a seventh aspect, includes a processor and a memory, the memory stores programs or instructions that can run on the processor, and the programs or instructions are executed by the processor When realizing the steps of the method as described in the third aspect.
  • a network side device including a processor and a communication interface, wherein the communication interface is used to obtain first information, and the first information includes a channel prediction result and model parameters of a target-specific operator model, or Including the model parameters of the target-specific operator model; wherein, the target-specific operator model is a specific operator model constructed by the UE based on historical channel estimation results, and the channel prediction result is the model parameters and historical channel estimation of the UE based on the target-specific operator model The obtained channel result of result prediction; the processor is used for performing channel prediction according to the first information.
  • a ninth aspect provides a communication system, a user equipment UE and a network side device, the UE can be used to perform the steps of the channel prediction method described in the first aspect, and the network side device can be used to perform the steps in the third aspect The steps of the channel prediction method.
  • a readable storage medium is provided, and a program or an instruction is stored on the readable storage medium, and when the program or instruction is executed by a processor, the steps of the method described in the first aspect are implemented, or the steps of the method as described in the first aspect are implemented, or the The steps of the method described in the third aspect.
  • a chip in an eleventh aspect, includes a processor and a communication interface, the communication interface is coupled to the processor, and the processor is used to run a program or an instruction to implement the method described in the first aspect. method, or implement the method as described in the third aspect.
  • a computer program/program product is provided, the computer program/program product is stored in a storage medium, and the computer program/program product is executed by at least one processor to implement the The steps of the channel prediction method, or the steps of implementing the channel prediction method as described in the third aspect.
  • the UE may send the first information to the network-side device (for the network-side device to perform channel prediction); the network-side device may receive the first information and perform channel prediction according to the first information; the first information includes The channel prediction result and the model parameters of the target-specific operator model, or including the model parameters of the target-specific operator model; wherein the channel prediction result is predicted by the UE based on the model parameters of the target-specific operator model and historical channel estimation results
  • the target specific operator model is a specific operator model constructed by the UE based on historical channel estimation results; the network side device performs channel prediction according to the first information.
  • the network side device can perform channel prediction according to the first information after receiving the first information; that is, the channel prediction method provided in the embodiment of the present application can be based on the model parameters sent by the UE, Alternatively, the channel is indirectly predicted based on the model parameters and channel prediction results sent by the UE, so that the computational complexity of channel prediction through a specific operator model can be reduced.
  • FIG. 1 is a schematic structural diagram of a communication system provided by an embodiment of the present invention
  • FIG. 2 is a schematic diagram of a channel prediction method provided by an embodiment of the present invention.
  • FIG. 3 is a schematic diagram of sampling in a historical time-domain channel through a sliding window in a channel prediction method provided by an embodiment of the present invention
  • FIG. 4 is a schematic diagram of a channel prediction flow chart of a channel prediction method provided by an embodiment of the present invention.
  • FIG. 5 is one of the structural schematic diagrams of a channel prediction device provided by an embodiment of the present invention.
  • FIG. 6 is the second structural schematic diagram of a channel prediction device provided by an embodiment of the present invention.
  • FIG. 7 is a schematic diagram of a hardware structure of a communication device provided by an embodiment of the present application.
  • FIG. 8 is a schematic diagram of a hardware structure of a UE provided in an embodiment of the present application.
  • FIG. 9 is a schematic diagram of a hardware structure of a network side device provided by an embodiment of the present application.
  • first, second and the like in the specification and claims of the present application are used to distinguish similar objects, and are not used to describe a specific sequence or sequence. It is to be understood that the terms so used are interchangeable under appropriate circumstances such that the embodiments of the application are capable of operation in sequences other than those illustrated or described herein and that "first" and “second” distinguish objects. It is usually one category, and the number of objects is not limited. For example, there may be one or more first objects.
  • “and/or” in the description and claims means at least one of the connected objects, and the character “/” generally means that the related objects are an "or” relationship.
  • LTE Long Term Evolution
  • LTE-Advanced LTE-Advanced
  • LTE-A Long Term Evolution-Advanced
  • CDMA Code Division Multiple Access
  • TDMA Time Division Multiple Access
  • FDMA Frequency Division Multiple Access
  • OFDMA Orthogonal Frequency Division Multiple Access
  • SC-FDMA Single-carrier Frequency Division Multiple Access
  • system and “network” in the embodiments of the present application are often used interchangeably, and the described technology can be used for the above-mentioned system and radio technology, and can also be used for other systems and radio technologies.
  • the following description describes the New Radio (New Radio, NR) system for example purposes, and uses NR terminology in most of the following descriptions, but these techniques can also be applied to applications other than NR system applications, such as the 6th generation (6th Generation , 6G) communication system.
  • 6G 6th generation
  • Fig. 1 shows a block diagram of a wireless communication system to which the embodiment of the present application is applicable.
  • the wireless communication system includes a terminal 11 and a network side device 12 .
  • the terminal 11 can be a mobile phone, a tablet computer (Tablet Personal Computer), a laptop computer (Laptop Computer) or a notebook computer, a personal digital assistant (Personal Digital Assistant, PDA), a palmtop computer, a netbook, a super mobile personal computer (ultra-mobile personal computer, UMPC), mobile Internet device (Mobile Internet Device, MID), augmented reality (augmented reality, AR) / virtual reality (virtual reality, VR) equipment, robot, wearable device (Wearable Device) , vehicle equipment (VUE), pedestrian terminal (PUE), smart home (home equipment with wireless communication functions, such as refrigerators, TVs, washing machines or furniture, etc.), game consoles, personal computers (personal computers, PCs), teller machines or self-service Wearable devices include: smart watches, smart bracelets, smart headphones, smart glasses, smart jewelry (
  • the network side device 12 may include an access network device or a core network device, where the access network device 12 may also be called a radio access network device, a radio access network (Radio Access Network, RAN), a radio access network function, or Wireless access network unit.
  • RAN Radio Access Network
  • RAN Radio Access Network
  • Wireless access network unit Wireless access network unit
  • the access network device 12 may include a base station, a WLAN access point, or a WiFi node, etc., and the base station may be called a Node B, an evolved Node B (eNB), an access point, a Base Transceiver Station (Base Transceiver Station, BTS), a radio Base station, radio transceiver, Basic Service Set (BSS), Extended Service Set (ESS), Home Node B, Home Evolved Node B, Transmitting Receiving Point (TRP) or all As long as the same technical effect is achieved, the base station is not limited to a specific technical vocabulary. It should be noted that in this embodiment of the application, only the base station in the NR system is used as an example for introduction, and The specific type of the base station is not limited.
  • X [x 1 , x 2 , . . . , x N ].
  • X is the attribute set of phenomenon P
  • x N is the Nth attribute of phenomenon P.
  • a linear model can be written in vector form as:
  • W [w 1 ,w 2 ,...,w N ]
  • the linear model f(X) is a real-valued function on the real number field
  • the above-mentioned linear model f(X) is called a linear regression model.
  • the parameters W and b of the linear model f(X) can be solved, then the phenomenon P can be deduced and predicted through the linear model f(X).
  • the time-varying channel process can be regarded as a time-autoregressive moving average process, and it can be considered as a completely predictable process when the signal-to-noise ratio is sufficient.
  • a linear model can be used for recursive prediction; at this time, the linear model is a time filter that inputs historical channel estimation results and outputs channels at equal time intervals in the future.
  • the channel can be modeled as a linear process of order P as:
  • H(n) represents the channel information at the nth moment
  • P is the order of the linear model
  • w(n) is white noise
  • the filter coefficient of the linear model is called linear model parameters.
  • the parameters of the linear model (specifically, the parameter values) must first be obtained.
  • the least square method is used as an example to construct an overdetermined equation system:
  • a [a 1 ,a 2 ,...,a n ] T
  • w [w n+1 ,w n+2 ,...,w n+N ] T ;
  • H n represents the channel information at the n-th moment
  • Y is the channel information matrix at the time that needs to be fitted
  • X is the channel information matrix used to fit the historical moment at the moment in Y, and its dimension is N ⁇ n
  • a is the linear coefficient of the linear model
  • w is usually called the residual.
  • the channel sample contains channel information for 10 consecutive moments, and these 10 moments are assumed to be t1, t2, t3, t4, t5, t6, t7, t8, t9, t10;
  • the least squares estimate of the coefficient a of the linear model is:
  • the residual w represents the difference between the theoretical and actual values, and the coefficient a of the linear model when the sum of the residual squares reaches the minimum indicates that the linear model fits the data sample best, and the channel prediction at the future moment can be realized, and the residual square and the formula (equation) is:
  • the coefficient equation of the linear model can be determined based on the residual sum of squares formula (equation). It can be seen that the number of columns of matrix X is still missing in this equation, that is, the order P of the linear model; the number of columns of matrix X can also be understood as how many historical channel samples input into the linear model need to be used for future channel prediction .
  • a good linear model usually requires a small residual error, and the model is relatively simple, that is, the order of the linear model is required to be low, so that some order determination criteria can be used to compare the superiority of linear models with different orders. inferior.
  • Commonly used order-setting criteria include the final forecast error criterion FPE (Final Prediction Error, FPE), Akaike information criterion AIC (Akaike information criterion, AIC) and Bayesian information criterion BIC (Bayesian information criterion, BIC) criterion function.
  • the advantage of using a linear model for channel prediction is that the model is very simple, and the linear model only needs endogenous variables instead of other exogenous variables; thus, the modeling of the linear model is relatively simple.
  • the shortcomings of the linear model are also obvious, specifically: when using a linear model for channel prediction in a complex environment, the model parameters will be large, and the amount of calculation will be too large; further, the current channel prediction based on the linear model has certain limitations , can only predict the channel at a time in the future that is equally spaced from the historical input data.
  • the UE may send to the network side device first information including model parameters of a specific operator model constructed by the UE based on historical channel estimation results and/or channel prediction results based on the model parameters , so that the network-side device can perform channel prediction according to the first information after receiving the first information, that is, the network-side device can indirectly predict the channel based on the model parameters and/or channel prediction results sent by the UE, thereby reducing Computational complexity of the model for channel prediction.
  • FIG. 2 shows a flowchart of a channel prediction method provided in an embodiment of the present application.
  • the signal transmission method provided by the embodiment of the present application may include the following steps 201 and 203 .
  • Step 201 the UE sends first information to a network side device.
  • the first information may be used by the network side device to perform channel prediction, and the first information includes the channel prediction result and the model parameters of the target-specific operator model, or includes the model parameters of the target-specific operator model.
  • the channel prediction result is a channel result predicted by the UE based on historical channel estimation results and model parameters of a target-specific operator model
  • the target-specific operator model is a specific operator model constructed by the UE based on historical channel estimation results.
  • the target-specific operator model can be a linear model, a polynomial fitting model, an oane algorithm model, a music algorithm model, or any other model that can be used for channel prediction, which can be based on actual use Requirements are determined, and this embodiment of the application does not make a limitation.
  • the above historical channel estimation result is a channel estimation result obtained by the UE using conventional means (such as estimating the channel by measuring CSI-RS/DMRS).
  • the UE may construct a target-specific operator model with multiple historical channel estimation results.
  • the above model parameters may include the first model order and the first model coefficient; the first model order may be used by the network side device to determine the target coefficient prediction component, and the first model coefficient may be used for The network side device performs channel prediction based on the target coefficient prediction component.
  • the above-mentioned model parameters may be called linear filter parameters, and the linear filter parameters may include the order of the linear filter (that is, the first model order) and coefficients (ie the first model coefficients).
  • the UE performs channel prediction based on the target-specific operator model, and reports the obtained channel prediction result and the model parameters of the target-specific operator model, so as to ensure that the model parameters adopted by the UE and the network-side device remain unified, which facilitates
  • a network-side device (such as a base station) selects an appropriate target coefficient prediction component to predict model coefficients at a future time; after obtaining the model coefficients at a future time, the network-side device can predict a channel at a future time based on the obtained model coefficients.
  • the target coefficient prediction component may be at least one of the following: a coefficient prediction component constructed through one or more cell identifiers, a coefficient prediction component constructed through one or more transmission and reception point TRP identifiers , a coefficient prediction component constructed by one or more channel feature information, a coefficient prediction component constructed by a beam identifier in a cell, and a coefficient prediction component constructed by different geographical locations.
  • the above-mentioned channel characteristic information refers to a precoding matrix indicator PMI (Precoding Matrix Indicator, PMI) and a channel quality indicator CQI ( Channel quality indication, CQI), rank indication RI (RANK Indicator, RI), etc.
  • the network side device may construct UEs with the same channel characteristic information as a coefficient prediction component.
  • these cell identities include the cell identities of the cells of the UE.
  • these TRP identifiers include the TRP identifier of the UE.
  • the network side device may construct the coefficient prediction component in units of one or more cell identities; or, may construct the coefficient prediction component in units of one or more TRP identities.
  • the target coefficient prediction component may be any of the following: a linear coefficient prediction component specific to each RB, a coefficient prediction component specific to each sub-band, coefficient prediction for multiple RBs or multiple sub-bands component, wideband coefficient prediction component.
  • the frequency domain prediction granularity of the target coefficient prediction component is RB; in the case where the target coefficient prediction component is a specific linear coefficient prediction component for each subband , the frequency-domain prediction granularity of the target coefficient prediction component is the subband; in the case where the target coefficient prediction component is a linear coefficient prediction component specific to multiple RBs, the frequency-domain prediction granularity of the target coefficient prediction component is multiple RBs; in When the target coefficient prediction component is a linear coefficient prediction component specific to multiple subbands, the frequency domain prediction granularity of the target coefficient prediction component is multiple subbands; when the target coefficient prediction component is a broadband linear coefficient prediction component, the target coefficient The frequency domain prediction granularity of the prediction component is broadband.
  • the target coefficient prediction component may be a coefficient prediction component based on a neural network (such as a convolutional neural network), or any artificial intelligence (AI) network based on other learning capabilities.
  • the coefficient prediction component can be specifically determined according to actual usage requirements, and is not limited in this embodiment of the present application.
  • the above-mentioned channel prediction results may be the number of UEs whose first model order is less than or equal to the historical channel estimation results (the historical channel estimation results based on which the target operator model is constructed for the UE) (for convenience description, hereinafter referred to as the target number), the channel prediction is obtained based on the first model parameters and the historical channel estimation results.
  • whether the first model order is less than or equal to the target number is a channel prediction condition for whether the UE performs channel prediction based on the target specific operator model. If the first model order is less than or equal to the target number, the UE may perform channel prediction based on the first model parameters and the historical channel estimation result, and send the channel prediction result and the model parameters of the target specific operator to the network side device; if the first If the model order is greater than the target number, the UE does not perform channel prediction based on the first model parameters and the historical channel estimation results this time, and does not report the model parameters of the target specific operator model to the network side device, so the UE needs to rebuild the specific operator model. operator model.
  • the UE may perform channel prediction at equal intervals by using the first model parameters and historical channel estimation results. It can be understood that, limited by the characteristics of the linear model, the prediction time of the UE cannot be specified arbitrarily, and only the channel at the time equal to the interval between the current prediction time and the historical data (that is, the above-mentioned historical channel estimation results) can be predicted, and only The channel at a time can be predicted in the future; then the UE can send the channel prediction result of this prediction and the model parameters of the target specific operator to the network side device.
  • the first information may also include the model coefficient prediction granularity expected by the UE, and the model coefficient prediction granularity is used by the network side device to predict the model coefficient; where the model coefficient prediction granularity may include the following At least one item: time domain granularity, frequency domain granularity.
  • the network side device can refer to the model prediction granularity expected by the UE to predict the model coefficient, thereby The accuracy of channel prediction performed by the network side device can be ensured.
  • step 201 may specifically be implemented through the following step 201a or step 201b.
  • Step 201a the UE sends the first information to the network side device on the target resource.
  • the target resource may include at least one of the following: radio resource control RRC (Radio Resource Control, RRC) pre-configured fixed resources, medium access control-control unit MAC CE (Medium Access Control-Control Element, MAC CE) indicated Resources, resources indicated by Downlink Control Information DCI (Downlink Control Information DCI).
  • RRC Radio Resource Control
  • MAC CE Medium Access Control-Control Element, MAC CE
  • step 201b the UE sends the first information to the network side device on the resource for sending the CSI measurement information.
  • the UE sends the first information to the network side device on the resource for sending the CSI measurement information, which can avoid extra time delay and resource configuration caused by reporting the first information for signaling interaction.
  • the flexibility of sending the first information can be improved.
  • Step 202 the network side device acquires first information.
  • the acquisition of the first information by the network side device may specifically be: the network side device receives the first information sent by the UE; or, the network side device may use other means to obtain the first information (such as the first information pre-stored in the network-side device); the details may be determined according to actual usage requirements, and are not limited in this embodiment of the present application.
  • Step 203 the network side device performs channel prediction according to the first information.
  • the above-mentioned model parameters include the first model order and the first model coefficient
  • the above-mentioned step 203 can be specifically implemented through the following steps A to C.
  • Step A the network side device determines the target coefficient prediction component corresponding to the first model order according to the first model order.
  • the order corresponding to the target coefficient prediction component is the same as the order of the first model.
  • the order corresponding to the target coefficient prediction component refer to the relevant description of the order corresponding to the coefficient prediction component in the following embodiments for details, and details are not repeated here to avoid repetition.
  • Step B The network side device uses the target coefficient prediction component to predict the model coefficients according to the first model coefficients to obtain the second model coefficients.
  • the network side device may input the first model coefficient into the target coefficient prediction component, so as to predict the model coefficient at a future time through the target coefficient prediction component and the first model coefficient, and obtain the second Two model coefficients. That is, the second model coefficient is a model coefficient at a future moment.
  • the time-domain granularity of the second model coefficient can be any of the following: one time slot, multiple time slots, remaining time slots of the current frame, CSI measurement time, multiple CSI measurements Time, each time slot in the CSI measurement period, each time slot in multiple CSI measurement periods; and/or, the frequency domain granularity of the second model coefficient can be any of the following: Resource Block RB (Resource Block, RB) , subband, broadband.
  • Resource Block RB Resource Block, RB
  • subband broadband
  • the channel prediction method provided in the embodiment of the present application introduces the target coefficient prediction component to realize the linear model at multiple time points at equal intervals The linear coefficient is predicted, and then the channel state at multiple moments with equal time intervals can be obtained, which can improve the accuracy of channel prediction.
  • step B in a case where the first information further includes the model coefficient prediction granularity expected by the UE, the above step B may specifically be implemented through the following step B1.
  • Step B1 the network side device uses the target coefficient prediction component to predict the model coefficients according to the first model coefficient, the model coefficient prediction granularity expected by the UE, and other model coefficients expected by the UE.
  • the granularity is used to predict the model coefficients to obtain the second model coefficients.
  • the base station when the first information also includes the model coefficient prediction granularity expected by the UE, the base station (network side device) can use the first model parameter, the model coefficient prediction granularity expected by the UE, and other UEs at this time
  • the linear coefficient prediction requirements of other UEs that is, the linear model prediction granularity expected by other UEs
  • the reasonable granularity allocation is performed, and the prediction of the second model coefficient is performed; that is, the network side device can synthesize the first model coefficient, the model coefficient prediction granularity expected by the UE, and the
  • the model coefficient prediction granularity expected by other UEs is used to predict model parameters, so that the accuracy and rationality of the model coefficients performed by the network side equipment can be improved.
  • the network side device may first combine the UE expected model coefficient prediction granularity with the target coefficient prediction Comparing the target prediction granularity of the components, if the model coefficient prediction granularity expected by the UE matches the target prediction granularity (for example, both are the same), the network side device can predict the model coefficient according to the model coefficient prediction granularity expected by the UE, to obtain Second model coefficients. If the model coefficient prediction granularity desired by the UE does not match (for example, the two are different) from the target prediction granularity, the network side device may predict the model coefficients according to the target prediction granularity to obtain the second model coefficients.
  • the network side device can determine to perform Prediction granularity for model coefficient predictions.
  • the frequency domain granularity is arranged in order from wide to narrow: broadband, subband, RB; that is, the order of frequency domain granularity is: broadband>subband>RB.
  • step B may specifically be implemented through the following step B2.
  • Step B2 when the UE's expected model coefficient prediction granularity does not match the target prediction granularity, the network side device uses the target coefficient prediction component to predict the model coefficients according to the first model coefficient and the target prediction granularity, and obtains the second model coefficient .
  • Step B2 is exemplarily described below in conjunction with five situations in which the UE's expected model coefficient prediction granularity does not match the target prediction granularity, specifically the following (a-e).
  • UE expected model coefficient prediction granularity is subband/RB in broadband, target prediction granularity is broadband;
  • the second model coefficient is each subband/RB in the broadband Forecast results for shared broadband.
  • UE expected model coefficient prediction granularity is broadband
  • target prediction granularity is subband in broadband
  • the second model coefficient is the channel quality indicator CQI in all subbands in the wideband (Channel quality indication, CQI) The prediction result on the lowest sub-band;
  • UE expected model coefficient prediction granularity is RB in subband
  • target prediction granularity is subband.
  • the second model coefficient is that each RB in the subband shares the subband The predicted results of the belt.
  • UE expected model coefficient prediction granularity is sub-band
  • target prediction granularity is RB in sub-band.
  • the second model coefficient is the one with the lowest CQI among all RBs in the subband RB's prediction results
  • the model coefficient prediction granularity desired by the UE is broadband
  • the target prediction granularity is RBs under subbands in the broadband.
  • the second model coefficient uses the lowest CQI among all RBs in the wideband The result of the RB.
  • Step B will be exemplarily described below in combination with specific examples.
  • the base station that is, the network side device
  • receives the first information and the frequency domain granularity expected by the UE is a subband; the base station can decide whether to set the prediction granularity as the terminal expected according to the actual situation of the current forecasted service demand
  • the base station can use the target coefficient prediction component to perform model coefficient prediction at the granularity desired by the UE; if the target prediction granularity determined by the base station is broadband, that is, the target prediction granularity is broadband and UE desired If the subbands do not match, the base station can perform channel prediction that each subband in the broadband shares the broadband, that is, the obtained prediction result is the prediction result that each subband in the broadband shares the broadband; that is, each subband shares the broadband, and each Utilize wideband for slot-level model coefficient prediction.
  • the network side device in the case that the UE's desired model coefficient prediction granularity does not match the target prediction granularity, since the network side device can perform model coefficient prediction based on the target prediction granularity, the network side device can be improved. The probability of success predicted by the model coefficients.
  • Step C the network side device performs channel prediction according to the second model coefficient and the first model order.
  • the network side device may perform channel prediction according to the second model coefficient, the first model order, and the channel prediction result sent by the UE.
  • the network side device can estimate As a result, the channel is predicted for multiple time instants at equal time intervals in the future.
  • the network side device can determine the target coefficient prediction component corresponding to the first model order based on the obtained first model order; and use the target coefficient prediction component to perform model Prediction of the coefficients to obtain the second model coefficients; and perform channel prediction according to the second model coefficients and the first model order; that is, the network side device can indirectly perform Channel prediction; therefore, compared with the scheme in which the network side equipment directly uses a specific operator model for channel prediction in the related art, the channel prediction method provided in the embodiment of the present application indirectly predicts the channel by predicting model coefficients, which reduces the complexity of channel prediction Spend.
  • the UE can send to the network side device the first information including the model parameters of the specific operator model constructed by the UE based on the historical channel estimation results and/or the channel prediction results based on the model parameters. information, so the network side device can perform channel prediction according to the first information after receiving the first information; that is, the channel prediction method provided in the embodiment of the present application can indirectly predict the channel based on the model parameters and/or channel prediction results sent by the UE, Therefore, the computational complexity of performing channel prediction through a specific operator model can be reduced.
  • the channel prediction method provided in the embodiment of the present application may further include the following steps 204 and 205.
  • Step 204 the network side device sends the second information to the UE.
  • Step 205 the UE receives the second information.
  • the second information may be used to indicate the minimum quantity threshold of historical channel estimation results required by the UE to construct the target specific operator model.
  • the minimum number threshold is an experience value obtained by the network side device during offline training of the coefficient prediction component.
  • the minimum quantity threshold is determined by the network side device according to at least one order value, at least one order value is the order value corresponding to at least one order group, and the number of orders in the at least one order group satisfies the preset condition, and each order The array includes multiple degrees with the same degree value.
  • the above-mentioned preset conditions include any of the following: the number of orders in the above-mentioned at least one order group is greater than or equal to the preset number threshold, and the number of orders in the above-mentioned at least one order group is The first M order arrays among the N order arrays arranged from small to large; wherein, M is the same as the number of at least one order array, and M is a positive integer.
  • the N orders include any of the following:
  • the network side device can use the model coefficients of multiple specific operator models with the same order to perform offline training of the coefficient prediction component to obtain a coefficient prediction component; the coefficient prediction component corresponds to The order is the same as that of the plurality of specific operator models.
  • the UE after receiving the second information, the UE can use conventional means to perform channel estimation to collect historical channel estimation results; when the number of collected historical channel estimation results reaches the above minimum threshold, the UE One can start trying to build a target specific operator model based on all the historical channel estimation results collected.
  • the second information may include at least one of the following: a parameter or a set of parameters pre-configured by RRC, MAC CE, and DCI.
  • the minimum quantity threshold when RRC pre-configures a parameter, the minimum quantity threshold may be indicated only by this parameter, and this parameter is the minimum quantity threshold.
  • the set of parameters can be a set of order values, respectively: 7, 9, and 11, the second information can also include MAC CE or DCI, so that the set of parameters can be obtained from the set of parameters through MAC CE and DCI A parameter is determined as the minimum number threshold.
  • the above-mentioned minimum quantity threshold may be included in the MAC CE or DCI, and the UE can directly obtain the minimum quantity threshold by decoding the DCI and MAC CE ;
  • MAC CE or DCI can bear an indication information, so as to indicate the minimum quantity threshold through the indication information.
  • the difference between MAC CE and DCI lies in the delay, and the delay of MAC CE is greater than the delay of DCI.
  • RRC>MAC CE>DCI that is, the higher the signaling delay to the upper layer, the greater the amount of information that can be carried at the same time.
  • the network side device can use the second information UE to construct the minimum number threshold of historical channel estimation results required for the target-specific operator model, after the UE receives the second information, it can use the collected historical channel
  • the target-specific operator model is created after the number of estimation results is greater than or equal to the minimum number threshold, so that the overhead of the UE can be saved and the accuracy of the target-specific operator model can be improved.
  • the channel prediction method provided in the embodiment of the present application may further include the following step 206 .
  • Step 206 the network side device determines a target coefficient prediction module according to the target information.
  • the target coefficient prediction module may include multiple coefficient prediction components, and the multiple coefficient prediction components may include the target coefficient prediction component.
  • the target information may include at least one of the following: geographic location information of the UE, CSI measurement information sent by the UE, tracking area identifier TAI corresponding to the UE, beam coverage of the UE, beam identifier of the UE, and scene information corresponding to the UE.
  • the scene information corresponding to the UE may be the scene information of the geographic location indicated by the geographic location information of the UE.
  • the scene information of the geographic location may indicate at least one of the following: a dense building group scene, an open scene (such as a highway), and a town scene.
  • the target information is different, and the target coefficient prediction module determined by the network side device may also be different, so the target coefficient prediction component is also different, which can be determined according to actual use requirements, and is not limited in this embodiment of the application.
  • the network-side device can divide (classify) the coefficient prediction components built/constructed by the network-side device based on the factors that affect the channel state to form multiple coefficient prediction modules , each coefficient prediction module includes multiple coefficient prediction components.
  • the coefficient prediction component can construct a coefficient prediction module according to one of the following standards:
  • CSI measurement information (results) with large differences, such as CQI and PMI
  • a coefficient prediction module of different CSI measurement information is constructed.
  • the network-side device uses the factors that can most affect the channel state as the standard to construct coefficient prediction modules under different standards, and each coefficient prediction module can contain linear coefficients under different model orders Forecast component.
  • network-side devices can build coefficient prediction modules with different model orders.
  • the network side device can build one or more of the following coefficient prediction modules:
  • Each RB's unique coefficient prediction module is suitable for dense building complex scenarios
  • Broadband coefficient prediction module suitable for open scenes.
  • the coefficient prediction module can be constructed based on the factors that most affect the channel conversion conditions, rather than the terminal-specific (specific) coefficient prediction module. Generality of coefficient prediction modules.
  • the target coefficient prediction module determined by the network side device for predicting model coefficients will also change correspondingly, which improves prediction flexibility and prediction accuracy.
  • the channel prediction method provided in the embodiment of the present application can perform channel prediction on each time slot in the future interval .
  • the execution process of the channel prediction method provided by the embodiment of the present application will be exemplarily described below by taking a specific operator model (such as a target-specific operator model) as a linear model as an example. .
  • the solid line box 30 is a sliding window,
  • the window size can be configured; it is also assumed that a sliding window with a size of T (Note: K>>T) slides in the historical channel of UE i . Sliding one slot each time and sliding once, the time domain channel sampling value in the current sliding window is used to construct the overdetermined equations of the linear model of the data sample, and the time domain channel sampling interval is related to the CSI-RS measurement of the RRC configuration on the network side The period is consistent, and the order P (T>P) and linear coefficient a of the linear model are calculated.
  • the specific steps are:
  • Step 1 assuming that at time t, the sampling interval is g, the total number of samples is n, and the window size is T, then within the sliding window, specifically the bold realization in Figure 3, the historical channel sample set Y1 in block 30 is:
  • Y 1 [H tT ,H t-T+1*g+1*1 ,...,H t-T+(n-1)*g+(n-1)*1 ];
  • Step 2 after the calculation of the linear coefficient a P2 of the historical channel sample set Y 1 is completed, the sliding window is moved forward by 1 slot, and after moving the sliding window by 1 slot, the inside of the sliding window, specifically, the implementation box in Figure 3
  • the channel sampling value within 31 is used as another historical channel sample set Y 2 ; as in step 1, a linear model is constructed according to the historical channel sample set Y 2 , and the optimal order P 2 of the linear model is obtained according to the criterion function, and then Calculate the linear coefficient under the historical channel sample set Y2
  • Step 3 the linear coefficient of the historical channel sample set Y 2 After the calculation is completed, the sliding window continues to move forward by 1 slot. Without loss of generality, when the sliding window moves to the t+kth moment, the historical channel sample set Y N within the sliding window, specifically the dashed box 32 in FIG. 3 is:
  • Y N [H t+kT ,H t+k-T+g*1+1*1 ,...,H t+k-T+(n-1)*g+(n-1)*1 ];
  • the linear coefficients of N linear models can be obtained, which are respectively a P1 , a P2 , . . . , a PN .
  • each coefficient prediction component is obtained by offline training of the linear coefficient set under the same order, and different coefficient prediction components are obtained by offline training of the linear coefficient set under different orders.
  • the base station can use the time sliding window in the historical time domain channel to construct the linear coefficient set of the linear model, put the linear coefficient set into the AI network model for offline training, and learn any prediction Linear model coefficients for moments.
  • the coefficient prediction module of the offline training of the base station in an open scene is a broadband coefficient prediction module.
  • Figure 4 is a schematic flow chart of channel prediction using a linear model, and the specific steps are as follows:
  • the UE determines the model parameters of the linear model (ie, the target-specific operator model).
  • the base station may indicate to the UE a reference value N of channel estimation collection quantity through RRC signaling (that is, the second information), that is, the number of historical channel estimation results required by the UE to construct the target specific operator model Minimum quantity threshold, N is a positive integer.
  • the UE can use the N historical channel estimation results to try to construct the overdetermined equations of the linear model; then, the UE can determine according to the criterion function of information theory The order P of the linear model (ie the first model order). After determining the order P, the UE can first judge whether the order P satisfies the channel prediction condition, and if the order P meets the channel prediction condition, then calculate the linear coefficient a of the linear model, otherwise it does not calculate the linear coefficient a of the linear model.
  • the UE can use the model parameters of the linear model and N historical channel estimation results to perform channel prediction at equal intervals (limited by the linear model, the prediction time cannot be specified arbitrarily, and can only be predicted from The channel at the moment equal to the interval between the historical data from the current prediction moment, and the channel at an interval moment is predicted in the future), specifically, as shown in Figure 4, assuming that the Nth historical channel estimation results of the N historical channels The channel estimation result is the channel result at time t1, then the UE can perform channel prediction at time t2 according to the model parameters of the linear model and N historical channel estimation results, that is, the channel prediction result at time t2 can be obtained.
  • Parameter reporting that is, sending the first information to the network side device.
  • the UE may report the model parameters of the linear model to the network side device, where the model parameters may include the order P of the linear model and the linear coefficient a of the linear model.
  • the UE can report the CSI measurement information together with the channel prediction results and the model parameters of the linear model (that is, the first information) It is also reported to the base station.
  • the UE can report information (that is, the first - information) carries the linear coefficient prediction frequency domain granularity required by the UE (ie, the linear coefficient prediction granularity expected by the UE), and the linear coefficient prediction frequency domain granularity is a subband.
  • the base station performs linear coefficient prediction.
  • the base station After the base station receives the first information sent by the UE, it can determine the target coefficient prediction module and the target coefficient according to which beam coverage the UE is in (that is, the beam coverage of the UE) or the beam identifier to which the UE belongs (that is, the beam identifier of the UE). A number of coefficient prediction components are included in the prediction module.
  • the base station determines which linear coefficient prediction component (that is, the target coefficient prediction component) in the target coefficient prediction module to use according to the order P in the model parameters reported by the UE, for example, the target coefficient prediction component is the coefficient prediction under the order P components.
  • the target coefficient prediction component is the coefficient prediction under the order P components.
  • the base station inputs the linear coefficient a of the linear model reported by the UE into the target coefficient prediction component, that is, uses the linear coefficient prediction component finally determined by the base station to perform the next interval of the current linear parameter prediction moment (such as time t2 in Figure 4)
  • the linear coefficient prediction at time (for example, time t3 in FIG. 4 ) obtains a linear coefficient (ie, the second model coefficient).
  • the information reported by the UE includes the frequency domain granularity of UE demand prediction, that is, the UE expected linear coefficient frequency domain prediction granularity, so that the base station can decide whether to determine the predicted frequency domain granularity as the terminal according to the actual situation of the current predicted service demand desired granularity.
  • the base station may determine to perform channel prediction at the model coefficient prediction granularity expected by the UE; however , since the target coefficient prediction component is the coefficient prediction component in the broadband coefficient prediction module, that is, the prediction granularity (also called the target prediction granularity) of the target coefficient prediction component is broadband, so the base station can determine that the target prediction granularity is consistent with the model expected by the UE
  • the coefficient prediction granularity (that is, the subband) does not match, so that the base station can use the target prediction granularity and the first model coefficient to perform slot-level linear coefficient prediction based on the target coefficient prediction component to obtain the second model coefficient (which is a linear coefficient ).
  • the base station uses the target coefficient prediction component to predict the linear coefficient at the next interval time (such as time t3 in Figure 4) according to the linear coefficient a of the linear model reported by the UE; then, the base station can use the prediction information reported by the UE (such as The channel result at time t2 in Figure 4 predicted by the UE based on the linear model) and the channel information reported by the historical measurement (such as the N historical channel estimation results in Figure 4) form the input samples required by the order P, for the next interval time channel for prediction.
  • the prediction information reported by the UE such as The channel result at time t2 in Figure 4 predicted by the UE based on the linear model
  • the channel information reported by the historical measurement such as the N historical channel estimation results in Figure 4
  • the channel information reported by historical measurement may include at least one of the following: channel results predicted by UE based on model parameters and N historical channel estimation results, N historical channel estimation results; channel results reported by other UE historical measurements.
  • the required base station can construct model parameters based on the constructed model parameters and these For the solution of channel prediction at a future moment based on information information, since the channel prediction method provided by the embodiment of the present application can directly predict the second model coefficient through the target coefficient prediction component, the complexity of channel prediction can be simplified.
  • the base station since the base station can indirectly predict the channel through the target coefficient prediction component capable of predicting linear coefficients, the complexity of channel prediction is reduced.
  • the channel prediction method provided in the embodiment of the present application may also be executed by a channel prediction device.
  • the channel prediction device provided in the embodiment of the present application is described by taking the channel prediction method performed by the channel prediction device as an example.
  • Fig. 5 shows a possible structural diagram of a channel prediction device involved in the embodiment of the present application.
  • the channel prediction apparatus 50 may include: a sending module 51 .
  • the sending module 51 is configured to send the first information to the network side device, the first information includes the channel prediction result and the model parameters of the target-specific operator model, or includes the model parameters of the target-specific operator model; wherein, the channel prediction result is the channel result predicted by the UE based on the model parameters of the target-specific operator model and historical channel estimation results, and the target-specific operator model is a specific operator model constructed by the UE based on the historical channel estimation results.
  • the above model parameters include a first model order and a first model coefficient
  • the first model order is used by the network side device to determine the target coefficient prediction component
  • the first model coefficient is used by the network side device based on The target coefficient prediction component performs channel prediction.
  • the above channel prediction result is obtained by the UE performing channel prediction based on the first model parameters and historical channel estimation results when the first model order is less than or equal to the number of historical channel estimation results.
  • the above channel prediction apparatus further includes a receiving module.
  • the receiving module is configured to receive the second information sent by the network side device before the sending module 51 sends the first information to the network side device, and the second information is used to indicate the historical channel estimation results required by the UE to construct the target specific operator model Minimum quantity threshold;
  • the second information includes at least one of the following: a parameter or a set of parameters pre-configured by RRC, MAC CE, and DCI.
  • the first information above further includes the model coefficient prediction granularity expected by the UE, and the model coefficient prediction granularity is used by the network side device to predict the model coefficient;
  • model coefficient prediction granularity includes at least one of the following: time domain granularity and frequency domain granularity.
  • the above sending module 51 is specifically configured to send the first information to the network side device on the target resource, and the target resource includes at least one of the following: RRC pre-configured fixed resources, MAC CE indicated resources, resources indicated by DCI;
  • the sending module 51 is specifically configured to send the first information to the network side device on the resource for sending the CSI measurement information.
  • An embodiment of the present application provides a channel prediction device.
  • the channel prediction device can send to the network side device information including model parameters of a specific operator model constructed by the channel prediction device based on historical channel estimation results and/or channel prediction results based on the model parameters.
  • the first information so the network side device can perform channel prediction according to the first information after receiving the first information; that is, the channel prediction method provided in the embodiment of the present application can be based on the model parameters and the specific operator model constructed by the channel prediction device /or indirectly predict the channel based on the channel prediction result predicted by the specific operator model, so that the computational complexity of channel prediction through the specific operator model can be reduced.
  • the channel measurement apparatus in this embodiment of the present application may be an electronic device, such as an electronic device with an operating system, or a component in the electronic device, such as an integrated circuit or a chip.
  • the electronic device may be a terminal, or other devices other than the terminal.
  • the terminal may include, but not limited to, the types of terminal 11 listed above, network attached storage (Network Attached Storage, NAS), etc., which are not specifically limited in this embodiment of the present application.
  • Network Attached Storage Network Attached Storage
  • the channel prediction device in this embodiment of the present application may be a device or a UE, and may also be a component, an integrated circuit, or a chip in the UE.
  • the channel measurement device provided in the embodiment of the present application can realize various processes implemented by the UE in the method embodiments in FIG. 2 to FIG. 5 and achieve the same technical effect. To avoid repetition, details are not repeated here.
  • the channel prediction device 60 may include: an acquisition module 61 and a prediction module 62 .
  • An acquisition module configured to acquire first information, where the first information includes a channel prediction result and model parameters of a target-specific operator model, or includes model parameters of a target-specific operator model; wherein, the channel prediction result is the UE based on a target-specific operator
  • the model parameters of the model and the channel results predicted by the historical channel estimation results, the target specific operator model is a specific operator model constructed by the UE based on the historical channel estimation results; the prediction module is used to perform channel prediction according to the first information obtained by the acquisition module .
  • the above model parameters include a first model order and a first model coefficient
  • the prediction module includes a determination sub-module and a prediction sub-module.
  • the determination sub-module is used to determine the target coefficient prediction component corresponding to the first model order according to the first model order;
  • the prediction sub-module is used to use the target coefficient prediction component determined by the determination sub-module to perform the calculation according to the first model coefficient Predicting the model coefficients to obtain second model coefficients; and performing channel prediction according to the second model coefficients and the first model order.
  • the determination submodule is further configured to determine the target coefficient prediction module according to the target information before determining the target coefficient prediction component corresponding to the first model order according to the first model order,
  • the target coefficient prediction module includes multiple coefficient prediction components, and the multiple coefficient prediction components include the target coefficient prediction component;
  • the target information includes at least one of the following: geographic location information of the UE, CSI measurement information sent by the UE, TAI corresponding to the UE, beam coverage of the UE, beam identifier of the UE, and scene information corresponding to the UE.
  • the first information above further includes a model coefficient prediction granularity expected by the UE, and the model coefficient prediction granularity includes at least one of the following: time domain granularity and frequency domain granularity;
  • the prediction sub-module is specifically used to use the target coefficient prediction component to predict the model coefficients according to the first model coefficients, the UE expected model coefficient prediction granularity and other UE expected model coefficient prediction granularities, and obtain the second model coefficients.
  • the time-domain granularity of the second model coefficient is any of the following: one time slot, multiple time slots, remaining time slots of the current frame, next CSI measurement moment, multiple CSI Measurement time, each time slot in a CSI measurement period, and each time slot in multiple CSI measurement periods;
  • the frequency-domain granularity of the second model coefficients is any of the following: resource block RB, sub-band, and broadband.
  • the above prediction submodule is specifically configured to use the target coefficient prediction component to use the first model coefficient and the target prediction granularity when the UE's expected model coefficient prediction granularity does not match the target prediction granularity Predicting the model coefficients to obtain the second model coefficients;
  • the target prediction granularity is the prediction granularity of the target coefficient prediction component.
  • the second model coefficient is shared by each subband/RB in the broadband forecast results for broadband;
  • the second model coefficient is the prediction result on the subband with the lowest channel quality indicator CQI among all subbands in the wideband; or ,
  • the second model coefficient is the prediction result of each RB in the subband sharing the subband;
  • the second model coefficient is the prediction result of the RB with the lowest CQI among all RBs in the subband;
  • the second model coefficients use the result of the RB with the lowest CQI among all RBs in the wideband.
  • the above-mentioned target coefficient prediction component is any one of the following: a coefficient prediction component constructed by one or more cell identifiers, a coefficient prediction component constructed by one or more transmission and reception point TRP identifiers.
  • the above channel prediction device may further include: a sending module
  • the sending module is configured to send second information to the UE before the acquiring module acquires the first information, and the second information is used to indicate a minimum number threshold of historical channel estimation results required by the UE to construct a target-specific operator model; wherein, the second The information includes at least one of the following: a parameter or a set of parameters pre-configured by the radio resource control RRC, MAC CE, and DCI.
  • An embodiment of the present application provides a channel prediction device.
  • the channel prediction device can receive the first information sent by the UE including model parameters of a specific operator model constructed by the UE based on historical channel estimation results and/or channel prediction results based on the model parameters. , so the channel prediction device can perform channel prediction according to the first information after receiving the first information; that is, the channel prediction method provided in the embodiment of the present application can be based on the model parameters of the specific operator model constructed by the UE and/or based on the specific
  • the channel prediction result predicted by the operator model indirectly predicts the channel, so that the computational complexity of channel prediction through a specific operator model can be reduced.
  • the channel measurement apparatus in this embodiment of the present application may be an electronic device, such as an electronic device with an operating system, or a component in the electronic device, such as an integrated circuit or a chip.
  • the electronic device may be a terminal, or other devices other than the terminal.
  • the terminal may include, but not limited to, the types of terminal 11 listed above, network attached storage (Network Attached Storage, NAS), etc., which are not specifically limited in this embodiment of the present application.
  • Network Attached Storage Network Attached Storage
  • the channel prediction device in this embodiment of the present application may be a device or a UE, and may also be a component, an integrated circuit, or a chip in the UE.
  • the channel measurement device provided in the embodiment of the present application can implement various processes implemented by the network side equipment in the method embodiments shown in FIG. 2 to FIG. 5 , and achieve the same technical effect. To avoid repetition, details are not repeated here.
  • this embodiment of the present application also provides a communication device 200, including a processor 201 and a memory 202, and the memory 202 stores programs or instructions that can run on the processor 201, such as
  • the communication device 200 is a UE
  • the program or instruction is executed by the processor 201
  • each step of the UE in the above channel prediction method embodiment can be realized, and the same technical effect can be achieved.
  • the communication device 200 is a network-side device
  • the program or instruction is executed by the processor 201
  • the various steps of the network-side device in the above channel prediction method embodiment can be achieved, and the same technical effect can be achieved. To avoid repetition, it is not repeated here repeat.
  • the embodiment of the present application also provides a UE, including a processor and a communication interface, wherein the processor is used to construct a target-specific operator model, and/or predict based on the model parameters of the target-specific operator model and historical channel estimation results The channel result is obtained, and the channel prediction is performed to obtain the channel prediction result; the target specific operator model is a specific operator model constructed by the UE based on the historical channel estimation result; the communication interface is used to send the first information to the network side device, the first The information includes the model parameters and/or channel prediction results.
  • FIG. 8 is a schematic diagram of a hardware structure of a UE implementing an embodiment of the present application.
  • the UE1000 includes but is not limited to: at least one of a radio frequency unit 1001, a network module 1002, an audio output unit 1003, an input unit 1004, a sensor 1005, a display unit 1006, a user input unit 1007, an interface unit 1008, a memory 1009, and a processor 1010. part parts.
  • the terminal 1000 can also include a power supply (such as a battery) for supplying power to various components, and the power supply can be logically connected to the processor 1010 through the power management system, so as to manage charging, discharging, and power consumption through the power management system. Management and other functions.
  • a power supply such as a battery
  • the UE structure shown in FIG. 8 does not limit the UE, and the UE may include more or fewer components than shown in the figure, or combine some components, or arrange different components, which will not be repeated here.
  • the input unit 1004 may include a graphics processing unit (Graphics Processing Unit, GPU) 10041 and a microphone 10042, and the graphics processor 10041 can be used by the image capture device (such as the image data of the still picture or video obtained by the camera) for processing.
  • the display unit 1006 may include a display panel 10061, and the display panel 10061 may be configured in the form of a liquid crystal display, an organic light emitting diode, or the like.
  • the user input unit 1007 includes at least one of a touch panel 10071 and other input devices 10072 .
  • the touch panel 10071 is also called a touch screen.
  • the touch panel 10071 may include two parts, a touch detection device and a touch controller.
  • Other input devices 10072 may include, but are not limited to, physical keyboards, function keys (such as volume control buttons, switch buttons, etc.), trackballs, mice, and joysticks, which will not be repeated here.
  • the radio frequency unit 1001 may transmit it to the processor 1010 for processing; in addition, the radio frequency unit 1001 may send the uplink data to the network side device.
  • the radio frequency unit 1001 includes, but is not limited to, an antenna, an amplifier, a transceiver, a coupler, a low noise amplifier, a duplexer, and the like.
  • the memory 1009 can be used to store software programs or instructions as well as various data.
  • the memory 1009 may mainly include a first storage area for storing programs or instructions and a second storage area for storing data, wherein the first storage area may store an operating system, an application program or instructions required by at least one function (such as a sound playing function, image playback function, etc.), etc.
  • memory 1009 may include volatile memory or nonvolatile memory, or, memory 1009 may include both volatile and nonvolatile memory.
  • the non-volatile memory can be read-only memory (Read-Only Memory, ROM), programmable read-only memory (Programmable ROM, PROM), erasable programmable read-only memory (Erasable PROM, EPROM), electronically programmable Erase Programmable Read-Only Memory (Electrically EPROM, EEPROM) or Flash.
  • ROM Read-Only Memory
  • PROM programmable read-only memory
  • Erasable PROM Erasable PROM
  • EPROM erasable programmable read-only memory
  • Electrical EPROM Electrical EPROM
  • EEPROM electronically programmable Erase Programmable Read-Only Memory
  • Volatile memory can be random access memory (Random Access Memory, RAM), static random access memory (Static RAM, SRAM), dynamic random access memory (Dynamic RAM, DRAM), synchronous dynamic random access memory (Synchronous DRAM, SDRAM), double data rate synchronous dynamic random access memory (Double Data Rate SDRAM, DDRSDRAM), enhanced synchronous dynamic random access memory (Enhanced SDRAM, ESDRAM), synchronous connection dynamic random access memory (Synch link DRAM , SLDRAM) and Direct Memory Bus Random Access Memory (Direct Rambus RAM, DRRAM).
  • RAM Random Access Memory
  • SRAM static random access memory
  • DRAM dynamic random access memory
  • DRAM synchronous dynamic random access memory
  • SDRAM double data rate synchronous dynamic random access memory
  • Double Data Rate SDRAM Double Data Rate SDRAM
  • DDRSDRAM double data rate synchronous dynamic random access memory
  • Enhanced SDRAM, ESDRAM enhanced synchronous dynamic random access memory
  • Synch link DRAM , SLDRAM
  • Direct Memory Bus Random Access Memory Direct Rambus
  • the processor 1010 may include one or more processing units; optionally, the processor 1010 integrates an application processor and a modem processor, wherein the application processor mainly processes operations related to the operating system, user interface, and application programs, etc., Modem processors mainly process wireless communication signals, such as baseband processors. It can be understood that the foregoing modem processor may not be integrated into the processor 1010 .
  • the radio frequency unit 1001 is configured to send first information to the network side device, the first information is used for the network side device to perform channel prediction, and the first information includes a channel prediction result and model parameters of a target-specific operator model, Or include model parameters of a target-specific operator model; wherein, the channel prediction result is a channel result predicted by the processor 1010 based on the model parameters of the target-specific operator model and historical channel estimation results, and the target-specific operator model is A specific operator model constructed by the processor 1010 based on historical channel estimation results.
  • An embodiment of the present application provides a UE.
  • the UE can send the first information including the model parameters of the specific operator model constructed by the UE based on the historical channel estimation results and/or the channel prediction results based on the model parameters to the network side device. Therefore, the network After receiving the first information, the side device may perform channel prediction according to the first information; that is, the channel prediction method provided in the embodiment of the present application may be based on the model parameters of the specific operator model constructed by the UE and/or based on the specific operator model The predicted channel prediction result indirectly predicts the channel, so that the computational complexity of channel prediction through a specific operator model can be reduced.
  • the above model parameters include a first model order and a first model coefficient
  • the first model order is used by the network side device to determine the target coefficient prediction component
  • the first model coefficient is used by the network side device based on The target coefficient prediction component performs channel prediction.
  • the radio frequency unit 1001 is further configured to receive second information sent by the network side device before sending the first information to the network side device, and the second information is used to instruct the UE to build a target-specific operator model The minimum number threshold of required historical channel estimation results; wherein, the second information includes at least one of the following: a parameter or a set of parameters pre-configured by Radio Resource Control RRC, MAC CE, and DCI.
  • the embodiment of the present application also provides a network side device, including a processor and a communication interface, wherein the communication interface is used to obtain first information, and the first information includes at least One item; wherein, the target-specific operator model is a specific operator model constructed by the UE based on historical channel estimation results, and the channel prediction result is a channel result predicted by the UE based on model parameters of the target-specific operator model and historical channel estimation results;
  • the processor is configured to perform channel prediction according to the first information.
  • the network-side device embodiment corresponds to the above-mentioned network-side device method embodiment, and each implementation process and implementation mode of the above-mentioned method embodiment can be applied to this network-side device embodiment, and can achieve the same technical effect.
  • the embodiment of the present application also provides a network side device.
  • the network side equipment is a base station 700 including: an antenna 71 , a radio frequency device 72 , a baseband device 73 , a processor 74 and a memory 75 .
  • the antenna 71 is connected to a radio frequency device 72 .
  • the radio frequency device 72 receives information through the antenna 71, and sends the received information to the baseband device 73 for processing.
  • the baseband device 73 processes the information to be sent and sends it to the radio frequency device 72
  • the radio frequency device 72 processes the received information and sends it out through the antenna 71 .
  • the method performed by the network side device in the above embodiments may be implemented in the baseband device 73, where the baseband device 73 includes a baseband processor.
  • the baseband device 73 can include at least one baseband board, for example, a plurality of chips are arranged on the baseband board, as shown in FIG.
  • the program executes the network device operations shown in the above method embodiments.
  • the network side device may also include a network interface 76, such as a common public radio interface (common public radio interface, CPRI).
  • a network interface 76 such as a common public radio interface (common public radio interface, CPRI).
  • the network side device 700 in this embodiment of the present invention also includes: instructions or programs stored in the memory 75 and operable on the processor 74, and the processor 74 calls the instructions or programs in the memory 75 to execute the various programs shown in FIG.
  • the method of module execution achieves the same technical effect, so in order to avoid repetition, it is not repeated here.
  • the embodiment of the present application also provides a readable storage medium.
  • the readable storage medium stores programs or instructions.
  • the program or instructions are executed by the processor, the various processes of the above-mentioned channel prediction method embodiments can be achieved, and the same To avoid repetition, the technical effects will not be repeated here.
  • the processor is the processor in the UE or the network side device mentioned in the above embodiments.
  • the readable storage medium includes a computer-readable storage medium, such as a computer read-only memory ROM, a random access memory RAM, a magnetic disk or an optical disk, and the like.
  • the embodiment of the present application further provides a chip, the chip includes a processor and a communication interface, the communication interface is coupled to the processor, and the processor is used to run programs or instructions to implement the above channel prediction method embodiment Each process can achieve the same technical effect, so in order to avoid repetition, it will not be repeated here.
  • the chip mentioned in the embodiment of the present application may also be called a system-on-chip, a system-on-chip, a system-on-a-chip, or a system-on-a-chip.
  • the embodiment of the present application further provides a computer program/program product, the computer program/program product is stored in a storage medium, and the computer program/program product is executed by at least one processor to implement the above embodiment of the channel prediction method
  • the computer program/program product is stored in a storage medium, and the computer program/program product is executed by at least one processor to implement the above embodiment of the channel prediction method
  • the embodiment of the present application also provides a communication system, including: a UE and a network side device, the UE can be used to perform the steps performed by the UE in the above channel prediction method embodiment, and the network side device can be used to perform the above The steps performed by the network side device in the embodiment of the channel prediction method.
  • the term “comprising”, “comprising” or any other variation thereof is intended to cover a non-exclusive inclusion such that a process, method, article or apparatus comprising a set of elements includes not only those elements, It also includes other elements not expressly listed, or elements inherent in the process, method, article, or device. Without further limitations, an element defined by the phrase “comprising a " does not preclude the presence of additional identical elements in the process, method, article, or apparatus comprising that element.
  • the scope of the methods and devices in the embodiments of the present application is not limited to performing functions in the order shown or discussed, and may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved. Functions are performed, for example, the described methods may be performed in an order different from that described, and various steps may also be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.
  • the methods of the above embodiments can be implemented by means of software plus a necessary general-purpose hardware platform, and of course also by hardware, but in many cases the former is better implementation.
  • the technical solution of the present application can be embodied in the form of computer software products, which are stored in a storage medium (such as ROM/RAM, magnetic disk, etc.) , CD-ROM), including several instructions to make a terminal (which may be a mobile phone, a computer, a server, an air conditioner, or a network device, etc.) execute the methods described in the various embodiments of the present application.

Landscapes

  • Engineering & Computer Science (AREA)
  • Physics & Mathematics (AREA)
  • Electromagnetism (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Quality & Reliability (AREA)
  • Mobile Radio Communication Systems (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

本发明实施例公开了一种信道预测方法、装置、UE及系统，涉及通信技术领域。该方法包括：UE可以向网络侧设备发送第一信息；网络侧设备可以接收第一信息，并根据第一信息进行信道预测；第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；其中，信道预测结果为所述UE基于目标特定算子模型的模型参数和历史信道估计结果预测得到的信道结果，目标特定算子模型为UE基于历史信道估计结果构建的特定算子模型；网络侧设备根据第一信息进行信道预测。

Description

信道预测方法、装置、UE及系统
相关申请的交叉引用
本申请主张在2021年11月17日在中国提交的中国专利申请号202111364923.1的优先权，其全部内容通过引用包含于此。
技术领域
本发明实施例涉及通信技术领域，尤其涉及一种信道预测方法、装置、UE及系统。
背景技术
目前，对时变信道可以使用特定算子模型(例如线性模型、多项式拟合模型等)进行递推预测，特定算子模型是一个时间滤波器，输入历史的信道估计，输出为未来等时间间隔时刻的信道；并且由于特定算子模型只需要内生变量而不需要借助其他外生变量，因此运用特定算子模型进行信道预测的建模比较简单。
然而，由于在复杂环境中运用特定算子模型进行信道预测时，特定算子模型的模型参数较多，因此导致采用特定算子模型进行信道预测的计算复杂度较高。
发明内容
本发明实施例提供一种信道预测方法、装置、UE及系统，可以解决采用特定算子模型进行信道预测的计算复杂度较高的问题。
第一方面，提供了一种信道预测方法，应用于用户设备UE(User Equipment，UE)，该方法可以包括：UE向网络侧设备发送第一信息，第一信息用于网络侧设备进行信道预测，第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；其中，信道预测结果为UE基于目标特定算子模型的模型参数和历史信道估计结果预测得到的信道结果，目标特定算子模型为UE基于历史信道估计结果构建的特定算子模型。
第二方面，提供了一种信道预测装置，包括：发送模块。发送模块，用于向网络侧设备发送第一信息，第一信息用于网络侧设备进行信道预测，所述第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；其中，信道预测结果为UE基于目标特定算子模型的模型参数和历史信道估计结果预测得到的信道结果，目标特定算子模型为UE基于历史信道估计结果构建的特定算子模型。
第三方面，提供了一种信道预测方法，应用于网络侧设备，该方法包括：网络侧设备获取第一信息，第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；其中，信道预测结果为所述UE基于目标特定算子模型的模型参数和历史信道估计结果预测得到的信道结果，目标特定算子模型为UE基于历史信道估计结果构建的特定算子模型；网络侧设备根据第一信息进行信道预测。
第四方面，提供了一种信道预测装置，包括：获取模块和预测模块。获取模块，用于获取第一信息，第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；其中，信道预测结果为UE基于目标特定算子模型的模型参数和历史信道估计结果预测得到的信道结果，目标特定算子模型为UE基于历史信道估计结果构建的特定算子模型；预测模块，用于根据获取模块获取的第一信息进行信道预测。
第五方面，提供了一种用户设备UE，该包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如第一方面所述的方法的步骤。
第六方面，提供了一种用户设备UE，包括处理器及通信接口，其中，所述处理器用于：构建目标特定算子模型，或用于构建目标特定算子模型和基于目标特定算子模型的模型参数和历史信道估计结果预测得到的信道结果，进行信道预测，得到信道预测结果；目标特定算子模型为UE基于历史信道估计结果构建的特定算子模型；所述通信接口用于向网络侧设备发送第一信息，第一信息中包括该模型参数和/或信道预测结果。
第七方面，提供了一种网络侧设备，该网络侧设备包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如第三方面所述的方法的步骤。
第八方面，提供了一种网络侧设备，包括处理器及通信接口，其中，所述通信接口用于获取第一信息，第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；其中，目标特定算子模型为UE基于历史信道估计结果构建的特定算子模型，该 信道预测结果为UE基于目标特定算子模型的模型参数和历史信道估计结果预测得到的信道结果；处理器用于根据第一信息进行信道预测。
第九方面，提供了一种通信系统，用户设备UE及网络侧设备，所述UE可用于执行如第一方面所述的信道预测方法的步骤，所述网络侧设备可用于执行如第三方面所述的信道预测方法的步骤。
第十方面，提供了一种可读存储介质，所述可读存储介质上存储程序或指令，所述程序或指令被处理器执行时实现如第一方面所述的方法的步骤，或者实现如第三方面所述的方法的步骤。
第十一方面，提供了一种芯片，所述芯片包括处理器和通信接口，所述通信接口和所述处理器耦合，所述处理器用于运行程序或指令，实现如第一方面所述的方法，或实现如第三方面所述的方法。
第十二方面，提供了一种计算机程序/程序产品，所述计算机程序/程序产品被存储在存储介质中，所述计算机程序/程序产品被至少一个处理器执行以实现如第一方面所述的信道预测方法的步骤，或者实现如第三方面所述的信道预测方法的步骤。
在本申请实施例中，UE可以向网络侧设备发送第一信息(用于网络侧设备进行信道预测)；网络侧设备可以接收第一信息，并根据第一信息进行信道预测；第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；其中，信道预测结果为所述UE基于目标特定算子模型的模型参数和历史信道估计结果预测得到的信道结果，目标特定算子模型为UE基于历史信道估计结果构建的特定算子模型；网络侧设备根据第一信息进行信道预测。通过该方案，由于UE可以向网络侧设备发送包括UE基于历史信道估计结果构建的特定算子模型的模型参数，或包括UE基于历史信道估计结果构建的特定算子模型的模型参数和UE基于该模型参数的信道预测结果的第一信息，因此网络侧设备可以在接收到第一信息后，根据第一信息进行信道预测；即本申请实施例提供的信道预测方法可以基于UE发送的模型参数，或者，基于UE发送的模型参数和信道预测结果间接预测信道，从而可以减少通过特定算子模型进行信道预测的计算复杂度。
附图说明
图1为本发明实施例提供的一种通信系统的架构示意图；
图2为本发明实施例提供的一种信道预测方法的示意图；
图3为本发明实施例提供的一种信道预测方法中通过滑动窗口在历史时域信道中采样的示意图；
图4为本发明实施例提供的一种信道预测方法的信道预测流程示意图；
图5为本发明实施例提供的一种信道预测装置的结构示意图之一；
图6为本发明实施例提供的一种信道预测装置的结构示意图之二；
图7是本申请实施例提供的一种通信设备的硬件结构示意图；
图8是本申请实施例提供的一种UE的硬件结构示意图；
图9是本申请实施例提供的一种网络侧设备的硬件结构示意图。
具体实施方式
下面将结合本申请实施例中的附图，对本申请实施例中的技术方案进行清楚描述，显然，所描述的实施例是本申请一部分实施例，而不是全部的实施例。基于本申请中的实施例，本领域普通技术人员所获得的所有其他实施例，都属于本申请保护的范围。
本申请的说明书和权利要求书中的术语“第一”、“第二”等是用于区别类似的对象，而不用于描述特定的顺序或先后次序。应该理解这样使用的术语在适当情况下可以互换，以便本申请的实施例能够以除了在这里图示或描述的那些以外的顺序实施，且“第一”、“第二”所区别的对象通常为一类，并不限定对象的个数，例如第一对象可以是一个，也可以是多个。此外，说明书以及权利要求中“和/或”表示所连接对象的至少其中之一，字符“/”一般表示前后关联对象是一种“或”的关系。
值得指出的是，本申请实施例所描述的技术不限于长期演进型(Long Term Evolution，LTE)/LTE的演进(LTE-Advanced，LTE-A)系统，还可用于其他无线通信系统，诸如码分多址(Code Division Multiple Access，CDMA)、时分多址(Time Division Multiple Access，TDMA)、频分多址(Frequency Division Multiple Access，FDMA)、正交频分多址(Orthogonal Frequency Division Multiple Access，OFDMA)、单载波频分多址(Single-carrier Frequency Division Multiple Access，SC-FDMA)和其他系统。本申请实施例中的术语“系统”和“网络”常被可互换地使用，所描述的技术既可用于以上提及的系统和无线电技术，也可用于其他系统和无线电技术。以下描述出于 示例目的描述了新空口(New Radio，NR)系统，并且在以下大部分描述中使用NR术语，但是这些技术也可应用于NR系统应用以外的应用，如第6代(6th Generation，6G)通信系统。
图1示出本申请实施例可应用的一种无线通信系统的框图。无线通信系统包括终端11和网络侧设备12。其中，终端11可以是手机、平板电脑(Tablet Personal Computer)、膝上型电脑(Laptop Computer)或称为笔记本电脑、个人数字助理(Personal Digital Assistant，PDA)、掌上电脑、上网本、超级移动个人计算机(ultra-mobile personal computer，UMPC)、移动上网装置(Mobile Internet Device，MID)、增强现实(augmented reality，AR)/虚拟现实(virtual reality，VR)设备、机器人、可穿戴式设备(Wearable Device)、车载设备(VUE)、行人终端(PUE)、智能家居(具有无线通信功能的家居设备，如冰箱、电视、洗衣机或者家具等)、游戏机、个人计算机(personal computer，PC)、柜员机或者自助机等终端侧设备，可穿戴式设备包括：智能手表、智能手环、智能耳机、智能眼镜、智能首饰(智能手镯、智能手链、智能戒指、智能项链、智能脚镯、智能脚链等)、智能腕带、智能服装等。需要说明的是，在本申请实施例并不限定终端11的具体类型。网络侧设备12可以包括接入网设备或核心网设备，其中，接入网设备12也可以称为无线接入网设备、无线接入网(Radio Access Network,RAN)、无线接入网功能或无线接入网单元。接入网设备12可以包括基站、WLAN接入点或WiFi节点等，基站可被称为节点B、演进节点B(eNB)、接入点、基收发机站(Base Transceiver Station，BTS)、无线电基站、无线电收发机、基本服务集(Basic Service Set，BSS)、扩展服务集(Extended Service Set，ESS)、家用B节点、家用演进型B节点、发送接收点(Transmitting Receiving Point，TRP)或所述领域中其他某个合适的术语，只要达到相同的技术效果，所述基站不限于特定技术词汇，需要说明的是，在本申请实施例中仅以NR系统中的基站为例进行介绍，并不限定基站的具体类型。
根据泰勒展开公式可以知道，任何非线性的东西都可以被线性的东西来拟合，所以理论上线性模型可以模拟物理世界的绝大多数现象。
假设现象P有N个属性来描述：
X＝[x 1,x 2,...,x N]。其中，X为现象P的属性集合，x N为现象P的第N个属性。
线性模型：为试图学得一个通过N个属性的线性组合来进行线性P的预测的函数，即线性模型可以为：f(X)＝(w 1x 1+b 1)+(w 2x 2+b 2)+...+(w Nx N+b N)，w N和b N分别为现象P的第N个属性的线性系数和偏移值。
线性模型以向量形式可以写成：
f(X)＝WX T+b
W＝[w 1,w 2,...,w N]
b＝[b 1,b 2,...,b N]
其中，若线性模型f(X)为实数域上的实值函数，那么上述线性模型f(X)被称为线性回归模型。理论上，若是能够解出线性模型f(X)的参数W和b，那么现象P可以通过该线性模型f(X)进行推演预测。
时变信道过程可以看作时自回归滑动平均过程，在信噪比足够的情况下可以认为完全可预测过程。对时变信道可以使用线性模型进行递推预测；此时，线性模型是一个时间滤波器，输入历史的信道估计结果，输出为未来等时间间隔时刻的信道。
信道可以用一个P阶的线性过程建模为：
Figure PCTCN2022132642-appb-000001
其中，H(n)表示第n个时刻的信道信息，P为线性模型的阶数，w(n)为白噪声，
Figure PCTCN2022132642-appb-000002
为线性模型的滤波器系数，线性模型的系数和阶数称为线性模型参数。
要利用线性模型进行信道预测必须先要求得线性模型的参数(具体为参数值)，求线性模型的参数值有很多方法，这里以最小二乘法为例，构造超定方程组：
Y＝Xa+w。
Figure PCTCN2022132642-appb-000003
Figure PCTCN2022132642-appb-000004
a＝[a 1,a 2,...,a n] T,w＝[w n+1,w n+2,...,w n+N] T；
其中，H n表示第n个时刻的信道信息，Y为需要进行拟合时刻信道信息矩阵，维度为N×1，X为用于拟合Y中时刻的历史时刻的信道信息矩阵，维度为N×n；a为线性模型的线性系数，w通常称为残差。其中，n＝P，
Figure PCTCN2022132642-appb-000005
可以理解，对于上述公式(1)，假设比如信道样本中包含连续10个时刻的信道信息，假设这10个时刻为t1，t2，t3，t4，t5，t6，t7，t8，t9，t10；又假设线性模型阶数设定为4，那么：需要利用4个历史时刻的信道信息，去拟合得到第10个时刻的信道信息，从而第10个时刻的信道信息可以表示为：H10＝[H6，H7，H8，H9]*a+w，也可以表示为H10＝[H5H6，H7，H8]*a+w，还可以表示为H10＝[H4，H5，H6，H7]*a+w，以此类推。
线性模型的系数a的最小二乘估计为：
a＝(X HX) -1X HY。
通过超定方程组，可知道残差w为：
w＝Y-Xa。
残差w代表理论与实际值之间的差别，使残差平方和达到最小时的线性模型的系数a，说明线性模型拟合数据样本最好，即可实现未来时刻的信道预测，残差平方和公式(方程)为：
Figure PCTCN2022132642-appb-000006
即可以基于残差平方和公式(方程)，确定线性模型的系数方程。可以看出，该方程中还缺少矩阵X的列数，即线性模型的阶数P；矩阵X的列数同样也可以理解为输入线性模型的历史信道样本需要多少个用于未来时刻的信道预测。
可以理解，一个好的线性模型通常要求残差较小，同时模型相对简单，即是要求线性模型的阶数较低，从而可以通过一些定阶准则来比较不同阶数情况下的线性模型的优劣。常用的定阶准则有最终预报误差准则FPE(Final Prediction Error，FPE)、赤池信息量准则AIC(Akaike information criterion，AIC)和贝叶斯信息准则BIC(Bayesian information criterion，BIC)准则函数。
结合上述论述可知，运用线性模型进行信道预测的其优点是模型十分简单，线性模型只需要内生变量而不需要借助其他外生变量；从而线性模型建模比较简单。然而，线性模型的缺点也很明显，具体为：在复杂环境中运用线性模型进行信道预测时模型参数会很大，计算量会偏大；进一步地，目前基于线性模型的信道预测存在一定局限性，只能预测同历史输入数据等间隔的未来一个时刻的信道。
在本申请实施例提供的信道预测方法中，UE可以向网络侧设备发送包括UE基于历史信道估计结果构建的特定算子模型的模型参数和/或基于该模型参数的信道预测结果的第一信息，从而网络侧设备可以在接收到第一信息后，根据第一信息进行信道预测，即网络侧设备可以基于UE发送的模型参数和/或信道预测结果间接预测信道，从而可以减少通过特定算子模型进行信道预测的计算复杂度。
下面结合附图，通过具体的实施例及其应用场景对本申请实施例提供的信息处理方法进行详细地说明。
本申请实施例提供一种信号传输方法，图2示出了本申请实施例提供的一种信道预测方法的流程图。如图2所示，本申请实施例提供的信号传输方法可以包括下述的步骤201和步骤203。
步骤201、UE向网络侧设备发送第一信息。
其中，第一信息可以用于网络侧设备进行信道预测，第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数。其中，该信道预测结果为UE基于历史信道估计结果和目标特定算子模型的模型参数预测得到的信道结果，目标特定算子模型为UE基于历史信道估计结果构建的特定算子模型。
可选地，本申请实施例中，目标特定算子模型可以为线性模型、多项式拟合模型、esprit算法模型、music算法模型，或其他任意可以用于进行信道预测的模型，具体可以根据实际使用需求确定，本申请实施例不作限定。
可选地，本申请实施例中，上述历史信道估计结果为UE采用常规手段(如通过测量CSI-RS/DMRS来估计信道)进行估计得到的信道估计结果。
可选地，本申请实施例中，UE可以多个历史信道估计结果构建目标特定算子模型。
可选地，本申请实施例中，上述模型参数可以包括第一模型阶数和第一模型系数；第一模型阶数可以用于网络侧设备确定目标系数预测组件，第一模型系数可以用于网络侧设备基于目标系数预测组件进行信道预测。
可选地，本申请实施例中，当目标特定算子模型为线性模型时，上述模型参数可以为称为线性滤波器参数，该线性滤波器参数可以包括线性滤波器的阶数(即第一模型阶数)和系数(即第一模型系数)。
本申请实施例中，UE基于目标特定算子模型进行信道预测，并上报得到的信道预测结果和目标特定算子模型的模型参数，从而可以保证UE和网络侧设备采用的模型参数保持统一，便于网络侧设备(例如基站)选择合适的目标系数预测组件进行未来时刻的模型系数预测；网络侧设备在得到未来时刻的模型系数后，可以基于得到的模型系数预测未来时刻上的信道。
可选地，本申请实施例中，目标系数预测组件可以为以下至少一项：通过一个或多个小区标识构建的系数预测组件、通过一个或多个发送和接收点TRP标识构建的系数预测组件、通过一个或多个的信道特征信息构建的系数预测组件、通过一个小区内的波束标识构建的系数预测组件、通过不同的地理位置构建的系数预测组件。
可选地，本申请实施例中，上述信道特征信息指的是通过信道状态信息CSI(Channel State Information，CSI)反馈得到的预编码矩阵指示PMI(Precoding Matrix Indicator，PMI)、信道质量指示CQI(Channel quality indication，CQI)、秩指示RI(RANK Indicator，RI)等。
可以理解，本申请实施例中，网络侧设备可以将相同信道特征信息的UE构建为一个系数预测组件。
需要说明的是，上述地理位置可以人为划分，不对地理位置的划分作具体划分方法限定。
可选地，本申请实施例中，当目标系数预测组件为通过一个或多个小区标识构建的系数预测组件时，这些小区标识中包括UE的小区的小区标识。当目标系数预测组件为通过一个或多个TRP标识构建的系数预测组件时，这些TRP标识中包括UE的TRP标识。
可选地，本申请实施例中，网络侧设备可以以一个或多个小区标识构建为单位，构建系数预测组件；或者，可以以一个或多个TRP标识为单位，构建系数预测组件。
可选地，本申请实施例中，目标系数预测组件可以为以下任一项：每个RB特有的线性系数预测组件、每个子带特有的系数预测组件、多个RB或多个子带的系数预测组件、宽带的系数预测组件。
其中，在目标系数预测组件为每个RB特有的线性系数预测组件的情况下，目标系数预测组件的频域预测粒度为RB；在目标系数预测组件为每个子带特有的线性系数预测组件的情况下，目标系数预测组件的频域预测粒度为该子带；在目标系数预测组件为多个RB特有的线性系数预测组件的情况下，目标系数预测组件的频域预测粒度为多个RB；在目标系数预测组件为多个子带特有的线性系数预测组件的情况下，目标系数预测组件的频域预测粒度为多个子带；在目标系数预测组件为宽带的线性系数预测组件的情况下，目标系数预测组件的频域预测粒度为宽带。
可选地，本申请实施例中，目标系数预测组件可以为基于神经网络(例如卷积神经网络)的系数预测组件，或为基于其他具有学习能力的任意人工智能AI(Artificial Intelligence，AI)网络的系数预测组件，具体可以根据实际使用需求确定，本申请实施例不作限定。
可选地，本申请实施例中，上述信道预测结果可以为UE在第一模型阶数小于或等于历史信道估计结果(为UE构建目标算子模型基于的历史信道估计结果)的数量(为了便于描述，以下称为目标数量)的情况下，基于第一模型参数和该历史信道估计结果进行信道预测得到的。
可以理解，本申请实施例中，第一模型阶数是否小于或等于目标数量为UE是否基于目标特定算子模型进行信道预测的信道预测条件。如果第一模型阶数小于或等于目标数量，则UE可以基于第一模型参数和该历史信道估计结果进行信道预测，并项网络侧设备发送信道预测结果和目标特定算子的模型参数；如果第一模型阶数大于目标数量，则UE本次不基于第一模型参数和该历史信道估计结果进行信道预测，且不向网络侧设备上报目标特定算子模型的模型参数，从而UE需要重新构建特定算子模型。
示例性地，假设目标特定算子模型为线性模型，那么当第一模型阶数小于或等于目标数量时，UE可以通过第一模型参数和历史信道估计结果，进行等间隔时间的信道预测。可以理解，受线性模型的特性的限制，UE预测的时间不可任意指定，只能预测从当前预测时刻起与历史数据(即上述历史信道估计结果)之间的间隔相等的时刻的信道，且只能往后预测一个时刻的信道；然后UE可以向网络侧设备发送本次预测的信道预测结果和目标特定算子的模型参数。
可选地，本申请实施例中，第一信息还可以包括UE期望的模型系数预测粒度，该模型系数预测粒度用于网络侧设备进行模型系数的预测；其中，该模型系数预测粒度可以包括以下至少一项：时域粒度、频域粒度。
本申请实施例中，由于第一信息中包括UE期望的模型系数预测粒度，因此当网络侧设备接收到第一信息之后，网络侧设备可以参考UE期望的模型预测粒度进行模型系数的预测，从而可以确保网络侧设备进行信道预测的准确性。
可选地，本申请实施例中，上述步骤201具体可以通过下述的步骤201a或步骤201b实现。
步骤201a、UE在目标资源上，向网络侧设备发送第一信息。
其中，目标资源可以包括以下至少一项：无线资源控制RRC(Radio Resource Control，RRC)预配置固定的资源、媒体接入控制-控制单元MAC CE(Medium Access Control-Control Element，MAC CE)指示的资源、下行控制信息DCI(Downlink Control Information DCI)指示的资源。
步骤201b、UE在发送CSI测量信息的资源上，向网络侧设备发送第一信息。
需要说明的是，本申请实施例中，UE在发送CSI测量信息的资源上，向网络侧设备发送第一信息可以避免因上报第一信息进行信令交互造成额外时延和资源配置。
本申请实施例中，由于可以采用在不同的资源上，向网络侧设备发送第一信息，因此可以提高发送第一信息的灵活性。
步骤202、网络侧设备获取第一信息。
对于第一信息的描述，具体可以参见上述实施例中对第一信息的相关描述，为了避免重复，此处不再赘述。
可选地，本申请实施例中，网络侧设备获取第一信息具体可以为：网络侧设备接收UE发送的第一信息；或者，网络侧设备可以采用其他手段获取第一信息(例如第一信息预先存储在网络侧设备中的)；具体可以根据实际使用需求确定，本申请实施例不作限定。
步骤203、网络侧设备根据第一信息进行信道预测。
可选地，本申请实施例中，上述模型参数中包括第一模型阶数和第一模型系数，上述步骤203具体可以通过下述的步骤A至步骤C实现。
步骤A、网络侧设备根据第一模型阶数，确定与第一模型阶数对应的目标系数预测组件。
可选地，本申请实施例中，目标系数预测组件对应的阶数与第一模型阶数相同。对于目标系数预测组件对应的阶数的描述，具体参见下述实施例中对系数预测组件对应的阶数的相关描述，为了避免重复，此处不予赘述。
步骤B、网络侧设备采用目标系数预测组件，根据第一模型系数进行模型系数的预测，得到第二模型系数。
可选地，本申请实施例中，网络侧设备可以将第一模型系数输入目标系数预测组件中，以通过目标系数预测组件和第一模型系数，对未来时刻上的模型系数进行预测，得到第二模型系数。即第二模型系数为未来时刻上的模型系数。
可选地，本申请实施例中，第二模型系数的时域粒度可以为以下任一项：一个时隙、多个时隙、当前帧的剩余时隙、CSI的测量时刻、多个CSI测量时刻、CSI测量周期内每个时隙、多个CSI测量周期内每个时隙；和/或，第二模型系数的频域粒度可以为以下任一项：资源块RB(Resource Block，RB)、子带、宽带。
本申请实施例中，在不改变目标特定算子模型(具体为线性模型)的前提下，本申请实施例提供的信道预测方法引入目标系数预测组件，实现对等间隔的多个时刻的线性模型的线性系数进行预测，进而可以得到等时间间隔的多个时刻的信道状态，如此可以提升信道预测精度。
可选地，本申请实施例中，在第一信息还包括UE期望的模型系数预测粒度的情况下，上述步骤B具体可以通过下述的步骤B1实现。
步骤B1、在第一信息还包括UE期望的模型系数预测粒度的情况下，网络侧设备采用目标系数预测组件，根据第一模型系数、UE期望的模型系数预测粒度和其他UE期望的模型系数预测粒度，进行模型系数的预测，得到第二模型系数。
对于UE期望的模型系数预测粒度的描述，具体可以参见上述步骤101中对UE期望的模型系 数预测粒度的相关描述，为了避免重复，此处不再赘述。
本申请实施例中，在第一信息还包括UE期望的模型系数预测粒度的情况下，基站(网络侧设备)可以根据第一模型参数、UE期望的模型系数预测粒度和此时来自的其他UE的线性系数预测需求(即其他UE期望的线性模型预测粒度)进行合理的粒度分配，并进行第二模型系数的预测；即网络侧设备可以综合第一模型系数、UE期望的模型系数预测粒度和其他UE期望的模型系数预测粒度，进行模型参数预测，从而可以提高网络侧设备进行模型系数的准确性和合理性。
可选地，本申请实施例中，如果第一信息中包括UE期望的模型系数预测粒度，那么网络侧设备在确定目标系数预测组件之后，可以先将UE期望的模型系数预测粒度与目标系数预测组件的目标预测粒度比对，若UE期望的模型系数预测粒度与目标预测粒度匹配(例如两者相同)，则网络侧设备可以根据UE期望的模型系数预测粒度，进行模型系数的预测，以得到第二模型系数。若UE期望的模型系数预测粒度与目标预测粒度不匹配(例如两者不同)，则网络侧设备可以根据目标预测粒度，进行模型系数的预测，以得到第二模型系数。
需要说明的是，本申请实施例中，当UE期望的模型系数预测粒度与目标预测粒度不匹配时，网络侧设备可以按照向上(宽粒度)或向下(窄粒度)兼容的策略，确定进行模型系数预测的预测粒度。
可以理解，本申请实施例中，频域粒度按照由宽至窄的顺序排列为：宽带、子带、RB；即频域粒度的大小顺序为：宽带>子带>RB。
可选地，本申请实施例中，上述步骤B具体可以通过下述的步骤B2实现。
步骤B2、网络侧设备在UE期望的模型系数预测粒度与目标预测粒度不匹配的情况下，采用目标系数预测组件，根据第一模型系数和目标预测粒度进行模型系数的预测，得到第二模型系数。
下面结合UE期望的模型系数预测粒度与目标预测粒度不匹配的5种情形，具体为下述的(a～e)，对步骤B2进行示例性地说明。
a，UE期望的模型系数预测粒度为宽带中的子带/RB，目标预测粒度为宽带；
可选地，本申请实施例中，在UE期望的模型系数预测粒度为宽带中的子带/RB、且目标预测粒度为宽带的情况下，第二模型系数为宽带内的每个子带/RB共用宽带的预测结果。
b，UE期望的模型系数预测粒度为宽带，目标预测粒度为宽带中的子带。
可选地，本申请实施例中，在UE期望的模型系数预测粒度为宽带、且目标预测粒度为宽带中的子带的情况下，第二模型系数为宽带内所有子带中信道质量指示CQI(Channel quality indication，CQI)最低的子带上的预测结果；
c，UE期望的模型系数预测粒度为子带中的RB，目标预测粒度为子带。
可选地，本申请实施例中，在UE期望的模型系数预测粒度为子带中的RB、且目标预测粒度为子带的情况下，第二模型系数为子带内的每个RB共用子带的预测结果。
d，UE期望的模型系数预测粒度为子带，目标预测粒度为子带中的RB。
可选地，本申请实施例中，在UE期望的模型系数预测粒度为子带、且目标预测粒度为子带中的RB的情况下，第二模型系数为子带内所有RB中CQI最低的RB的预测结果；
e，UE期望的模型系数预测粒度为宽带，目标预测粒度为宽带中的子带下的RB。
可选地，本申请实施例中，在UE期望的模型系数预测粒度为宽带、且目标预测粒度为宽带中的子带下的RB的情况下，第二模型系数使用宽带内所有RB中CQI最低的RB的结果。
下面再结合具体示例对步骤B进行示例性地说明。
示例性地，假设基站(即网络侧设备)收到第一信息中，UE期望的频域粒度为子带；基站可以根据当前预测业务需求的实际情况决定是否要将预测粒度定为终端所期望的粒度(即子带)。具体的，若确定的目标预测粒度充足，则基站可以采用目标系数预测组件，进行UE期望的粒度的模型系数预测；若基站确定的目标预测粒度为宽带，即目标预测粒度为宽带与UE期望的子带不匹配，则基站可以进行宽带内的每个子带共用宽带的信道预测，即得到的预测结果为宽带内的每个子带共用宽带的预测结果；也就是说，每个子带共用宽带，各自利用宽带进行时隙(slot)级别的模型系数预测。
本申请实施例中，在所述UE期望的模型系数预测粒度与所述目标预测粒度不匹配的情况下，由于网络侧设备可以基于目标预测粒度，进行模型系数预测，因此可以提高网络侧设备进行模型系数预测的成功概率。
步骤C、网络侧设备根据第二模型系数和第一模型阶数，进行信道预测。
可选地，本申请实施例中，网络侧设备可以根据第二模型系数、第一模型阶数和UE发送的信道预测结果，进行信道预测。
可选地，本申请实施例中，当目标特定算子模型为线性模型时，网络侧设备可以根据第二模型系数、第一模型阶数、UE发送的信道预测结果、UE发送的历史信道估计结果，预测未来等时间间隔的多个时刻的信道。
本申请实施例中，由于网络侧设备可以基于获取的第一模型阶数，确定与第一模型阶数对应的目标系数预测组件；且采用目标系数预测组件，根据所述第一模型系数进行模型系数的预测，得到第二模型系数；并根据第二模型系数和第一模型阶数，进行信道预测；即网络侧设备可以根据网络侧设备预测的模型系数和UE预测的模型阶数，间接进行信道预测；因此相比于相关技术中网络侧设备直接采用特定算子模型进行信道预测的方案，本申请实施例提供的信道预测方法通过预测模型系数方式来间接预测信道，减少了信道预测的复杂度。
在本申请实施例提供的信道预测方法中，由于UE可以向网络侧设备发送包括UE基于历史信道估计结果构建的特定算子模型的模型参数和/或基于该模型参数的信道预测结果的第一信息，因此网络侧设备可以在接收到第一信息后，根据第一信息进行信道预测；即本申请实施例提供的信道预测方法可以基于UE发送的模型参数和/或信道预测结果间接预测信道，从而可以减少通过特定算子模型进行信道预测的计算复杂度。
可选地，本申请实施例中，在上述步骤201之前，本申请实施例提供的信道预测方法还可以包括下述的步骤204和步骤205。
步骤204、网络侧设备向UE发送第二信息。
步骤205、UE接收第二信息。
其中，本申请实施例中，第二信息可以用于指示UE构建目标特定算子模型所需的历史信道估计结果的最小数量门限。
可选地，本申请实施例中，最小数量门限为网络侧设备在进行系数预测组件的离线训练时得到的经验值。该最小数量门限由网络侧设备根据至少一个阶数值确定，至少一个阶数值为至少一个阶数组对应的阶数值，该至少一个阶数组中的阶数的数量均满足预设条件，且每个阶数组包括的多个阶数的阶数值相同。
可选地，本申请实施例中，上述预设条件包括以下任一项：上述至少一个阶数组中的阶数的数量大于或等于预设数量阈值、上述至少一个阶数组中为按照阶数数量由小至大排列的N个阶数组中的前M个阶数组；其中，M与该至少一个阶数组的数量相同，M为正整数。
可选地，本申请实施例中，假设满足预设条件的至少一个阶数组中的阶数为N个阶数中的阶数，那么：该N个阶数包括以下任一项：
网络侧设备对UE的波束对应的系数预测组件进行离线训练时采用的所有阶数；
网络侧设备对UE所处地理区域对应的系数预测组件进行离线训练时采用的所有阶数；
网络侧设备对UE对应的跟踪区域TA对应的系数预测组件进行离线训练时采用的所有阶数；
网络侧设备对同一个信道特征信息对应的系数预测组件进行离线训练时采用的所有阶数。
需要说明的是，本申请实施例中，网络侧设备可以采用阶数相同的多个特定算子模型的模型系数，进行系数预测组件的离线训练，得到一个系数预测组件；该系数预测组件对应的阶数与该多个特定算子模型的阶数相同。
可以理解，本申请实施例中，UE在接收到第二信息之后，可以采用常规手段进行信道估计，以收集历史信道估计结果；当收集的历史信道估计结果的数量达到上述最小数量门限时，UE可以开始尝试基于收集的所有历史信道估计结果构建目标特定算子模型。
可选地，本申请实施例中，第二信息可以包括以下至少一项：RRC预配置的一个参数或一套参数、MAC CE、DCI。
可选地，本申请实施例中，当RRC预配置一个参数时，最小数量门限可以仅通过该参数进行指示，该参数即为最小数量门限。当RRC预配置一套参数，例如该套参数可以为一组阶数值，分别为：7、9、11时，第二信息还可以包括MAC CE或DCI，以通过MAC CE和DCI从该套参数中确定一个参数作为最小数量门限。
可选地，本申请实施例中，当第二信息包括：MAC CE或DCI时，上述最小数量门限可以包含在MAC CE或DCI中，UE解码DCI和MAC CE即可直接可以直接获取最小数量门限；例如，MAC CE或DCI上可以承载一个指示信息，以通过该指示信息指示最小数量门限。
本申请实施例中，MAC CE和DCI的区别在于延时，且MAC CE的延时>DCI的延时。
另外，在时延上，RRC>MAC CE>DCI，即越往上层的信令延时越高，同时可携带的信息量越多。
本申请实施例中，由于网络侧设备可以通过第二信息UE构建目标特定算子模型所需的历史 信道估计结果的最小数量门限，因此在UE接收到第二信息之后，可以在收集的历史信道估计结果的数量大于或等于该最小数量门限后再创建目标特定算子模型，从而可以节省UE的开销，并提高目标特定算子模型的准确性。
可选地，本申请实施例中，在上述步骤C之前，本申请实施例提供的信道预测方法还可以包括下述的步骤206。
步骤206、网络侧设备根据目标信息，确定目标系数预测模块。
本申请实施例中，目标系数预测模块中可以包括多个系数预测组件，该多个系数预测组件中可以包括目标系数预测组件。
其中，目标信息可以包括以下至少一项：UE的地理位置信息、UE发送的CSI测量信息、UE对应的跟踪区域标识TAI、UE的波束覆盖范围、UE的波束标识、UE对应的场景信息。
可选地，本申请实施例中，UE对应的场景信息可以为UE的地理位置信息指示的地理位置的场景信息。其中，地理位置的场景信息可以指示以下至少一项：密集建筑群场景、开阔场景(例如高速路)、乡镇场景。
可以理解，本申请实施例中，目标信息不同，网络侧设备确定的目标系数预测模块也可能不同，从而目标系数预测组件也不同，具体可以根据实际使用需求确定，本申请实施例不作限定。
为了更好地理解本申请实施例提供的信道预测方法，下面对网络侧设备对构建系数预测模型的方式进行示例性地说明。
i，考虑到系数预测模块的通用性和预测准确性，网络侧设备可以以影响信道状态的因素为标准对网络侧设备搭建/构建的系数预测组件划分(归类)，组成多个系数预测模块，每个系数预测模块中包括多个系数预测组件。系数预测组件可以根据以下的一种为标准构建系数预测模块：
根据波束覆盖范围或UE所属波束标识划分不同波束的系数预测模块；
根据TAI构建不同TA的系数预测模块；
根据UE的地理位置信息，构建不同地理位置上的系数预测模块；
根据有差别较大的CSI测量信息(结果)，如CQI、PMI，构建不同CSI测量信息的系数预测模块。
可选地，本申请实施例中，网络侧设备以最能影响信道状态的因素为标准，构建不同标准下的系数预测模块，且每个系数预测模块中可以包含不同模型阶数下的线性系数预测组件。
ii，考虑到不同模型阶数对模型系数预测的影响，网络侧设备可以构建不同模型阶数的系数预测模块。
iii，考虑到不同场景下的信道预测，网络侧设备可以构建以下的一种或多种的系数预测模块：
每个RB特有的系数预测模块，适用于密集建筑群场景；
每个子带特有的系数预测模块，适用于乡镇场景；
多个RB或多个子带的系数预测模块，适用于开阔场景；
宽带的系数预测模块，适用于开阔场景。
需要说明的是，UE的信道状态会随着周围环境的不同有明显变化，能够以最影响信道变换条件的因素为标准构建系数预测模块，而非终端特定(specific)的系数预测模块，增加了系数预测模块的通用性。
进一步地，随着UE的移动和环境的变化，网络侧设备确定的用于预测模型系数的目标系数预测模块也会对应变化，提高了预测灵活性和预测精度。
可选地，本申请实施例中，若已知预测间隔内的每个时隙上的模型系数，则本申请实施例提供的信道预测方法可以进行未来间隔内的每个时隙上的信道预测。
为了便于理解本申请实施例提供的信道预测方法，下面以特定算子模型(例如目标特定算子模型)为线性模型为例，对本申请实施例提供的信道预测方法的执行过程进行示例性地说明。
一、系数预测组件训练数据集构建：
为了简化描述，以一个beam下的单个UE历史时间的t+k个信道来简述本方案。
假设UE i的历史时域信道为：H＝[H 1,H 2,...H t,...,H t+k]，如图3所示，实线框30是一个滑动窗口，窗口大小可以配置；又假设以T(注：K>>T)大小的滑动窗口在UE i的历史信道中进行滑动。每次滑动一个slot且每滑动一次，则以当前滑动窗口内的时域信道采样值为数据样本构造线性模型的超定方程组，时域信道采样的间隔与网络侧RRC配置的CSI-RS测量周期保持一致，计算该线性模型的阶数P(T>P)及线性系数a。具体步骤为：
步骤1，假设在t时刻，采样间隔为g，采样总数为n，窗口大小为T，则滑动窗口内，具体 为图3中的加粗实现方框30内的历史信道样本集合Y 1为：
Y 1＝[H t-T,H t-T+1*g+1*1,...,H t-T+(n-1)*g+(n-1)*1]；
根据历史信道样本集合Y 1构造线性模型，根据准则函数得到线性模型的最优阶数P 1，然后计算历史信道样本集合Y 1下的线性系数：
Figure PCTCN2022132642-appb-000007
其中，
Figure PCTCN2022132642-appb-000008
为最优阶数P 1下的第1个线性系数元素，
Figure PCTCN2022132642-appb-000009
中总共P 1个线性系数元素。
步骤2，历史信道样本集合Y 1的线性系数a P2计算完成后，滑动窗口向前推移1个slot，在将滑动窗口移动1个slot之后，将滑动窗口内，具体为图3中实现方框31内的信道采样值作为另外一个历史信道样本集合Y 2；同步骤1一样，根据历史信道样本集合Y 2构造一个线性模型，并根据准则函数得到该线性模型的最优阶数P 2，然后计算历史信道样本集合Y 2下的线性系数
Figure PCTCN2022132642-appb-000010
Figure PCTCN2022132642-appb-000011
其中，
Figure PCTCN2022132642-appb-000012
为最优阶数P 2下的第1个线性系数元素，
Figure PCTCN2022132642-appb-000013
中总共P 2个线性系数元素。
步骤3，历史信道样本集合Y 2的线性系数
Figure PCTCN2022132642-appb-000014
计算完成后，滑动窗口继续向前推移1个slot。不失一般性地，当滑窗移动到第t+k时刻时，滑动窗口内，具体为图3中的虚线方框32内的历史信道样本集合Y N为：
Y N＝[H t+k-T,H t+k-T+g*1+1*1,...,H t+k-T+(n-1)*g+(n-1)*1]；
根据历史信道样本集合Y N构造一个线性模型，并根据准则函数得到该线性模型的最优阶数P N，然后计算历史信道样本集合Y N下的线性系数
Figure PCTCN2022132642-appb-000015
为：
Figure PCTCN2022132642-appb-000016
如此，可以得到N个线性模型的线性系数，分别为a P1，a P2，......，a PN。
二、系数预测组件离线训练，也称为系数预测组件创建：
将N个线性模型的线性系数中相同阶数对应的线性系数组合在一起，构造不同阶数下的线性系数集合。将每个线性阶数下的线性系数按照滑动窗口对应时刻的先后进行排序(例如：若时刻j<k，则滑动窗口在时刻j时计算的线性系数a j和滑动窗口在时刻k时计算的线性系数a k的排列顺序为[a j,a k])。然后将排序后的不同阶数下的线性系数集合分别作为AI网络模型的输入进行训练，以得到不同阶数下的系数预测组件。
具体的，假设组合后阶数P k下的线性系数集合为
Figure PCTCN2022132642-appb-000017
且线性系数集合
Figure PCTCN2022132642-appb-000018
中包括S个线性系数，则
Figure PCTCN2022132642-appb-000019
可表示为：
Figure PCTCN2022132642-appb-000020
其中，
Figure PCTCN2022132642-appb-000021
表示在线性阶数P k下第s个线性系数，
Figure PCTCN2022132642-appb-000022
可表示为：
Figure PCTCN2022132642-appb-000023
其中，
Figure PCTCN2022132642-appb-000024
表示的是在阶数P k下第s个线性系数的第p k个元素。可以理解，对于一个线性系数模型，线性系数模型的线性阶数与该线性系数模型的线性系数的元素个数相同，每个元素为一个数值。
将
Figure PCTCN2022132642-appb-000025
作为AI网络模型的训练数据集输入到AI网络模型中进行训练，得到泛化的系数预测组件。
可以看出，每个系数预测组件由相同阶数下的线性系数集合离线训练得到，且不同系数预测组件由不同阶数下的线性系数集合离线训练得到。
需要说明的是，本申请实施例中，基站可以在历史时域信道中，利用时间滑窗，构造线性模 型的线性系数集合，将线性系数集合放入AI网络模型中进行离线训练，学习任意预测时刻的线性模型系数。
三、信道预测过程：
假设网络侧设备为处于开阔场景的基站(如高速路)，由于该基站附近的障碍物少，多数是视距(LOS)的传播条件，因此多径效应不明显，不同RB之间的信道也不会有很大变化，因此处于开阔场景下的基站离线训练的系数预测模块为宽带的系数预测模块。
图4为运用线性模型进行信道预测的流程示意图，具体步骤如下：
1. UE确定线性模型(即目标特定算子模型)的模型参数。
a)UE接入基站后，基站可以通过由RRC信令(即第二信息)指示UE一个信道估计收集数量参考值N，即UE构建所述目标特定算子模型所需的历史信道估计结果的最小数量门限，N为正整数。
b)如图4所示，UE收集的历史信道估计结果的数量达到N后，UE可以利用N个历史信道估计结果尝试构造线性模型的超定方程组；然后，UE可以根据信息论的准则函数确定线性模型的阶数P(即第一模型阶数)。在确定阶数P之后，UE可以先判断阶数P是否满足信道预测条件，若阶数P满足信道预测条件，则计算线性模型的线性系数a，否则不计算线性模型的线性系数a。
具体的：
i，若P>N，则确定线性模型不满足信道预测条件，从而UE本次不进行线性模型的信道预测，具体为不进行线性模型的线性系数的计算。
ii，若P≤N，则确定线性模型满足信道预测条件，从而UE可以计算线性模型的线性系数a。
2. UE基于线性模型的信道预测。
UE确定线性模型的模型参数之后，UE可以通过线性模型的模型参数和N个历史信道估计结果，进行等间隔时间的信道预测(受线性模型的限制，预测的时间不可任意指定，只能预测从当前预测时刻起与历史数据之间的间隔相等的时刻的信道，且往后预测一个间隔时刻的信道)，具体的，如图4所示，假设N个历史信道估计结果中的第N个历史信道估计结果为t1时刻的信道结果，那么UE可以根据线性模型的模型参数和N个历史信道估计结果，进行t2时刻的信道预测，即可以得到t2时刻的信道预测结果。
3. 参数上报，即向网络侧设备发送第一信息。
UE可以在确定线性模型的线性系数之后，向网络侧设备上报线性模型的模型参数，该模型参数可以包括线性模型的阶数P和线性模型的线性系数a。
a)在上报阶段，为避免上报参数的信令交互造成额外的时延和资源配置，因此UE可以在进行CSI测量信息上报时，连同信道预测结果和线性模型的模型参数(即第一信息)也上报给基站。
b)由于终端处于高速移动，多普勒频移会造成信道在不同时刻下的相位和/或频率有较大改变，因此为了使基站更加准确地进行信道预测，UE可以在上报信息(即第一信息)中携带UE需求的线性系数预测频域粒度(即UE期望的线性系数预测粒度)，且该线性系数预测频域粒度为子带。
4. 基站进行线性系数预测。
a)基站接收UE发送的第一信息之后，可以根据UE处于哪个beam的覆盖范围(即UE的波束覆盖范围)或UE所属波束标识(即UE的波束标识)，确定目标系数预测模块，目标系数预测模块中包括多个系数预测组件。
b)基站根据UE上报的模型参数中的阶数P，确定使用目标系数预测模块中的哪个线性系数预测组件(即目标系数预测组件)，例如，目标系数预测组件为阶数P下的系数预测组件。
c)基站将UE上报的线性模型的线性系数a输入目标系数预测组件中，即使用基站最终确定的线性系数预测组件，进行当前线性参数预测时刻(例如图4中的t2时刻)的下一间隔时刻(例如图4中的t3时刻)上的线性系数预测，得到一个线性系数(即第二模型系数)。
d)UE上报的信息中包括UE需求预测的频域粒度，即UE期望的线性系数频域预测粒度，从而基站可以根据当前预测业务需求的实际情况，决定是否要将预测频域粒度确定为终端所期望的粒度。
具体的，若基站可调度的处理单元资源(即目标预测粒度)充足，例如目标预测粒度与UE期望的模型系数预测粒度相同，则基站可以确定进行UE期望的模型系数预测粒度的信道预测；然而，由于目标系数预测组件为宽带的系数预测模块中的系数预测组件，即目标系数预测组件的预测粒度(也可以称为目标预测粒度)为宽带，因此基站可以确定目标预测粒度与UE期望的模 型系数预测粒度(即子带)不匹配，从而基站可以基于目标系数预测组件，利用目标预测粒度和第一模型系数，进行时隙级别的线性系数预测，以得到第二模型系数(为一个线性系数)。
5. 线性系数(例如AI系数)预测下的信道预测。
基站采用目标系数预测组件，根据UE上报的线性模型的线性系数a，预测出下一个间隔时刻(例如图4中的t3时刻)上的线性系数；然后，基站可以利用UE上报的预测信息(例如UE基于线性模型预测的图4中的t2时刻的信道结果)和历史测量上报的信道信息(例如图4中的N个历史信道估计结果)组成阶数P要求的输入样本，对下一个间隔时刻的信道进行预测。
其中，历史测量上报的信道信息可以包括以下至少一项：UE基于模型参数和N个历史信道估计结果预测得到的信道结果、N个历史信道估计结果；其他UE历史测量上报的信道结果。
本申请实施例中，相比于相关技术中需求基站可以根据t2时刻(时隙)和更早的信道信息(例如N个历史信道估计结果)构建的模型参数后在基于构建的模型参数和这些信息信息进行未来时刻的信道预测的方案，由于本申请实施例提供的信道预测方法可以直接通过目标系数预测组件预测出第二模型系数，因此可以简化信道预测的复杂度。
本申请实施例中，由于基站可以通过能够预测线性系数的目标系数预测组件来间接预测信道，从而减少了信道预测的复杂度。
本申请实施例提供的信道预测方法，执行主体还可以为信道预测装置。本申请实施例中以信道预测装置执行信道预测方法为例，说明本申请实施例提供的信道预测装置。
图5示出了本申请实施例中涉及的信道预测装置的一种可能的结构示意图。如图5所示，该信道预测装置50可以包括：发送模块51。
其中，发送模块51，用于向网络侧设备发送第一信息，第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；其中，信道预测结果为UE基于目标特定算子模型的模型参数和历史信道估计结果预测得到的信道结果，目标特定算子模型为UE基于历史信道估计结果构建的特定算子模型。
在一种可能的实现方式中，上述模型参数包括第一模型阶数和第一模型系数，第一模型阶数用于网络侧设备确定目标系数预测组件，第一模型系数用于网络侧设备基于目标系数预测组件进行信道预测。
在一种可能的实现方式中，上述信道预测结果为UE在第一模型阶数小于或等于历史信道估计结果的数量的情况下，基于第一模型参数和历史信道估计结果进行信道预测得到的。
在一种可能的实现方式中，上述信道预测装置还包括接收模块。接收模块，用于在发送模块51向网络侧设备发送第一信息之前，接收网络侧设备发送的第二信息，第二信息用于指示UE构建目标特定算子模型所需的历史信道估计结果的最小数量门限；
其中，第二信息包括以下至少一项：RRC预配置的一个参数或一套参数、MAC CE、DCI。
在一种可能的实现方式中，上述第一信息还包括UE期望的模型系数预测粒度，模型系数预测粒度用于网络侧设备进行模型系数的预测；
其中，模型系数预测粒度包括以下至少一项：时域粒度、频域粒度。
在一种可能的实现方式中，上述发送模块51，具体用于在目标资源上，向网络侧设备发送第一信息，目标资源包括以下至少一项：RRC预配置固定的资源、MAC CE指示的资源、DCI指示的资源；
或者，
发送模块51，具体用于在发送CSI测量信息的资源上，向网络侧设备发送第一信息。
本申请实施例提供一种信道预测装置，信道预测装置可以向网络侧设备发送包括信道预测装置基于历史信道估计结果构建的特定算子模型的模型参数和/或基于该模型参数的信道预测结果的第一信息，因此网络侧设备可以在接收到第一信息后，根据第一信息进行信道预测；即本申请实施例提供的信道预测方法可以基于信道预测装置构建的特定算子模型的模型参数和/或基于该特定算子模型预测的信道预测结果间接预测信道，从而可以减少通过特定算子模型进行信道预测的计算复杂度。
本申请实施例中的信道测量装置可以是电子设备，例如具有操作系统的电子设备，也可以是电子设备中的部件，例如集成电路或芯片。该电子设备可以是终端，也可以为除终端之外的其他设备。示例性的，终端可以包括但不限于上述所列举的终端11的类型，网络附属存储器(Network Attached Storage，NAS)等，本申请实施例不作具体限定。
本申请实施例中的信道预测装置可以是装置或UE，也可以是UE中的部件、集成电路、或芯片。
本申请实施例提供的信道测量装置能够实现图2至图5的方法实施例中UE实现的各个过程，并达到相同的技术效果，为避免重复，这里不再赘述。
图6示出了本申请实施例中涉及的信道预测装置的一种可能的结构示意图。如图6所示，信道预测装置60可以包括：获取模块61和预测模块62。获取模块，用于获取第一信息，第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；其中，信道预测结果为UE基于目标特定算子模型的模型参数和历史信道估计结果预测得到的信道结果，目标特定算子模型为UE基于历史信道估计结果构建的特定算子模型；预测模块，用于根据获取模块获取的第一信息进行信道预测。
在一种可能的实现方式中，上述模型参数包括第一模型阶数和第一模型系数；预测模块包括确定子模块和预测子模块。确定子模块，用于根据第一模型阶数，确定与第一模型阶数对应的目标系数预测组件；预测子模块，用于采用确定子模块确定的目标系数预测组件，根据第一模型系数进行模型系数的预测，得到第二模型系数；并根据第二模型系数和第一模型阶数，进行信道预测。
在一种可能的实现方式中，上述确定子模块，还用于在根据第一模型阶数，确定与第一模型阶数对应的目标系数预测组件之前，根据目标信息，确定目标系数预测模块，目标系数预测模块中包括多个系数预测组件，多个系数预测组件中包括目标系数预测组件；
其中，目标信息包括以下至少一项：UE的地理位置信息、UE发送的CSI测量信息、UE对应的TAI、UE的波束覆盖范围、UE的波束标识、UE对应的场景信息。
在一种可能的实现方式中，上述第一信息还包括UE期望的模型系数预测粒度，模型系数预测粒度包括以下至少一项：时域粒度、频域粒度；
预测子模块，具体用于采用目标系数预测组件，根据第一模型系数、UE期望的模型系数预测粒度和其他UE期望的模型系数预测粒度，进行模型系数的预测，得到第二模型系数。
在一种可能的实现方式中，上述第二模型系数的时域粒度为以下任一项：一个时隙、多个时隙、当前帧的剩余时隙、下一个CSI的测量时刻、多个CSI测量时刻、CSI测量周期内每个时隙、多个CSI测量周期内每个时隙；
和/或，
第二模型系数的频域粒度为以下任一项：资源块RB、子带、宽带。
在一种可能的实现方式中，上述预测子模块，具体用于在UE期望的模型系数预测粒度与目标预测粒度不匹配的情况下，采用目标系数预测组件，根据第一模型系数和目标预测粒度进行模型系数的预测，得到第二模型系数；
其中，目标预测粒度为目标系数预测组件的预测粒度。
在一种可能的实现方式中，在UE期望的模型系数预测粒度为宽带中的子带/RB、且目标预测粒度为宽带的情况下，第二模型系数为宽带内的每个子带/RB共用宽带的预测结果；或者，
在UE期望的模型系数预测粒度为宽带、且目标预测粒度为宽带中的子带的情况下，第二模型系数为宽带内所有子带中信道质量指示CQI最低的子带上的预测结果；或者，
在UE期望的模型系数预测粒度为子带中的RB、且目标预测粒度为子带的情况下，第二模型系数为子带内的每个RB共用子带的预测结果；或者，
在UE期望的模型系数预测粒度为子带、且目标预测粒度为子带中的RB的情况下，第二模型系数为子带内所有RB中CQI最低的RB的预测结果；或者，
在UE期望的模型系数预测粒度为宽带、且目标预测粒度为宽带中的子带下的RB的情况下，第二模型系数使用宽带内所有RB中CQI最低的RB的结果。
在一种可能的实现方式中，上述目标系数预测组件为以下任一项：通过一个或多个小区标识构建的系数预测组件、通过一个或多个发送和接收点TRP标识构建的系数预测组件。
在一种可能的实现方式中，上述信道预测装置还可以包括：发送模块；
发送模块，用于在获取模块获取第一信息之前，向UE发送第二信息，第二信息用于指示UE构建目标特定算子模型所需的历史信道估计结果的最小数量门限；其中，第二信息包括以下至少一项：无线资源控制RRC预配置的一个参数或一套参数、MAC CE、DCI。
本申请实施例提供一种信道预测装置，信道预测装置可以接收UE发送的包括UE基于历史信道估计结果构建的特定算子模型的模型参数和/或基于该模型参数的信道预测结果的第一信息，因此信道预测装置可以在接收到第一信息后，根据第一信息进行信道预测；即本申请实施例提供的信道预测方法可以基于UE构建的特定算子模型的模型参数和/或基于该特定算子模型预测的信道预测结果间接预测信道，从而可以减少通过特定算子模型进行信道预测的计算复杂度。
本申请实施例中的信道测量装置可以是电子设备，例如具有操作系统的电子设备，也可以是电子设备中的部件，例如集成电路或芯片。该电子设备可以是终端，也可以为除终端之外的其他设备。示例性的，终端可以包括但不限于上述所列举的终端11的类型，网络附属存储器(Network Attached Storage，NAS)等，本申请实施例不作具体限定。
本申请实施例中的信道预测装置可以是装置或UE，也可以是UE中的部件、集成电路、或芯片。
本申请实施例提供的信道测量装置能够实现图2至图5的方法实施例中网络侧设备实现的各个过程，并达到相同的技术效果，为避免重复，这里不再赘述。
可选的，如图7所示，本申请实施例还提供一种通信设备200，包括处理器201和存储器202，存储器202上存储有可在所述处理器201上运行的程序或指令，例如，该通信设备200为UE时，该程序或指令被处理器201执行时实现上述信道预测方法实施例中UE的各个步骤，且能达到相同的技术效果。该通信设备200为网络侧设备时，该程序或指令被处理器201执行时实现上述信道预测方法实施例中网络侧设备的各个步骤，且能达到相同的技术效果，为避免重复，这里不再赘述。
本申请实施例还提供一种UE，包括处理器及通信接口，其中，所述处理器用于构建目标特定算子模型，和/或，基于目标特定算子模型的模型参数和历史信道估计结果预测得到的信道结果，进行信道预测，得到信道预测结果；目标特定算子模型为UE基于历史信道估计结果构建的特定算子模型；所述通信接口用于向网络侧设备发送第一信息，第一信息中包括该模型参数和/或信道预测结果。该UE实施例与上述UE侧方法实施例对应，上述UE侧方法实施例的各个实施过程和实现方式均可适用于该UE实施例中，且能达到相同的技术效果。具体地，图8为实现本申请实施例的一种UE的硬件结构示意图。
该UE1000包括但不限于：射频单元1001、网络模块1002、音频输出单元1003、输入单元1004、传感器1005、显示单元1006、用户输入单元1007、接口单元1008、存储器1009以及处理器1010等中的至少部分部件。
本领域技术人员可以理解，终端1000还可以包括给各个部件供电的电源(比如电池)，电源可以通过电源管理系统与处理器1010逻辑相连，从而通过电源管理系统实现管理充电、放电、以及功耗管理等功能。图8中示出的UE结构并不构成对UE的限定，UE可以包括比图示更多或更少的部件，或者组合某些部件，或者不同的部件布置，在此不再赘述。
应理解的是，本申请实施例中，输入单元1004可以包括图形处理单元(Graphics Processing Unit，GPU)10041和麦克风10042，图形处理器10041对在视频捕获模式或图像捕获模式中由图像捕获装置(如摄像头)获得的静态图片或视频的图像数据进行处理。显示单元1006可包括显示面板10061，可以采用液晶显示器、有机发光二极管等形式来配置显示面板10061。用户输入单元1007包括触控面板10071以及其他输入设备10072中的至少一种。触控面板10071，也称为触摸屏。触控面板10071可包括触摸检测装置和触摸控制器两个部分。其他输入设备10072可以包括但不限于物理键盘、功能键(比如音量控制按键、开关按键等)、轨迹球、鼠标、操作杆，在此不再赘述。
本申请实施例中，射频单元1001接收来自网络侧设备的下行数据后，可以传输给处理器1010进行处理；另外，射频单元1001可以向网络侧设备发送上行数据。通常，射频单元1001包括但不限于天线、放大器、收发信机、耦合器、低噪声放大器、双工器等。
存储器1009可用于存储软件程序或指令以及各种数据。存储器1009可主要包括存储程序或指令的第一存储区和存储数据的第二存储区，其中，第一存储区可存储操作系统、至少一个功能所需的应用程序或指令(比如声音播放功能、图像播放功能等)等。此外，存储器1009可以包括易失性存储器或非易失性存储器，或者，存储器1009可以包括易失性和非易失性存储器两者。其中，非易失性存储器可以是只读存储器(Read-Only Memory，ROM)、可编程只读存储器(Programmable ROM，PROM)、可擦除可编程只读存储器(Erasable PROM，EPROM)、电可擦除可编程只读存储器(Electrically EPROM，EEPROM)或闪存。易失性存储器可以是随机存取存储器(Random Access Memory，RAM)，静态随机存取存储器(Static RAM，SRAM)、动态随机存取存储器(Dynamic RAM，DRAM)、同步动态随机存取存储器(Synchronous DRAM，SDRAM)、双倍数据速率同步动态随机存取存储器(Double Data Rate SDRAM，DDRSDRAM)、增强型同步动态随机存取存储器(Enhanced SDRAM，ESDRAM)、同步连接动态随机存取存储器(Synch link DRAM，SLDRAM)和直接内存总线随机存取存储器(Direct Rambus RAM，DRRAM)。本申请实施例中的存储器1009包括但不限于这些和任意其它适合类型的存储器。
处理器1010可包括一个或多个处理单元；可选的，处理器1010集成应用处理器和调制解调处理器，其中，应用处理器主要处理涉及操作系统、用户界面和应用程序等的操作，调制解调处理器主要处理无线通信信号，如基带处理器。可以理解的是，上述调制解调处理器也可以不集成到处理器1010中。
其中，射频单元1001，用于向网络侧设备发送第一信息，所述第一信息用于所述网络侧设备进行信道预测，第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；其中，所述信道预测结果为处理器1010基于目标特定算子模型的模型参数和历史信道估计结果预测得到的信道结果，所述目标特定算子模型为处理器1010基于历史信道估计结果构建的特定算子模型。
本申请实施例提供一种UE，UE可以向网络侧设备发送包括UE基于历史信道估计结果构建的特定算子模型的模型参数和/或基于该模型参数的信道预测结果的第一信息，因此网络侧设备可以在接收到第一信息后，根据第一信息进行信道预测；即本申请实施例提供的信道预测方法可以基于UE构建的特定算子模型的模型参数和/或基于该特定算子模型预测的信道预测结果间接预测信道，从而可以减少通过特定算子模型进行信道预测的计算复杂度。
在一种可能的实现方式中，上述模型参数包括第一模型阶数和第一模型系数，第一模型阶数用于网络侧设备确定目标系数预测组件，第一模型系数用于网络侧设备基于目标系数预测组件进行信道预测。
在一种可能的实现方式中，射频单元1001，还用于在向网络侧设备发送第一信息之前，接收网络侧设备发送的第二信息，第二信息用于指示UE构建目标特定算子模型所需的历史信道估计结果的最小数量门限；其中，第二信息包括以下至少一项：无线资源控制RRC预配置的一个参数或一套参数、MAC CE、DCI。
本申请实施例还提供一种网络侧设备，包括处理器及通信接口，其中，所述通信接口用于获取第一信息，第一信息包括目标特定算子模型的模型参数和信道预测结果中至少一项；其中，目标特定算子模型为UE基于历史信道估计结果构建的特定算子模型，该信道预测结果为UE基于目标特定算子模型的模型参数和历史信道估计结果预测得到的信道结果；处理器用于根据第一信息进行信道预测。该网络侧设备实施例与上述网络侧设备方法实施例对应，上述方法实施例的各个实施过程和实现方式均可适用于该网络侧设备实施例中，且能达到相同的技术效果。
具体地，本申请实施例还提供了一种网络侧设备。如图9所示，该网络侧设备为基站700包括：天线71、射频装置72、基带装置73、处理器74和存储器75。天线71与射频装置72连接。在上行方向上，射频装置72通过天线71接收信息，将接收的信息发送给基带装置73进行处理。在下行方向上，基带装置73对要发送的信息进行处理，并发送给射频装置72，射频装置72对收到的信息进行处理后经过天线71发送出去。
以上实施例中网络侧设备执行的方法可以在基带装置73中实现，该基带装置73包括基带处理器。
基带装置73例如可以包括至少一个基带板，该基带板上设置有多个芯片，如图9所示，其中一个芯片例如为基带处理器，通过总线接口与存储器75连接，以调用存储器75中的程序，执行以上方法实施例中所示的网络设备操作。
该网络侧设备还可以包括网络接口76，该接口例如为通用公共无线接口(common public radio interface，CPRI)。
具体地，本发明实施例的网络侧设备700还包括：存储在存储器75上并可在处理器74上运行的指令或程序，处理器74调用存储器75中的指令或程序执行图6所示各模块执行的方法，并达到相同的技术效果，为避免重复，故不在此赘述。
本申请实施例还提供一种可读存储介质，所述可读存储介质上存储有程序或指令，该程序或指令被处理器执行时实现上述信道预测方法实施例的各个过程，且能达到相同的技术效果，为避免重复，这里不再赘述。
其中，所述处理器为上述实施例中所述的UE或网络侧设备中的处理器。所述可读存储介质，包括计算机可读存储介质，如计算机只读存储器ROM、随机存取存储器RAM、磁碟或者光盘等。
本申请实施例另提供了一种芯片，所述芯片包括处理器和通信接口，所述通信接口和所述处理器耦合，所述处理器用于运行程序或指令，实现上述信道预测方法实施例的各个过程，且能达到相同的技术效果，为避免重复，这里不再赘述。
应理解，本申请实施例提到的芯片还可以称为系统级芯片，系统芯片，芯片系统或片上系统芯片等。
本申请实施例另提供了一种计算机程序/程序产品，所述计算机程序/程序产品被存储在存储介质中，所述计算机程序/程序产品被至少一个处理器执行以实现上述信道预测方法实施例的各个过程，且能达到相同的技术效果，为避免重复，这里不再赘述。
本申请实施例还提供了一种通信系统，包括：UE及网络侧设备，所述UE可用于执行如上所述的信道预测方法实施例中UE执行的步骤，所述网络侧设备可用于执行如上所述的信道预测方法实施例中网络侧设备执行的步骤。
需要说明的是，在本文中，术语“包括”、“包含”或者其任何其他变体意在涵盖非排他性的包含，从而使得包括一系列要素的过程、方法、物品或者装置不仅包括那些要素，而且还包括没有明确列出的其他要素，或者是还包括为这种过程、方法、物品或者装置所固有的要素。在没有更多限制的情况下，由语句“包括一个……”限定的要素，并不排除在包括该要素的过程、方法、物品或者装置中还存在另外的相同要素。此外，需要指出的是，本申请实施方式中的方法和装置的范围不限按示出或讨论的顺序来执行功能，还可包括根据所涉及的功能按基本同时的方式或按相反的顺序来执行功能，例如，可以按不同于所描述的次序来执行所描述的方法，并且还可以添加、省去、或组合各种步骤。另外，参照某些示例所描述的特征可在其他示例中被组合。
通过以上的实施方式的描述，本领域的技术人员可以清楚地了解到上述实施例方法可借助软件加必需的通用硬件平台的方式来实现，当然也可以通过硬件，但很多情况下前者是更佳的实施方式。基于这样的理解，本申请的技术方案本质上或者说对现有技术做出贡献的部分可以以计算机软件产品的形式体现出来，该计算机软件产品存储在一个存储介质(如ROM/RAM、磁碟、光盘)中，包括若干指令用以使得一台终端(可以是手机，计算机，服务器，空调器，或者网络设备等)执行本申请各个实施例所述的方法。
上面结合附图对本申请的实施例进行了描述，但是本申请并不局限于上述的具体实施方式，上述的具体实施方式仅仅是示意性的，而不是限制性的，本领域的普通技术人员在本申请的启示下，在不脱离本申请宗旨和权利要求所保护的范围情况下，还可做出很多形式，均属于本申请的保护之内。

Claims (34)

  1. 一种信道预测方法，包括：
    用户设备UE向网络侧设备发送第一信息，所述第一信息用于所述网络侧设备进行信道预测，所述第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；
    其中，所述信道预测结果为所述UE基于所述模型参数和历史信道估计结果预测得到的信道结果，所述目标特定算子模型为所述UE基于历史信道估计结果构建的特定算子模型。
  2. 根据权利要求1所述的方法，其中，所述模型参数包括第一模型阶数和第一模型系数，所述第一模型阶数用于所述网络侧设备确定目标系数预测组件，所述第一模型系数用于所述网络侧设备基于所述目标系数预测组件进行信道预测。
  3. 根据权利要求2所述的方法，其中，所述信道预测结果为所述UE在所述第一模型阶数小于或等于所述历史信道估计结果的数量的情况下，基于所述第一模型参数和所述历史信道估计结果进行信道预测得到的。
  4. 根据权利要求3所述的方法，其中，所述UE向网络侧设备发送第一信息之前，所述方法还包括：
    所述UE接收所述网络侧设备发送的第二信息，所述第二信息用于指示所述UE构建所述目标特定算子模型所需的历史信道估计结果的最小数量门限；
    其中，所述第二信息包括以下至少一项：无线资源控制RRC预配置的一个参数或一套参数、媒体接入控制-控制单元MAC CE、下行控制信息DCI。
  5. 根据权利要求1至4中任一项所述的方法，其中，所述第一信息还包括所述UE期望的模型系数预测粒度，所述模型系数预测粒度用于所述网络侧设备进行模型系数的预测；
    其中，所述模型系数预测粒度包括以下至少一项：时域粒度、频域粒度。
  6. 根据权利要求1所述的方法，其中，所述UE向网络侧设备发送第一信息，包括：
    所述UE在目标资源上，向所述网络侧设备发送所述第一信息，所述目标资源包括以下至少一项：RRC预配置固定的资源、MAC CE指示的资源、DCI指示的资源；
    或者，
    所述UE在发送信道状态信息CSI测量信息的资源上，向所述网络侧设备发送所述第一信息。
  7. 一种信道预测装置，包括：发送模块；
    所述发送模块，用于向网络侧设备发送第一信息，所述第一信息用于所述网络侧设备进行信道预测，所述第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；
    其中，所述信道预测结果为所述UE基于所述模型参数和历史信道估计结果预测得到的信道结果，所述目标特定算子模型为所述UE基于历史信道估计结果构建的特定算子模型。
  8. 根据权利要求7所述的装置，其中，所述模型参数包括第一模型阶数和第一模型系数，所述第一模型阶数用于所述网络侧设备确定目标系数预测组件，所述第一模型系数用于所述网络侧设备基于所述目标系数预测组件进行信道预测。
  9. 根据权利要求8所述的装置，其中，所述信道预测结果为所述UE在所述第一模型阶数小于或等于所述历史信道估计结果的数量的情况下，基于所述第一模型参数和所述历史信道估计结果进行信道预测得到的。
  10. 根据权利要求9所述的装置，其中，还包括：接收模块；
    所述接收模块，用于在所述发送模块向所述网络侧设备发送第一信息之前，接收所述网络侧设备发送的第二信息，所述第二信息用于指示所述UE构建所述目标特定算子模型所需的历史信道估计结果的最小数量门限；
    其中，所述第二信息包括以下至少一项：无线资源控制RRC预配置的一个参数或一套参数、媒体接入控制-控制单元MAC CE、下行控制信息DCI。
  11. 根据权利要求7至10中任一项所述的装置，其中，所述第一信息还包括所述UE期望的模型系数预测粒度，所述模型系数预测粒度用于所述网络侧设备进行模型系数的预测；
    其中，所述模型系数预测粒度包括以下至少一项：时域粒度、频域粒度。
  12. 根据权利要求7所述的装置，其中，
    所述发送模块，具体用于在目标资源上，向所述网络侧设备发送所述第一信息，所述目标资源包括以下至少一项：RRC预配置固定的资源、MAC CE指示的资源、DCI指示的资源；
    或者，
    所述发送模块，具体用于在发送信道状态信息CSI测量信息的资源上，向所述网络侧设备发送所述第一信息。
  13. 一种信道预测方法，包括：
    网络侧设备获取第一信息，所述第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；
    其中，所述信道预测结果为所述UE基于所述模型参数和历史信道估计结果预测得到的信道结果，所述目标特定算子模型为所述UE基于历史信道估计结果构建的特定算子模型；
    所述网络侧设备根据所述第一信息进行信道预测。
  14. 根据权利要求13所述的方法，其中，所述模型参数包括第一模型阶数和第一模型系数；
    所述网络侧设备根据所述第一信息进行信道预测，包括：
    所述网络侧设备根据所述第一模型阶数，确定与所述第一模型阶数对应的目标系数预测组件；
    所述网络侧设备采用所述目标系数预测组件，根据所述第一模型系数进行模型系数的预测，得到第二模型系数；
    所述网络侧设备根据所述第二模型系数和所述第一模型阶数，进行信道预测。
  15. 根据权利要求14所述的方法，其中，所述网络侧设备根据所述第一模型阶数，确定与所述第一模型阶数对应的目标系数预测组件之前，所述方法还包括：
    所述网络侧设备根据目标信息，确定目标系数预测模块，所述目标系数预测模块中包括多个系数预测组件，所述多个系数预测组件中包括所述目标系数预测组件；
    其中，所述目标信息包括以下至少一项：所述UE的地理位置信息、所述UE发送的CSI测量信息、所述UE对应的跟踪区域标识TAI、UE的波束覆盖范围、UE的波束标识、所述UE对应的场景信息。
  16. 根据权利要求14所述的方法，其中，所述第一信息还包括所述UE期望的模型系数预测粒度，所述模型系数预测粒度包括以下至少一项：时域粒度、频域粒度；
    所述网络侧设备采用所述目标系数预测组件，根据所述第一模型系数进行模型系数的预测，得到第二模型系数，包括：
    所述网络侧设备采用所述目标系数预测组件，根据所述第一模型系数、所述UE期望的模型系数预测粒度和其他UE期望的模型系数预测粒度，进行模型系数的预测，得到所述第二模型系数。
  17. 根据权利要求14至16中任一项所述的方法，其中，所述第二模型系数的时域粒度为以下任一项：一个时隙、多个时隙、当前帧的剩余时隙、下一个信道状态信息CSI的测量时刻、多个CSI测量时刻、CSI测量周期内每个时隙、多个CSI测量周期内每个时隙；
    和/或，
    所述第二模型系数的频域粒度为以下任一项：资源块RB、子带、宽带。
  18. 根据权利要求14所述的方法，其中，所述网络侧设备采用所述目标系数预测组件，根据所述第一模型系数进行模型系数的预测，得到第二模型系数，包括：
    所述网络侧设备在所述UE期望的模型系数预测粒度与目标预测粒度不匹配的情况下，采用所述目标系数预测组件，根据所述第一模型系数和所述目标预测粒度进行模型系数的预测，得到所述第二模型系数；
    其中，所述目标预测粒度为所述目标系数预测组件的预测粒度。
  19. 根据权利要求18所述的方法，其中，在所述UE期望的模型系数预测粒度为宽带中的子带/RB、且所述目标预测粒度为宽带的情况下，所述第二模型系数为所述宽带内的每个子带/RB共用所述宽带的预测结果；
    在所述UE期望的模型系数预测粒度为宽带、且所述目标预测粒度为所述宽带中的子带的情况下，所述第二模型系数为所述宽带内所有子带中信道质量指示CQI最低的子带上的预测结果；
    在所述UE期望的模型系数预测粒度为子带中的RB、且所述目标预测粒度为所述子带的情况下，所述第二模型系数为所述子带内的每个RB共用所述子带的预测结果；
    在所述UE期望的模型系数预测粒度为子带、且所述目标预测粒度为所述子带中的RB的情况下，所述第二模型系数为所述子带内所有RB中CQI最低的RB的预测结果；
    在所述UE期望的模型系数预测粒度为宽带、且所述目标预测粒度为所述宽带中的子带下的RB的情况下，所述第二模型系数使用所述宽带内所有RB中CQI最低的RB的结果。
  20. 根据权利要求13所述的方法，其中，所述目标系数预测组件为以下任一项：通过一个或多个小区标识构建的系数预测组件、通过一个或多个发送和接收点TRP标识构建的系数预测组件。
  21. 根据权利要求13所述的方法，其中，所述网络侧设备获取第一信息之前，所述方法还包括：
    所述网络侧设备向所述UE发送第二信息，所述第二信息用于指示所述UE构建所述目标特定算子模型所需的历史信道估计结果的最小数量门限；
    其中，所述第二信息包括以下至少一项：无线资源控制RRC预配置的一个参数或一套参数、媒体接入控制-控制单元MAC CE、下行控制信息DCI。
  22. 一种信道预测装置，包括：获取模块和预测模块；
    所述获取模块，用于获取第一信息，所述第一信息包括信道预测结果和目标特定算子模型的模型参数，或包括目标特定算子模型的模型参数；其中，所述信道预测结果为所述UE基于所述模型参数和历史信道估计结果预测得到的信道结果，所述目标特定算子模型为所述UE基于历史信道估计结果构建的特定算子模型；
    所述预测模块，用于根据所述获取模块获取的所述第一信息进行信道预测。
  23. 根据权利要求22所述的装置，其中，所述模型参数包括第一模型阶数和第一模型系数；所述预测模块包括确定子模块和预测子模块；
    所述确定子模块，用于根据所述第一模型阶数，确定与所述第一模型阶数对应的目标系数预测组件；
    所述预测子模块，用于采用所述确定子模块确定的所述目标系数预测组件，根据所述第一模型系数进行模型系数的预测，得到第二模型系数；并根据所述第二模型系数和所述第一模型阶数，进行信道预测。
  24. 根据权利要求23所述的装置，其中，所述确定子模块，还用于在根据所述第一模型阶数，确定与所述第一模型阶数对应的目标系数预测组件之前，根据目标信息，确定目标系数预测模块，所述目标系数预测模块中包括多个系数预测组件，所述多个系数预测组件中包括所述目标系数预测组件；
    其中，所述目标信息包括以下至少一项：所述UE的地理位置信息、所述UE发送的CSI测量信息、所述UE对应的跟踪区域标识TAI、UE的波束覆盖范围、UE的波束标识、所述UE对应的场景信息。
  25. 根据权利要求23所述的装置，其中，所述第一信息还包括所述UE期望的模型系数预测粒度，所述模型系数预测粒度包括以下至少一项：时域粒度、频域粒度；
    所述预测子模块，具体用于采用所述目标系数预测组件，根据所述第一模型系数、所述UE期望的模型系数预测粒度和其他UE期望的模型系数预测粒度，进行模型系数的预测，得到所述第二模型系数。
  26. 根据权利要求23至25中任一项所述的装置，其中，所述第二模型系数的时域粒度为以下任一项：一个时隙、多个时隙、当前帧的剩余时隙、下一个信道状态信息CSI的测量时刻、多个CSI测量时刻、CSI测量周期内每个时隙、多个CSI测量周期内每个时隙；
    和/或，
    所述第二模型系数的频域粒度为以下任一项：资源块RB、子带、宽带。
  27. 根据权利要求23所述的装置，其中，所述预测子模块，具体用于在所述UE期望的模型系数预测粒度与目标预测粒度不匹配的情况下，采用所述目标系数预测组件，根据所述第一模型系数和所述目标预测粒度进行模型系数的预测，得到所述第二模型系数；
    其中，所述目标预测粒度为所述目标系数预测组件的预测粒度。
  28. 根据权利要求27所述的装置，其中，在所述UE期望的模型系数预测粒度为宽带中的子带/RB、且所述目标预测粒度为所述宽带的情况下，所述第二模型系数为所述宽带内的每个子带/RB共用所述宽带的预测结果；
    在所述UE期望的模型系数预测粒度为宽带、且所述目标预测粒度为所述宽带中的子带的情况下，所述第二模型系数为所述宽带内所有子带中信道质量指示CQI最低的子带上的预测结果；
    在所述UE期望的模型系数预测粒度为子带中的RB、且所述目标预测粒度为所述子带的情况下，所述第二模型系数为所述子带内的每个RB共用所述子带的预测结果；
    在所述UE期望的模型系数预测粒度为子带、且所述目标预测粒度为所述子带中的RB的情况下，所述第二模型系数为所述子带内所有RB中CQI最低的RB的预测结果；
    在所述UE期望的模型系数预测粒度为宽带、且所述目标预测粒度为所述宽带中的子带下的RB的情况下，所述第二模型系数使用所述宽带内所有RB中CQI最低的RB的结果。
  29. 根据权利要求22所述的装置，其中，所述目标系数预测组件为以下任一项：通过一个或 多个小区标识构建的系数预测组件、通过一个或多个发送和接收点TRP标识构建的系数预测组件。
  30. 根据权利要求22所述的装置，其中，还包括：发送模块；
    所述发送模块，用于在所述获取模块获取所述第一信息之前，向所述UE发送第二信息，所述第二信息用于指示所述UE构建所述目标特定算子模型所需的历史信道估计结果的最小数量门限；
    其中，所述第二信息包括以下至少一项：无线资源控制RRC预配置的一个参数或一套参数、媒体接入控制-控制单元MAC CE、下行控制信息DCI。
  31. 一种用户设备UE，包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如权利要求1至6中任一项所述的信道预测方法的步骤。
  32. 一种网络侧设备，包括处理器和存储器，所述存储器存储可在所述处理器上运行的程序或指令，所述程序或指令被所述处理器执行时实现如权利要求13至21中任一项所述的信道预测方法的步骤。
  33. 一种可读存储介质，所述可读存储介质上存储程序或指令，所述程序或指令被处理器执行时实现如权利要求1至6中任一项所述的信道预测方法，或者实现如权利要求13至21中任一项所述的信道预测方法的步骤。
  34. 一种通信系统，所述通信系统包括如权利要求7至12中任一项所述的信道预测装置以及如权利要求22至30中任一项所述的信道预测装置；或者，
    所述通信系统包括如权利要求31所述的用户设备UE以及如权利要求32所述的网络侧设备。
PCT/CN2022/132642 2021-11-17 2022-11-17 信道预测方法、装置、ue及系统 WO2023088387A1 (zh)

Applications Claiming Priority (2)

Application Number Priority Date Filing Date Title
CN202111364923.1A CN116137553A (zh) 2021-11-17 2021-11-17 一种信道预测方法、装置、ue及系统
CN202111364923.1 2021-11-17

Publications (1)

Publication Number Publication Date
WO2023088387A1 true WO2023088387A1 (zh) 2023-05-25

Family

ID=86332665

Family Applications (1)

Application Number Title Priority Date Filing Date
PCT/CN2022/132642 WO2023088387A1 (zh) 2021-11-17 2022-11-17 信道预测方法、装置、ue及系统

Country Status (2)

Country Link
CN (1) CN116137553A (zh)
WO (1) WO2023088387A1 (zh)

Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1921464A (zh) * 2005-08-22 2007-02-28 松下电器产业株式会社 频域信道预测算法
CN101960757A (zh) * 2008-02-28 2011-01-26 京瓷株式会社 信道信息预测系统以及信道信息预测方法
CN109076403A (zh) * 2016-04-29 2018-12-21 华为技术有限公司 无线网络中信道预测改善的测量模型优化
CN110635858A (zh) * 2019-04-09 2019-12-31 杭州电子科技大学 一种基于量子计算的自回归模型信道预测方法
CN112054825A (zh) * 2019-06-06 2020-12-08 华为技术有限公司 一种信道测量方法和通信装置
KR20210044645A (ko) * 2019-10-15 2021-04-23 삼성전자주식회사 무선 통신 시스템에서 채널을 추정하기 위한 위한 장치 및 방법
WO2021164033A1 (zh) * 2020-02-21 2021-08-26 华为技术有限公司 预测信道状态信息的方法及装置
US20210326726A1 (en) * 2020-04-16 2021-10-21 Qualcomm Incorporated User equipment reporting for updating of machine learning algorithms

Patent Citations (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN1921464A (zh) * 2005-08-22 2007-02-28 松下电器产业株式会社 频域信道预测算法
CN101960757A (zh) * 2008-02-28 2011-01-26 京瓷株式会社 信道信息预测系统以及信道信息预测方法
CN109076403A (zh) * 2016-04-29 2018-12-21 华为技术有限公司 无线网络中信道预测改善的测量模型优化
CN110635858A (zh) * 2019-04-09 2019-12-31 杭州电子科技大学 一种基于量子计算的自回归模型信道预测方法
CN112054825A (zh) * 2019-06-06 2020-12-08 华为技术有限公司 一种信道测量方法和通信装置
KR20210044645A (ko) * 2019-10-15 2021-04-23 삼성전자주식회사 무선 통신 시스템에서 채널을 추정하기 위한 위한 장치 및 방법
WO2021164033A1 (zh) * 2020-02-21 2021-08-26 华为技术有限公司 预测信道状态信息的方法及装置
US20210326726A1 (en) * 2020-04-16 2021-10-21 Qualcomm Incorporated User equipment reporting for updating of machine learning algorithms

Also Published As

Publication number Publication date
CN116137553A (zh) 2023-05-19

Similar Documents

Publication Publication Date Title
WO2023040887A1 (zh) 信息上报方法、装置、终端及可读存储介质
WO2023040885A1 (zh) 参数选择方法、参数配置方法、终端及网络侧设备
US20240267761A1 (en) Model request method, model request processing method, and related device
WO2022237822A1 (zh) 训练数据集获取方法、无线传输方法、装置及通信设备
CN116266216A (zh) 客户端筛选方法及装置、客户端及中心设备
WO2023040888A1 (zh) 数据传输方法及装置
WO2023088387A1 (zh) 信道预测方法、装置、ue及系统
Yin et al. Channel prediction with liquid time-constant networks: an online and adaptive approach
CN116415476A (zh) 模型的构建方法、装置及通信设备
CN115551073A (zh) 空间关系指示方法和设备
WO2024067281A1 (zh) Ai模型的处理方法、装置及通信设备
WO2023179540A1 (zh) 信道预测方法、装置及无线通信设备
WO2024017239A1 (zh) 数据采集方法及装置、通信设备
WO2023174325A1 (zh) Ai模型的处理方法及设备
WO2023134653A1 (zh) 通信网络预测方法、终端及网络侧设备
WO2023072239A1 (zh) 信道预测方法、装置、网络侧设备及终端
WO2024008111A1 (zh) 数据采集方法及装置
WO2023179617A1 (zh) 定位方法、装置、终端及网络侧设备
WO2024140445A1 (zh) 模型信息传输方法、装置及设备
WO2024093997A1 (zh) 确定模型适用性的方法、装置及通信设备
WO2024067280A1 (zh) 更新ai模型参数的方法、装置及通信设备
WO2023198094A1 (zh) 模型输入的确定方法及通信设备
WO2024032695A1 (zh) Csi预测处理方法、装置、通信设备及可读存储介质
WO2024208260A1 (zh) 基于ai的csi预测的数据采集方法及装置
WO2023174253A1 (zh) Ai模型的处理方法及设备

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: 22894918

Country of ref document: EP

Kind code of ref document: A1