WO2023179540A1

WO2023179540A1 - Channel prediction method and apparatus, and wireless communication device

Info

Publication number: WO2023179540A1
Application number: PCT/CN2023/082507
Authority: WO
Inventors: 吴建明; 李佳林
Original assignee: 维沃移动通信有限公司
Priority date: 2022-03-24
Filing date: 2023-03-20
Publication date: 2023-09-28
Also published as: CN116846493A

Abstract

The present application relates to the technical field of communications, and discloses a channel prediction method and apparatus, and a wireless communication device. The channel prediction method in embodiments of the present application comprises: a wireless communication device determines a target model from among at least one model on the basis of an auxiliary parameter set and a first parameter of a first target channel; and the wireless communication device predicts the first target channel on the basis of the target model, wherein each model is associated with one auxiliary parameter in the auxiliary parameter set, each auxiliary parameter is mapped with the Doppler frequency of one channel, and the first parameter is mapped with the Doppler frequency of the first target channel.

Description

Channel prediction method, device and wireless communication equipment

Cross-references to related applications

This application claims priority to Chinese Patent Application No. 202210303803.9 filed in China on March 24, 2022, the entire content of which is incorporated herein by reference.

Technical field

The present application belongs to the field of communication technology, and specifically relates to a channel prediction method, device and wireless communication equipment.

Background technique

In mobile wireless communications, the key to realizing multiple-input multiple-output (MIMO) transmission is: how to accurately feedback channel status information through the wireless communication receiving end (such as user equipment UE (User Equipment, UE)) CSI (Channel State Information, CSI) is provided to the wireless communication sending end (such as serving NR Node B (NR Node B, gNB) base station).

Specifically, the wireless communication receiving end can directly feed back CSI to the wireless communication sending end; more effectively, the wireless communication receiving end can predict the channel through a learning model, such as an artificial intelligence AI (Artificial Intelligence, AI) model, to predict the channel CSI for effective feedback. However, due to network complexity limitations, model transmission limitations, and the unpredictability of communication equipment, it is difficult for the network to train a switching learning model for each terminal; therefore, in related technologies, the network generally provides a generalized sum for all terminals. Cell-related learning models. However, it is difficult for generalized learning models to effectively improve the feedback performance of MIMO-CSI.

Contents of the invention

Embodiments of the present application provide a channel prediction method, device and wireless communication equipment, which can solve the problem that it is difficult for a generalized learning model to effectively improve the feedback performance of MIMO-CSI.

In a first aspect, a channel prediction method is provided, which is applied to a wireless communication device. The method includes: the wireless communication device determines a target model from at least one model based on an auxiliary parameter set and a first parameter of a first target channel; wireless communication The device predicts the first target channel based on the target model; wherein each model is associated with an auxiliary parameter in the auxiliary parameter set, each auxiliary parameter is mapped to the Doppler frequency of a channel, and the first parameter is mapped to the first Doppler frequency phase mapping of the target channel.

In a second aspect, a channel prediction device is provided. The device includes: a determination module and a prediction module; a determination module configured to determine a target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel; predict A module for predicting the first target channel based on the target model determined by the determination module; wherein each model is associated with an auxiliary parameter in the auxiliary parameter set, and each auxiliary parameter is mapped to the Doppler frequency of a channel , the first parameter is mapped to the Doppler frequency of the first target channel.

In a third aspect, a wireless communication device is provided. The wireless communication device includes a processor and a memory. The memory stores programs or instructions that can be run on the processor. The program or instructions are executed by the processor. When implementing the steps of the method described in the first aspect.

In a fourth aspect, a wireless communication device is provided, including a processor and a communication interface, wherein the processor is configured to determine a target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel; and Based on the target model, predict the first target channel; wherein each model is associated with an auxiliary parameter in the auxiliary parameter set, each auxiliary parameter is mapped to the Doppler frequency of a channel, and the first parameter is mapped to the first target The Doppler frequency of the channel is mapped, and the communication interface is used to obtain the first parameter.

In a fifth aspect, a readable storage medium is provided. Programs or instructions are stored on the readable storage medium. When the programs or instructions are executed by a processor, the steps of the method described in the first aspect are implemented.

In a sixth aspect, a chip is provided. The chip includes a processor and a communication interface. The communication interface is coupled to the processor. The processor is used to run programs or instructions to implement the method described in the first aspect. .

In a seventh 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 method described in the first aspect. Steps of the channel prediction method.

In this embodiment of the present application, the wireless communication device determines a target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel; and based on the target model, predicts the first target channel; wherein, Each model is associated with an auxiliary parameter in the auxiliary parameter set, each auxiliary parameter is mapped to the Doppler frequency of the one channel, and the first parameter is mapped to the Doppler frequency of the first target channel. Through this solution, when there is a need to predict the first target channel, the wireless communication device can determine the target model based on the first parameter mapped to the Doppler frequency of the first target channel, thereby ensuring assistance in target model association. The parameters are adapted to the Doppler frequency of the first target channel. For example, the auxiliary parameters used in training the target model (mapping with the Doppler frequency of a channel) match the first parameters, that is, the Doppler frequency feature can be used The target model that matches the Doppler frequency characteristics of the first target channel to be predicted predicts the first target channel. Therefore, compared with the generalized learning model used in related technologies, As for the channel prediction scheme, the channel prediction method provided by the embodiments of this application can improve the accuracy of channel prediction.

Description of the drawings

Figure 1 is one of the architectural schematic diagrams of a wireless communication system provided by an embodiment of the present application;

Figure 2 is a second architectural schematic diagram of a wireless communication system provided by an embodiment of the present application;

Figure 3 is a schematic diagram of the reasoning process based on split AI/ML;

Figure 4 is the AI functional framework related to RAN3;

Figure 5 is a schematic flowchart of a channel prediction method provided by an embodiment of the present application;

Figure 6 is a schematic diagram of the relationship between the autocorrelation function of the channel and the time standard deviation of the autocorrelation function;

Figure 7 is a schematic diagram of training a model related to the Doppler frequency f _k in the channel prediction method provided by the embodiment of the present application;

Figure 8 is a schematic diagram of the data preprocessing process of Doppler frequency 2f _k through a model related to Doppler frequency f _k in the channel prediction method provided by the embodiment of the present application;

Figure 9 is a schematic diagram of the data preprocessing process of Doppler frequency f _k /2 through a model related to Doppler frequency f _k in the channel prediction method provided by the embodiment of the present application;

Figure 10 is a schematic flow chart of the wireless communication device performing model training, data preprocessing and channel prediction based on the AI functional architecture in the channel prediction method provided by the embodiment of the present application;

Figure 11 is a schematic structural diagram of a channel prediction device provided by an embodiment of the present application;

Figure 12 is a schematic structural diagram of a wireless communication device provided by an embodiment of the present application;

Figure 13 is one of the schematic diagrams of the hardware structure of the wireless communication device provided by the embodiment of the present application;

Figure 14 is the second schematic diagram of the hardware structure of the wireless communication device provided by the embodiment of the present application.

Detailed ways

The technical solutions in the embodiments of the present application will be clearly described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, but not all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art fall within the scope of protection of this application.

The terms "first", "second", etc. in the description and claims of this application are used to distinguish similar objects and are not used to describe a specific order or sequence. It is to be understood that the terms so used are interchangeable under appropriate circumstances so that the embodiments of the present application can be practiced in sequences other than those illustrated or described herein, and that "first" and "second" are distinguished objects It is usually one type, and the number of objects is not limited. For example, the first object can be one or multiple. In addition, "and/or" in the description and claims indicates at least one of the connected objects, and the character "/" generally indicates that the related objects are in an "or" relationship.

It is worth pointing out that the technology described in the embodiments of this application is not limited to Long Term Evolution (LTE)/LTE Evolution (LTE-Advanced, LTE-A) systems, and can also be used in other wireless communication systems, such as code Code Division Multiple Access (CDMA), Time Division Multiple Access (TDMA), Frequency Division Multiple Access (FDMA), Orthogonal Frequency Division Multiple Access, OFDMA), Single-carrier Frequency Division Multiple Access (SC-FDMA) and other systems. The terms "system" and "network" in the embodiments of this application are often used interchangeably, and the described technology can be used not only for the above-mentioned systems and radio technologies, but also for other systems and radio technologies. The following description describes a New Radio (NR) system for example purposes, and NR terminology is used in much of the following description, but these techniques can also be applied to applications other than NR system applications, such as 6th ^generation Generation, 6G) communication system.

Figure 1 shows a block diagram of a wireless communication system to which embodiments of the present application are applicable. The wireless communication system includes a terminal 11 and a network side device 12. The terminal 11 may 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, or a super mobile personal computer. (ultra-mobile personal computer, UMPC), mobile Internet device (Mobile Internet Device, MID), augmented reality (AR)/virtual reality (VR) equipment, robots, wearable devices (Wearable Device) , vehicle-mounted 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 (PC), teller machines or self-service Terminal devices such as mobile phones, wearable devices include: smart watches, smart bracelets, smart headphones, smart glasses, smart jewelry (smart bracelets, smart bracelets, smart rings, smart necklaces, smart anklets, smart anklets, etc.), Smart wristbands, smart clothing, etc. It should be noted that the embodiment of the present application does not limit the specific type of the terminal 11. As shown in Figure 2, 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 a radio access network unit. The access network device 12 may include a base station, a WLAN access point or a WiFi node, etc. The base station may be called a Node B, an evolved Node B (eNB), an access point, a 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 Some other appropriate terminology in the above field, as long as the same technical effect is achieved, the base station is not limited to specific technical terms. It should be noted that in the embodiment of this application, only the base station in the NR system is used as an example for introduction, and The specific type of base station is not limited. Core network equipment may include but is not limited to at least one of the following: core network nodes, core network functions, mobility management entities (Mobility Management Entity, MME), access mobility management functions (Access and Mobility Management Function, AMF), session management functions (Session Management Function, SMF), User Plane Function (UPF), Policy Control Function (PCF), Policy and Charging Rules Function (PCRF), Edge Application Service Discovery function (Edge Application Server Discovery Function, EASDF), Unified Data Management (UDM), Unified Data Repository (UDR), Home Subscriber Server (HSS), centralized network configuration ( Centralized network configuration (CNC), Network Repository Function (NRF), Network Exposure Function (NEF), Local NEF (Local NEF, or L-NEF), Binding Support Function (Binding Support Function, BSF), application function (Application Function, AF), etc. It should be noted that in the embodiment of this application, only the core network equipment in the NR system is used as an example for introduction, and the specific type of the core network equipment is not limited.

First, some nouns or terms involved in the claims and description of this application will be explained below.

5G related AI technology

AI/Machine Learning (ML) is being used in a range of applications across industries. In mobile communication systems, mobile devices (e.g., smartphones, cars, robots) are increasingly using AI/ML models to replace traditional algorithms (e.g., speech recognition, image recognition, video processing) to more effectively support applications program. The 5G system can support at least the following three AI/ML operations 1), 2) and 3):

1), AI/ML operation splitting between AI/ML endpoints, specifically: AI operation splitting between AI endpoints, or ML operation splitting between ML endpoints.

2), the distribution and sharing of AI/ML models or data on the 5G system, specifically: the distribution and sharing of AI model data on the 5G system, or the distribution and sharing of ML models and data on the 5G system;

3), Distributed/federated learning on 5G systems.

Currently, AI/ML operations/models can be divided into multiple parts according to the current task and environment. For example, Figure 3 is a schematic diagram of the inference process based on split AI/ML. As shown in Figure 3, the AI/ML model is split into two partitions, namely the terminal device partition and the network partition.

Among them, the purpose of splitting the AI/ML model is to concentrate the computing-intensive and energy-intensive parts on the network side, while leaving the privacy-sensitive and delay-sensitive parts on the terminal device. The terminal device executes the operation/model to a specific part/level and then sends the intermediate data to the network side. The network executes the remaining parts/layers and feeds the inference results back to the device.

For specific requirements on traffic characteristics and performance of AI/ML model migration in 5G, please refer to relevant technologies. To avoid duplication, we will not go into details here.

Figure 4 shows the RAN3-related AI functional framework. As shown in Figure 4, the AI functional framework at least includes: data collection module, training module, model reasoning and behavior module (actor). The functions of the data collection module, training module, model reasoning and behavior module are described in detail below.

A data collection module (i.e., Data Collection) is a function that provides input data to the model training and model inference functions (or modules). The data collection module only performs some data preprocessing and cleaning, formatting and conversion, but does not perform AI/ML specific algorithm data preparation. Examples of input data may include measurements from user equipment UE or different network entities, feedback from actors, output from AI/ML models. The data collected by the data collection module includes training data and inference data; training data is the data required as input to the AI/ML model training function, while inference data is the data required as input to the AI/ML model inference function.

The model training module (i.e., Model Training) is a functional block that performs AI machine learning model training, validation, and testing. It can generate model performance metrics as part of the model testing process. If required, the model training function is also responsible for data preparation (e.g., data preprocessing and cleaning, formatting, and transformation) based on the training data provided by the data collection function. The model training module includes model deployment/update; model deployment/update is used to initially deploy the trained, verified and tested AI/ML model to the model inference function, or deliver the updated model to the model inference function.

A model inference module (i.e., Model Inference) is a functional block that provides AI/ML model inference output (e.g., prediction or decision-making). If required, the model inference function is also responsible for data preparation (e.g., data preprocessing and cleaning, formatting, and transformation) based on the inference data provided by the data collection function. The model inference module includes output; the output is the inference output of the AI/ML model produced by the model inference function.

Behavioral modules (i.e., Actors) are functional blocks that receive the output of the model inference function and trigger or perform corresponding actions. Behavior modules can trigger actions against other entities or themselves. Behavioral modules include feedback; feedback is information that may be needed to obtain training or inference data or performance feedback.

Space Channel Model (SCM)

The SCM model can be obtained through 12 steps, specifically:

Step 1: Set up environment, network layout and antenna array parameters.

Step 2: Assign propagation conditions, specifically line of sight (Line Of Sight, LOS) or non-line of sight (Not Line Of Sight, NLOS). It is worth noting that the propagation conditions of different base station and terminal links are uncorrelated.

Step 3: Model the path loss for each base station and terminal link.

Step 4: Generate large-scale parameters, such as taking into account delay spread (DS); angle spread (AOA, AOD, ZOA, ZOD); Ricean Factor (Ricean Factor); and shadow fading (SF). Specifically, Cholesky can be used to decompose large-scale parameter vectors to generate square root matrices.

The extended angle includes at least one of the following: angle-to-edge (AOA), AOD, ZOA, and ZOD. For detailed explanations of AOD, ZOA, and ZOD, please refer to relevant explanations of spatial channel models in related technologies.

Step 5: Randomly draw from the delay distribution to generate cluster delays.

Step 6: Generate a hypothetical single-slope exponential power delay curve and calculate the cluster power.

Step 7: Generate the angle of arrival (Angle Of Arrival, AOA) and departure angle for the azimuth and elevation angles.

Step 8: Perform intra-cluster ray coupling of azimuth and elevation angles, that is, randomly couple the AOD angle to the AOA angle within cluster n.

Step 9: Generate cross-polarization power ratio (XPR) for ray m of cluster n. XPR is lognormally distributed.

Step 10: Configure a random initial phase for ray m of cluster n and four different polarization combinations.

Step 11: Generate relevant channel coefficients for cluster n and the pair of u-th receive antenna unit and s-th transmit antenna unit.

Step 12: Assign path loss and shading to channel coefficients.

MIMO channel-related Doppler frequency estimation method

Due to the mobility of wireless communication equipment and the spatial characteristics of MIMO channels, accurate MIMO channel Doppler frequency estimation is very difficult. Therefore, among the many MIMO channel Doppler frequency estimation methods, the calculation of MIMO channel Doppler frequency before and after time is used. Correlation, rough estimation of Doppler frequency characteristics is relatively practical and effective.

If the channel response matrix of the MIMO channel at time t (time t can be represented by Orthogonal Frequency Division Multiplexing (OFDM) symbols or time slots, etc.) is H _t , then the MIMO channel at time t and time The autocorrelation function η(Δ _d ) of t-Δ _d can be calculated by the following formula 1, that is:

Among them, _Δd is the interval time of MIMO channel response, which can be represented by a time slot or OFDM symbol or time (such as millisecond).

It is worth noting that eta (Δ _d ) is the correlation parameter of the MIMO channel response and is not equivalent to the Doppler frequency of the channel. However, in general, eta (Δ _d ) has a one-to-one correspondence with the Doppler frequency of the channel. Relationship.

The channel prediction method, device and wireless communication equipment provided by the embodiments of the present application will be described in detail below with reference to the accompanying drawings through some embodiments and application scenarios.

In mobile wireless communication systems, the key to effectively realizing MIMO transmission is how to accurately feed back CSI to the wireless communication sending end (eg, gNB base station) through the wireless communication receiving end (eg, UE). Optionally, the wireless communication receiving end can directly feed back CSI to the wireless communication sending end; more effectively, the wireless communication receiving end relies on the trained AI model to perform channel prediction to monitor the channel compression process, thereby achieving MIMO channel-related CSI Provide more effective feedback.

In real networks, due to network complexity limitations, model transmission limitations and terminal unpredictability, it is difficult for the network to train AI models for each terminal. Generally, the network generally provides a generalized and cell-related AI model for all terminals. However, it is difficult for generalized AI models to effectively improve the feedback performance of MIMO-CSI information.

In the channel prediction method provided by the embodiments of the present application, model training can be carried out through channel auxiliary information (that is, the following auxiliary parameters, which can also be called channel Doppler frequency characteristics), and the channel to be predicted (for example, the first Select an appropriate model based on the channel auxiliary information (such as the first parameter described below) of the target channel), and predict the future channel response (also called channel response data) of the channel to be predicted based on the selected model. Since the model is trained based on the channel auxiliary information auxiliarily, only when the wireless communication device has the same channel auxiliary information, the model related to the channel auxiliary information will be trained or used for inference by the wireless communication device, so it can improve Accuracy and effectiveness of model use.

In the embodiment of the present application, the wireless communication device can receive a reference signal on the channel, and estimate the channel response of the corresponding channel through the reference signal, Furthermore, the wireless communication device can evaluate the channel Doppler frequency characteristics of the corresponding channel based on the reference signal. Then, on the one hand, for the corresponding channel Doppler frequency characteristics, the wireless communication device selects an AI model associated with the channel Doppler frequency characteristics, so that the wireless communication device can utilize the channel Doppler frequency characteristics and the estimated channel In response, the corresponding AI model is trained. On the other hand, for the corresponding channel Doppler frequency characteristics, the wireless communication device selects an AI model associated with the channel Doppler frequency characteristics. The wireless communication device uses the estimated channel response and the selected AI model to predict the future of the corresponding channel. The channel response is predicted. This can improve the accuracy of channel prediction.

An embodiment of the present application provides a channel prediction method. Figure 5 is a schematic flow chart of the channel prediction method provided by an embodiment of the present application. As shown in Figure 5, the channel prediction method provided by an embodiment of the present application may include the following steps 101 and Step 102. The method will be exemplified below by taking an example of a wireless communication device executing the method.

Step 101: The wireless communication device determines a target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel.

Step 102: The wireless communication device predicts the first target channel based on the target model.

Wherein, each model in the above-mentioned at least one model is associated with an auxiliary parameter in the auxiliary parameter set, each auxiliary parameter is mapped to the Doppler frequency of a channel, and the first parameter is mapped to the Doppler frequency of the first target channel. Phase mapping.

In the embodiment of the present application, the wireless communication device predicting the first target channel based on the target model may be called: the wireless communication device performs inference on the target model.

Optionally, in this embodiment of the present application, the wireless communication device may be a receiving end communication device or a sending end communication device. Alternatively, the wireless communication device may be a UE or a network side device, and the details may be determined according to actual usage requirements, which are not limited in the embodiments of this application.

It should be noted that in the embodiment of the present application, the association between the model and the auxiliary parameters can be understood as: the model is trained based on the auxiliary parameters, and each auxiliary parameter is mapped to the Doppler frequency of a channel, that is, the auxiliary parameters The Doppler frequency of the channel can be indicated; thus the model can accurately predict channels with the same auxiliary parameters, which can improve the accuracy of channel prediction.

In this embodiment of the present application, the above auxiliary parameter set includes at least one auxiliary parameter.

In the embodiment of the present application, mapping the auxiliary parameter to the Doppler frequency of the channel can be understood as: the auxiliary parameter may indicate the Doppler frequency of the channel, or the auxiliary parameter may be determined by the Doppler frequency of the channel. Correspondingly, mapping the first parameter to the newly arrived Doppler frequency of the first target can be understood as: the first parameter may indicate the newly arrived Doppler frequency of the first target, or the first parameter may be the newly arrived Doppler frequency of the first target. The Doppler frequency is determined.

Optionally, the model in the embodiment of the present application can be an AI model, an ML model, or any other model that can predict the channel. The specific model can be determined according to actual usage requirements, and is not limited in the embodiment of the present application.

It can be understood that in this embodiment of the present application, the above auxiliary parameter set includes at least one auxiliary parameter.

Optionally, in this embodiment of the present application, the auxiliary parameters associated with the target model match the first parameters. That is, the wireless communication device selects the target model based on the matching degree between the first parameter and the auxiliary parameter in the auxiliary parameter set.

Optionally, in the embodiment of the present application, in one way, the second auxiliary parameter can be the auxiliary parameter that has the greatest matching degree with the first parameter in the auxiliary parameter set, so that the optimal relationship between the target model and the target channel can be achieved. The degree of fitness enables the target model to accurately predict the first target channel. In another way, the second auxiliary parameter may be an auxiliary parameter whose matching degree with the first parameter in the auxiliary parameter set is greater than or equal to the preset matching degree. In this way, it is assumed that the auxiliary parameter set includes an auxiliary parameter whose matching degree with the first parameter is greater than or equal to the preset matching degree. Or multiple auxiliary parameters equal to the preset matching degree, then when the model associated with the auxiliary parameter with the largest matching degree to the first parameter among the multiple auxiliary parameters is occupied, the wireless communication device can select the matching first parameter The model corresponding to the auxiliary parameter with the next largest degree is determined as the target model; this can improve the flexibility of the determined model on the basis of ensuring the fitness between the determined model and the channel to be predicted.

Optionally, in this embodiment of the present application, each auxiliary parameter in the auxiliary parameter set may include at least one of the following: the Doppler frequency of a channel, the first autocorrelation function of a channel, and the first autocorrelation function of a channel. time standard deviation. The first parameter may include at least one of the following: a Doppler frequency of the first target channel, a second autocorrelation function of the first target channel, and a time standard deviation of the second autocorrelation function of the first target channel.

Optionally, in this embodiment of the present application, the wireless communication device can use the estimated channel response to calculate the autocorrelation function of the channel, thereby indirectly obtaining the Doppler frequency of the channel.

Optionally, in this embodiment of the present application, the wireless communication device may derive the time standard deviation of the autocorrelation function based on the autocorrelation function.

It can be understood that in the embodiment of the present application, the wireless communication device can at least be based on the Doppler frequency of the first target channel, the autocorrelation function, or the autocorrelation The time standard deviation of the function determines the target model. In actual implementation, the wireless communication device can also determine the target model based on any parameter that can be mapped to the Doppler frequency of the channel.

For details about the Doppler frequency estimation method of the above-mentioned one channel and the above-mentioned first target channel, please refer to the related description of the Doppler frequency estimation method related to the MIMO channel in the above terminology section. To avoid repetition, it will not be described again here.

In the embodiments of this application, the "Doppler frequency of the channel" may also be called the "Doppler frequency of the channel response". The two have the same meaning and can be interchanged. "The autocorrelation function of the channel" can also be called the "autocorrelation function of the channel response". The two have the same meaning and can be interchanged.

The channel response in the embodiment of the present application can be represented by a channel response matrix corresponding to the channel response.

Optionally, in the embodiment of the present application, for any channel, the Doppler frequency of the channel response H _t can be obtained through a variety of methods, where it is simple to obtain the Doppler frequency characteristics of the channel response matrix H _t An effective method is to calculate the correlation characteristics of the channel response matrix _Ht . Specifically, the correlation characteristics of the channel response matrix H _t can be effectively represented by the autocorrelation function η(Δ _d ) of the channel response matrix H _t . For details, see Formula 1 above.

Optionally, in this embodiment of the present application, after obtaining the autocorrelation function of the channel, the wireless communication device can derive the time standard deviation (Standard Deviation) related to the autocorrelation function based on the autocorrelation function, represented by σ.

The above step 101 will be described in detail below, taking each auxiliary parameter in the auxiliary parameter set as the standard deviation of the autocorrelation function of a channel as an example.

For example, the AI model for channel prediction is determined using the standard deviation σ as an auxiliary parameter. As shown in Figure 6, assume that the k-th AI model in at least one AI model is associated to an autocorrelation function η( _Δd ), the time standard deviation of the autocorrelation function η( _Δd ) is _σk , and wireless communication The device is UE. Then when the UE needs to predict the first target channel, the UE can first obtain the second autocorrelation function of the first target channel, and determine the time standard deviation σ _UE of the second autocorrelation function based on the second autocorrelation function. The UE may then compare σ _UE with the time standard deviation σ _k of the autocorrelation function mapped by the at least one AI model, respectively, to select the k-th model in the at least one AI model as the target model.

For example, the conditions that the selected kth AI model needs to meet are:
k＝arg min _j |σ _UE -σ _j |

Among them, j=1,2,...,K; K is the total number of AI models in at least one AI model.

It can be understood that in actual implementation, the model (such as an AI model) can also be directly associated with the time standard deviation.

Optionally, in the embodiment of this application, the auxiliary parameters associated with each model are specified by high-level configuration or protocol.

For example, taking each model in at least one model as an AI model, the time standard deviation σ _k associated with each AI model (k=1, 2,...K, K is the number of AI models) can be Through high-level network configuration, as shown in Table 1.

Table 1: Mapping relationship between AI model and corresponding time standard deviation σ _k

Optionally, in the embodiment of this application, the setting of the time standard deviation σ _k can be expressed by a proportional formula, such as σ _k =γσ _k-1 , where γ is a constant and can be set by a high-level network layer, such as γ = 1.5.

Optionally, in the embodiment of this application, the setting of the time standard deviation σ _k can also be expressed by a difference-related formula, such as σ _k =σ _k-1 + β, where β is a constant and can be set by a high-level setting, such as β =0.1.

Optionally, in this embodiment of the present application, the above-mentioned step 102 may be specifically implemented through the following step 102a.

Step 102a: The wireless communication device predicts the first target channel based on the target model and the first channel response.

Wherein, the first channel response is a channel response of the first target channel received or estimated by the receiving end communication device.

It is worth noting that when the wireless communication device is a sending communication device, the receiving communication device needs to report the received or estimated first channel response to the wireless communication device, and then the wireless communication device can based on the first channel response and the target model Predict the first target channel. When the wireless communication device is a receiving end communication device, the wireless communication device can directly predict the first target channel through the target model and the channel response received or estimated by the wireless communication device.

Optionally, in this embodiment of the present application, the number of first channel responses may be multiple, and the multiple first channel responses are continuous in the time domain.

For example, the first channel response may be the channel response matrix H t of the estimated first target channel within the quantized time domain range [nN, n], _t =n, n-1,..., nN, and N is a positive integer .

The description of estimating the first channel response will be described in detail in the following embodiments, and will not be described again in order to avoid repetition.

Optionally, in this embodiment of the present application, after the wireless communication device determines the target model, if the matching degree between the first parameter and the auxiliary parameter associated with the target model is greater than the matching threshold, the wireless communication device may adopt the following first method: A prediction method is used to predict the first target channel; if the matching degree between the first parameter and the auxiliary parameter associated with the target model is less than or equal to the matching threshold, the wireless communication device can use the following prediction method to predict the first target channel. The first target channel is predicted.

The first prediction method

The wireless communication device directly uses the first channel response as input data of the target model, and the output data of the target model is the predicted channel response to the first target channel.

The channel prediction method provided by the embodiment of the present application is exemplarily described below with reference to specific examples.

Illustratively, for a wireless communication device in the network, the target model is determined from at least one AI model, such as the kth AI model, according to the correlation characteristics (specifically, the time standard deviation) of the channel response matrix H _t of the MIMO channel. , where k=1,2,...,K; and the correlation characteristics mapped by the kth AI model and the correlation characteristics of the estimated channel response H _t matrix should be closest. Then, the wireless communication device can select the response matrix H _n , H _n-1 ,..., H _nN of the channel response of the first target channel as the first channel response, so that the first channel response can be used as the input of the kth AI model data; in this way, the future channel of the first target channel can be predicted through the channel response matrix H _n , H _n-1 ,..., H _nN and the kth AI model, and the predicted channel response can be obtained in, is the length of time for channel prediction through the kth AI model (also called the prediction interval).

The second prediction method

The wireless communication device uses the first channel response to perform data pre-processing (Pre-processing); and then uses the pre-processed channel response and the target model to predict the first target channel. Among them, the data preprocessing method for the first channel response is related to the target comparison result.

Optionally, in this embodiment of the present application, the above-mentioned step 102a may be specifically implemented through the following steps 102a1 and 102a2.

Step 102a1: When the matching degree between the auxiliary parameter associated with the target model and the first parameter is less than or equal to the first matching degree threshold, the wireless communication device performs preprocessing corresponding to the target comparison result on the first channel to obtain the second Channel response, target comparison result is the comparison result between the auxiliary parameter associated with the target model and the first parameter.

Step 102a2: The wireless communication device predicts the first target channel based on the target model and the second channel response.

Optionally, in this embodiment of the present application, the target comparison result is the ratio between the auxiliary parameter associated with the target model and the first parameter.

Optionally, in this embodiment of the present application, the wireless communication device uses an interpolation method or a data selection method to preprocess the first channel response based on the target comparison result.

For example, assuming that the target comparison result is the ratio between the auxiliary parameter associated with the target model and the first parameter, then: if the ratio is less than 1 or greater than 1, the wireless communication device can use the interpolation value based on the ratio. method to interpolate the first channel response. It can be understood that when the ratio is equal to or approximately equal to 1, it means that the matching degree of the first parameter and the auxiliary parameter associated with the target model is greater than or equal to the matching threshold, that is, there is no need to preprocess the first channel response.

For example, assume that the auxiliary parameter associated with the target model is the time standard deviation σ _k , the first parameter is the time standard deviation σ _UE , and Less than 1; and assume that the first channel response includes: channel response H _n ,...,H _nN ; then:

Assume that H _ni is the channel response of the first target channel at time ni. If it is necessary to predict the first target channel at time through the channel response H _ni channel response Then the time required to interpolate the value will be expressed as:

Among them, i=0,1,…,N.

Similarly, the time interval for channel prediction also needs to be adjusted, the adjusted time interval is expressed as: So that the preprocessed data corresponds in time to match the time interval of the channel response prediction of the selected model.

The wireless communication device can then transfer the time The channel response matrix of As input data, at time The channel response matrix of as output. It is worth noting that the input data is the channel response matrix at time t _n ,...,t _nN H _n ,…, H _nN and interpolated values are obtained.

For specific methods of interpolating the channel response matrix, please refer to relevant descriptions in related technologies. To avoid duplication, they will not be described again here.

The channel prediction method provided by the embodiment of the present application will be exemplified below with reference to specific examples.

Example 1, when the selected model (i.e., the target model) is associated with the channel Doppler frequency f _k (i.e., the auxiliary parameter associated with the target model), and the Doppler frequency of the channel to be predicted (i.e., the first target channel) is 2f _k (i.e. the first parameter). Since the ratio of Doppler frequency f _k to Doppler frequency 2f _k is equal to 1/2, the training data length (immediate domain sampling range) and prediction interval of the model adapted to the first target channel are divided into training of the target model The data length and prediction interval are 2 times, so that the wireless communication device can perform 1/2 interpolation processing on the first channel response to obtain the second channel response. Then, the wireless communication device can use the channel response data in the second channel response that is within the training data length of the target model as the target model inference input, and predict the channel response of the first target channel at a time corresponding to the prediction interval of the target model. .

Example 2, when the selected model (i.e., the target model) is related to the channel Doppler frequency f _k (i.e., the auxiliary parameter associated with the target model), and the Doppler frequency of the channel to be predicted (i.e., the first target channel) is f _k /2. Since the ratio of Doppler frequency f _k to Doppler frequency f _k /2 is equal to 2, the training data length (immediate domain sampling range) and prediction interval of the model adapted to the first target channel are divided into training of the target model Half of the data length and prediction interval, so that the wireless communication device can perform 50% selection/sampling processing on the channel response in the first channel response to obtain the second channel response. For example, the first channel response includes: channel response matrix H1, H2, H3, H4, H5, H6, H7 and H8, the second channel response includes: channel response matrices H1, H3, H5, H7. Alternatively, the second channel response may include channel response matrices H2, H4, H6, H8. Then, the wireless communication device can use the channel response matrix in the second channel response that is within the training data length of the target model as the inference input of the target model, and perform the channel response of the first target channel at the time corresponding to the prediction interval of the target model. predict.

Optionally, in the embodiment of the present application, the selection process of the channel response is associated with the selection of the channel response of the prediction interval. Figure 9 shows an example of the selection of the channel response at two different channel response prediction time points.

Optionally, in this embodiment of the present application, before the above step 101, the channel prediction method provided by the embodiment of the present application may further include the following step 103.

Step 103: The wireless communication device trains the target model based on the auxiliary parameters associated with the target model.

Optionally, in the embodiment of this application, the training of the model can be implemented through the UE or the serving NR node (NR Node B, gNB) or related communication equipment. That is, the wireless communication device may be a UE or a gNB.

In the embodiment of the present application, the wireless communication device trains the target model through channel auxiliary information (ie, auxiliary parameters), so that the target model can accurately predict a channel with parameters that match its associated auxiliary parameters.

It can be understood that in the embodiment of the present application, the target model is a model for a channel with specific Doppler frequency characteristics.

In the embodiments of the present application, different models can be trained according to different auxiliary parameters, so that the trained models will be more characteristic. Therefore, compared with the generalized model in the related art, the model trained using auxiliary parameters in the embodiment of the present application can more accurately predict channels with the same or corresponding Doppler frequency characteristics.

In order to facilitate understanding of the channel prediction method provided by the embodiment of the present application, the method of channel prediction based on auxiliary information is first described below.

In most cases, an AI supervised learning (Supervised Learning) model can be represented by a probability distribution function p(y|x;w), where x is a known data vector (can also be called a data label, that is, Label), y is the input data of the AI neural network, w is the AI neural network weight vector or coefficient vector, which needs to be obtained through AI model training.

It is worth noting that (y,x) can be regarded as the training data set for AI model training, which is represented as

Among them, (y _n ,x _n ) is the nth pair of input data samples, and N is the total number of training data samples in the training data set.

By simply using the Maximum Likelihood principle (Maximum Likelihood), that is, using the Cross Entropy between training data and model predictions as the cost function (Cost Function), the training of AI neural network parameters w can be done by solving the cost function To obtain the minimum value of J(w), where the cost function J(w) is expressed as the following formula 2:

If we assume that the training data set can be divided into training data subsets, and the k-th data subset is associated with the parameter z _k , then the training data set can be expressed as:

Among them, z can be regarded as an auxiliary parameter vector, z={z ₁ , z ₂ ,..., z _K }, and the training data subset (X _k , y _k ) can be expressed as:

in,

When the data subset (X _k , y _k ) and the data subset (X _i , Y _i ) are completely independent, that is, Where k≠i, the probability distribution function p(y|x;w) can be expressed as the following formula 3:

Therefore, the cross entropy between the training data and the training model can be expressed as the following formula 4:

Combining formula 4, it can be seen that if the auxiliary parameter vector z is known or can be estimated, the model associated with the auxiliary parameter z _k can be trained independently, and its cost function can be simplified to the following formula 5:

Among them, k=1,2,…,K, K is the total number of independent models.

It is worth noting that compared with the model method trained using Formula 2, the model method trained using Formula 5 is much simpler, and the model training will be more accurate, and the inference performance will be greatly improved.

It can be understood that the target model in the embodiment of this application is trained based on Formula 5. Among them, z _k can be the auxiliary parameters associated with the target model.

The following is an analysis and explanation of the principle that the model based on auxiliary parameter training can improve the accuracy of channel prediction.

Referring to the above description of the SCM model, steps 1 to 10 mainly generate static or semi-static parameters associated with the MIMO channel, that is, within a certain time range, the static or semi-static parameters will not change due to time or Changes occur due to changes in the environment. Step 11 generates relevant channel coefficients for cluster n and the pair of the u-th receiving antenna unit of the receiving-end communication device and the s-th transmitting antenna unit of the sending-end communication device, where u and s represent the index of the antenna. Specifically, for the nth cluster and mth ray, the relevant channel coefficient is given by the following formula 6:

in, Generate a channel response for the pair of receiving antenna unit u and transmitting antenna unit s in the nth cluster and mth ray, P _n is the received power of the nth cluster; M is the total number of rays in each cluster; F _{rx, u, θ} and F _{rx, u, φ} respectively represent the receiving antenna unit u in the direction of the spherical basis vector. and field pattern; F _tx,s,θ and F _tx,s,φ are respectively the transmitting antenna unit s in the direction of the spherical basis vector and field mode; is the position vector of the receiving antenna unit u, is the position vector of the transmitting antenna unit s, is the Cross Polarization Power Ratio in linear scale, and λ ₀ is the signal wavelength; is the moving speed of the mobile communication device; θ _n,m,ZOA is the zenith angle of the nth cluster and the mth ray (i.e., Zenith angle Of Arrival); φ _n,m,AOA is the zenith angle of the nth cluster and the mth ray Azimuth angle Of Arrival; is the phase between the zenith angles of the nth cluster and the mth ray; is the phase of the zenith angle and arrival azimuth angle of the nth cluster and mth ray; is the arrival azimuth angle and phase of the zenith angle of the nth cluster and the mth ray; θ _n,m,ZOD is the departure zenith angle (i.e., Zenith angle Of Departure) of the nth cluster and the mth ray; φ _{n , m, AOD} is the departure azimuth angle of the nth cluster and the mth ray (ie, Azimuth angle Of Departure). is the direction of the receiving antenna unit u; is the direction of the transmitting antenna unit s.

Therefore, combined with Equation 6, it can be seen that the channel response generated by the u-th receiving antenna unit and the s-th transmitting antenna unit pair in the n-th cluster is:

Among them, C _u,s,n,m are static parameters independent of time t, and f _n,m is the Doppler frequency.

Optionally, in this embodiment of the present application, the Doppler frequency of the channel is determined based on the angle of arrival of the reference signal on the channel.

Optionally, in this embodiment of the present application, when it is considered that either the sending end or the receiving end has mobility (for example, the UE in cellular network communication has mobility), the Doppler frequency f _n,m can be expressed as:

in, is the direction of the receiving antenna unit u, is a vector related to the UE’s movement speed v, expressed as: Among them, φ _v and θ _v are the traveling azimuth angle and elevation angle of the UE respectively.

Optionally, in the embodiment of this application, when considering that the sending end communication device and the receiving end communication device have mobility at the same time (for example, two UEs in V2X communication have mobility at the same time), then: the sending end communication device and the receiving end communication device Dual Mobility of terminal communication equipment must be considered. At this time, the Doppler frequency f _n,m can be expressed as:

in, and are the velocity correlation vectors of the sending end communication equipment and the receiving end communication equipment respectively, θ _v,tx and θ _v,rx are the traveling azimuth angles of the sending end and receiving end respectively, φ _v,tx and φ _{v,rx are the sending end and φ v,rx} respectively. The elevation angle of the receiving end; is the direction of the receiving antenna unit u; is the direction of the transmitting antenna unit s.

If the model is trained as a channel prediction, the channel prediction process is to predict the channel response at time t+Δ through the channel response at time t, and Δ is a relatively small time period, then the trained model is fully capable Obtain more accurate information related to the static parameters C _u,s,n,m in Equation 7. Therefore, a model for channel prediction can be trained based on auxiliary parameters mapped to the Doppler frequency of the channel, so that the model can better predict the channel.

Optionally, in this embodiment of the present application, the auxiliary parameters associated with the target model are mapped to the Doppler frequency of the second target channel, and the above step 103 can be implemented specifically through the following step 103a.

Step 103a: The wireless communication device trains the target model based on the auxiliary parameters associated with the target model and the channel response of the second target channel.

Optionally, in this embodiment of the present application, the second target channel and the first target channel may be the same or different.

It can be understood that the Doppler frequency characteristics of the first target channel match the Doppler frequency characteristics of the second target channel.

Optionally, in this embodiment of the present application, when the wireless communication device is a sending communication device, the channel response of the second target channel is reported by the receiving device, or the wireless communication device estimates it based on the information reported by the receiving device, which information is Information received by the receiving end device through the second target channel. When the wireless communication device is a receiving end communication device, the wireless communication device directly trains the target model through the received/estimated channel response.

It should be noted that in the embodiment of the present application, model training (training) is performed based on known channel responses. Specifically, it is assumed that the channel response of the second target channel may include a fourth channel response and a fifth channel response, and the fourth channel response is the second target channel at time arrive channel response. The fifth channel response is the channel response of the second target channel at time _tn . The wireless communication device may use the fourth channel response as input data for model training and the fifth channel response as a model training label. The model is trained or the AI model is updated by comparing the model's output data with the label data.

It is worth noting that the target model training needs to be carried out in advance and updated when needed. For example, when the response characteristics of the second target channel change, the target model can be updated.

The model training process is described in detail below.

1). For K AI models used for channel prediction, each AI model is associated with an auxiliary parameter mapped to a Doppler frequency (i.e., z _k expressed in Formula 5); for example, each AI model Associated with the autocorrelation function η(Δ _d ) of a channel or the time standard deviation of the autocorrelation function of the channel.

The following is an example of the correlation between each AI model and the time standard deviation of the autocorrelation function of a channel.

2). For a wireless communication device in the network, at time t, estimate the channel response matrix H _t of the second target channel, where t=n, n-1,..., nL.

3). According to the autocorrelation function of the channel response matrix H _t , estimate the time standard deviation of the channel response matrix H _t .

4). According to the time standard deviation of the channel response matrix H _t , select the AI model that needs to be trained (i.e., the target model), such as the k-th AI model, where k=1, 2,...,K.

It is worth noting that the time standard deviation associated with the selected kth AI model should be closest to the time standard deviation of the channel response matrix H _t .

4) Sampling the channel response (matrix) of the second target channel to obtain the channel response matrix The sampled channel response matrix as the input data of the kth AI model; and the channel response matrix H _n of the second target channel as the kth Label data of the AI model; to train or fine-tune the kth AI model through the input data and the label data. in, is the measurement interval for channel response prediction through the kth AI model.

In the embodiment of the present application, since the auxiliary parameters associated with the target model are mapped to the Doppler frequency of the second target channel, it can be ensured that the training data of the target model is adapted to the auxiliary parameters associated with the target model, thereby improving the target Model pertinence.

In the channel prediction method provided by the embodiment of the present application, when there is a need to predict the first target channel, since the wireless communication device can determine the target model based on the first parameter mapped to the Doppler frequency of the first target channel, This ensures that the auxiliary parameters associated with the target model are adapted to the Doppler frequency of the first target channel. For example, the auxiliary parameters used in training the target model (mapping with the Doppler frequency of a channel) match the first parameters, That is, a target model whose Doppler frequency characteristics match the Doppler frequency characteristics of the first target channel to be predicted can be used to predict the first target channel. Therefore, compared with the use of a generalized learning model in related technologies to predict the channel For prediction solutions, the channel prediction method provided by the embodiments of this application can improve the accuracy of channel prediction.

Optionally, in this embodiment of the present application, after the above-mentioned step 103, the channel prediction method provided by the embodiment of the present application may further include the following steps 104 and 105.

Step 104: The wireless communication device performs performance evaluation on the target model.

Step 105: When the performance evaluation result of the target model does not meet the performance requirements, the wireless communication device updates or fine-tunes the model parameters of the target model;

The model parameters of the target model include at least one of the following: a time domain sampling range of the target model, a prediction interval of the target model, and a sampling interval of the target model.

In the embodiment of the present application, the wireless communication device performs a performance evaluation on the target model as follows: the wireless communication device can associate the auxiliary parameters of the target model with the parameters of the channel that have been predicted (the parameters are related to the Doppler frequency of the channel that has been predicted). Phase mapping) is compared to obtain the first comparison result.

It can be understood that the performance evaluation of the model is performed after the channel prediction is completed through the model. Specifically, the predicted channel can be compared with the received channel at the prediction time of the predicted channel to obtain the first comparison result. That is, the performance of the model is evaluated for the same time by comparing the channel response received at that time with the channel response predicted based on the model for the channel at that time.

Optionally, in this embodiment of the present application, if the first comparison result is greater than or equal to the prediction tolerance threshold (preset, or protocol agreement), that is, it does not meet the performance evaluation requirements, the wireless communication device starts the process of updating the model parameters of the target model.

Optionally, in the embodiment of this application, the wireless communication device can evaluate the performance of the target model when the network environment changes, such as suddenly building a temporary building; or because the target model is not trained well enough. , determine whether the model parameters of the target model need to be updated or fine-tuned.

In the embodiment of the present application, since the wireless communication device evaluates the performance of the target model, and updates the model parameters of the target model when the evaluation result does not meet the performance requirements, it can be ensured that the target model can adapt to changes in the channel, so that it can Make accurate predictions for specific channels through the target model.

Optionally, in this embodiment of the present application, before the above step 101, the channel prediction method provided by the embodiment of the present application may further include the following step 106.

Step 106: The wireless communication device performs parameter estimation on the first target channel based on the target information to obtain the first parameters;

The target information includes at least one of the following: a reference signal on the first target channel, a third channel response of the first target channel estimated or received by the receiving end communication device.

Optionally, in this embodiment of the present application, the third channel response and the first channel response may be the same or different.

Optionally, in this embodiment of the present application, taking the target information as a reference signal on the first target channel as an example, the wireless communication device can estimate the channel response of the first target channel based on the reference signal, and then use the estimated channel response , perform channel estimation on the first target channel to obtain the first parameter.

Optionally, in the embodiment of the present application, taking the target information as the third channel response as an example, the wireless communication device selects the channel response matrix of the MIMO channel (i.e., the first target channel) at time t to be H _t (i.e., the third channel response), where t=n,n-1,...,nL. The wireless communication device may then determine an autocorrelation function of the first target channel based on the third channel response, and estimate a time standard deviation of the first target channel based on the autocorrelation function.

In the embodiment of the present application, since the first parameter is estimated by the wireless communication device based on target information, the accuracy of the first parameter can be improved, thereby improving the accuracy of determining the target model, thereby improving the accuracy of channel prediction.

For example, when the selected model (i.e., the target model) is associated with the channel Doppler frequency f _k (i.e., the auxiliary parameter associated with the target model), and the Doppler frequency of the channel that needs to be predicted (i.e., the first target channel) is 2f _k (that is, the first parameter); that is, the matching degree between the auxiliary parameter associated with the target model and the first parameter is less than or equal to the first matching degree threshold (for example, the first matching degree threshold is 100%).

Figure 7 shows a schematic diagram of the training of the model related to the Doppler frequency f _k . When training the model, the wireless communication device first sets the training data length (i.e. domain sampling range) and prediction interval related to the Doppler frequency _fk . Then, the wireless communication device collects the channel response data (i.e., the second channel response) within the training data length as the input data of the model, and collects the channel response matrix Hn of the prediction interval as the label of the model training, so that the model can be trained train.

Figure 8 shows a schematic diagram of the data preprocessing process of Doppler frequency 2f _k through a model related to Doppler frequency f _k . Among them, (a) in FIG. 8 is a schematic diagram of the first channel response, and (b) in FIG. 8 is a schematic diagram of the second channel response.

The wireless communication device first sets the training data length and prediction interval related to the Doppler frequency 2f _k . Since the ratio of the Doppler frequency f _k to the Doppler frequency 2f _k is equal to 1/2, the first target channel adaptation The training data length (immediate domain sampling range) and prediction interval of the model are divided into 2 times the training data length and prediction interval of the target model, so that the wireless communication device can perform 1/2 interpolation processing on the first channel response, and we get As shown in (b) of FIG. 8 in the second channel response, it can be seen that the number of matrices in the second channel response is twice the number of matrices in the first channel response. Then, the wireless communication device may use the channel response data (matrix) in the second channel response that is within the training data length of the target model as the target model inference input, and calculate the channel response data (matrix) of the first target channel at a time corresponding to the prediction interval of the target model. Response prediction.

Also for example, when the selected model (i.e., the target model) is associated with the channel Doppler frequency f _k (i.e., the auxiliary parameter associated with the target model), and the Doppler of the channel that needs to be predicted (i.e., the first target channel) The frequency is f _k /2; that is, the matching degree between the auxiliary parameter associated with the target model and the first parameter is less than or equal to the first matching threshold (for example, the first matching threshold is 100%). It is assumed that the model associated with the channel Doppler frequency f _k has been trained.

Figure 9 shows a schematic diagram of the data preprocessing process of Doppler frequency f _k /2 through a model related to Doppler frequency f _k ; due to the ratio of Doppler frequency f _k to Doppler frequency f _k /2 is equal to 2, so the training data length (i.e. domain sampling range) and prediction interval of the model adapted to the first target channel are divided into half of the training data length and prediction interval of the target model, so that (a) in Figure 9 or As shown in (b) in Figure 9, the wireless communication device can perform 50% selection/sampling processing on the channel response in the first channel response to obtain the second channel response, and the number of matrices in the second channel response is Half the number of matrices in a channel. Then, the wireless communication device can use the channel response data (matrix) in the second channel response that is within the training data length of the target model as the inference input of the target model, and perform the calculation of the first target channel at the time corresponding to the prediction interval of the target model. Channel response prediction.

Optionally, in this embodiment of the present application, as shown in FIG. 9 , the selection process of the channel response is associated with the selection of the channel response of the prediction interval.

Optionally, in the embodiment of the present application, as shown in Figure 10, it is a schematic flow chart of a wireless communication device performing model training, data preprocessing and channel prediction based on an AI functional architecture. It can be seen from Figure 10 that the AI functional architecture at least Including: data collection module, Doppler or related characteristic prediction module, data preprocessing module, model training module, model selection module and model inference module. Among them, the wireless communication device can collect the channel response of the channel (hereinafter referred to as channel A) through the data collection module, and then:

In one case, the wireless communication device can directly train a model associated with the corresponding auxiliary parameters through the model training module based on the channel response collected by the data collection module; then the wireless communication device can deploy or deploy the trained model through the model training module. Update to the model inference module, that is, add it to the model library, so that subsequent predictions of related channels can be made based on the trained model.

In another case, the wireless communication device can estimate the Doppler frequency-related parameters of channel A based on the channel response collected by the data collection module, such as estimating the time standard deviation σ _UE of the autocorrelation function of channel A; and then wirelessly communicate The device can determine a model associated with the same or corresponding time standard deviation based on the time standard deviation σ _UE through the model selection module. For example, the kth model; thus, the wireless communication device can train the selected model through the model training module, or predict channel A based on the selected model through the model inference module.

In another case, when the parameters related to the Doppler frequency of channel A match the auxiliary parameters associated with the selected model, but the matching degree between the two is less than or equal to the first matching threshold, the wireless communication device may first be based on The data preprocessing module preprocesses the channel response collected by the data collection module, and then uses the model inference module to predict channel A based on the preprocessed channel response and the selected model, and outputs the predicted channel response.

Further, after the wireless communication device predicts channel A based on the model inference module, the wireless communication device can feed back the performance of the k-th model to the model selection module through the model inference module.

It should be noted that in the embodiments of the present application, the above embodiments all take the training, inference and update of the target model of wireless communication equipment as an example. In actual implementation, the training, inference and update of the target model can be performed by different methods respectively. The execution of the communication device can be specifically determined according to actual usage requirements, and is not limited in the embodiments of this application.

For the channel prediction method provided by the embodiments of the present application, the execution subject may be a channel prediction device. In the embodiment of the present application, the channel prediction method performed by the channel prediction apparatus is used as an example to illustrate the channel prediction apparatus provided by the embodiment of the present application.

An embodiment of the present application provides a channel prediction device 110. Figure 11 shows a schematic structural diagram of the channel prediction device 110 provided by an embodiment of the present application. As shown in Figure 11, the channel prediction device 110 may include: a determination module 111 and Prediction module 112. Determining module 111 for determining based on the auxiliary parameter set and the first parameter of the first target channel, to determine a target model from at least one model; the prediction module 112 is configured to predict the first target channel based on the target model determined by the determination module 111 . Each model is associated with an auxiliary parameter in the auxiliary parameter set, each auxiliary parameter is mapped to the Doppler frequency of a channel, and the first parameter is mapped to the Doppler frequency of the first target channel.

In a possible implementation, each of the above auxiliary parameters includes at least one of the following: the Doppler frequency of the above one channel, the first autocorrelation function of the above one channel, and the time standard deviation of the first autocorrelation function.

In a possible implementation, the above-mentioned first parameter includes at least one of the following: the Doppler frequency of the first target channel, the second autocorrelation function of the first target channel, and the time standard deviation of the second autocorrelation function.

In a possible implementation, the auxiliary parameters associated with the above target model match the first parameters.

In a possible implementation manner, the above-mentioned second auxiliary parameter is the auxiliary parameter with the greatest matching degree between the auxiliary parameter set and the first parameter.

In a possible implementation, the above prediction module 112 is specifically used to predict the first target channel based on the target model and the first channel response; wherein the first channel response is the first signal received or estimated by the receiving end communication device. Channel response of the target channel.

In a possible implementation, the above-mentioned prediction module 112 is specifically configured to perform a prediction on the first channel with the target when the matching degree between the auxiliary parameters associated with the target model and the first parameter is less than or equal to the first matching threshold. The preprocessing corresponding to the comparison result obtains the second channel response. The target comparison result is the comparison result between the auxiliary parameters associated with the target model and the first parameter; the wireless communication device predicts the first target channel based on the target model and the second channel response. .

In a possible implementation manner, the above target comparison result is a ratio between the auxiliary parameter associated with the target model and the first parameter.

In a possible implementation, the above device may also include a training module. The trained module is used to train the target model based on the auxiliary parameters associated with the target model before the determining module 111 determines the target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel.

In a possible implementation, the auxiliary parameters associated with the target model are mapped to the Doppler frequency of the second target channel; the training module is specifically used to map the auxiliary parameters associated with the target model and the channel response of the second target channel based on the auxiliary parameters associated with the target model and the channel response of the second target channel. Train the target model.

In a possible implementation, the above device may also include an evaluation module and an update module; the evaluation module is used to train the target model after the training module based on the auxiliary parameters associated with the target model and the channel response of the second target channel, Perform performance evaluation on the target model; the updated module is used to update or fine-tune the model parameters of the target model when the performance evaluation results of the target model do not meet the performance requirements;

In a possible implementation, the above device further includes an estimation module; an estimation module configured to determine the target model based on the target model from at least one model before the determination module 111 determines the target model based on the auxiliary parameter set and the first parameter of the first target channel. information, perform parameter estimation on the first target channel, and obtain the first parameter;

In a possible implementation, the Doppler frequency of the channel is determined based on the arrival angle of the reference signal on the channel.

In a possible implementation, the auxiliary parameters associated with each of the above models are agreed upon by high-level configuration or protocols.

In the channel prediction device 110 provided in the embodiment of the present application, when it is necessary to predict the first target channel, the channel prediction device 110 can determine the target based on the first parameter mapped to the Doppler frequency of the first target channel. model, thereby ensuring that the auxiliary parameters associated with the target model are adapted to the Doppler frequency of the first target channel. For example, the auxiliary parameters used in training the target model (mapping with the Doppler frequency of a channel) are consistent with the first parameters. Matching, that is, the first target channel can be predicted using a target model whose Doppler frequency characteristics match the Doppler frequency characteristics of the first target channel to be predicted. Therefore, compared with the generalized learning model used in related technologies, As for the channel prediction scheme, the channel prediction method provided by the embodiments of this application can improve the accuracy of channel prediction.

The channel prediction device in the embodiment of the present application may be an electronic device, such as an electronic device with an operating system, or may be a component in the electronic device, such as an integrated circuit or chip. The electronic device may be a terminal or other devices other than the terminal. For example, terminals may include but are not limited to the types of terminals 11 listed above, and other devices may be servers, network attached storage (Network Attached Storage, NAS), etc., which are not specifically limited in the embodiment of this application.

The channel prediction device provided by the embodiments of the present application can implement each process implemented by the method embodiments in Figures 1 to 10 and achieve the same technical effect. To avoid duplication, details will not be described here.

The embodiment of the present application also provides a communication device 5000, as shown in Figure 12, including a processor 5001 and a memory 5002. The memory 5002 stores programs or instructions that can be run on the processor 5001. For example, the communication When the device 5000 is a UE, the program or instruction is processed by the processor 5001 During execution, each step of the above UE side method embodiment is implemented and the same technical effect can be achieved. To avoid duplication, it will not be repeated here. Or when the communication device 5000 is a network side device, the program or instruction is executed by the processor 5001 Each step of the above method embodiment of the first network side device or the second network side device is implemented, and the same technical effect can be achieved. To avoid duplication, the details will not be described here.

An embodiment of the present application also provides a wireless communication device, including a processor and a communication interface, wherein the processor is configured to determine a target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel; and Based on the target model, predict the first target channel; wherein each model is associated with an auxiliary parameter in the auxiliary parameter set, each auxiliary parameter is mapped to the Doppler frequency of a channel, and the first parameter is mapped to the first target The Doppler frequency of the channel is mapped, and the communication interface is used to obtain the first parameter. Each implementation process and implementation manner of the above method embodiment can be applied to this terminal embodiment, and can achieve the same technical effect.

Optionally, in this embodiment of the present application, the wireless communication device may be a UE or a network side device.

Optionally, in this embodiment of the present application, taking the wireless communication device as a UE as an example, FIG. 13 is one of the schematic diagrams of the hardware structure of a wireless communication device that implements the embodiment of the present application.

As shown in Figure 13, UE7000 includes but is not limited to: radio frequency unit 7001, network module 7002, audio output unit 7003, input unit 7004, sensor 7005, display unit 7006, user input unit 7007, interface unit 7008, memory 7009 and processor At least some parts of 7010 etc.

Those skilled in the art can understand that the UE7000 can also include a power supply (such as a battery) that supplies power to various components. The power supply can be logically connected to the processor 7010 through the power management system, thereby achieving management of charging, discharging, and power consumption management through the power management system. and other functions. The UE structure shown in Figure 13 does not constitute a limitation on the UE. The UE may include more or less components than shown in the figure, or combine certain components, or arrange different components, which will not be described again here.

It should be understood that in the embodiment of the present application, the input unit 7004 may include a graphics processing unit (Graphics Processing Unit, GPU) 7041 and a microphone 7042. The graphics processor 7041 is responsible for the image capture device (GPU) in the video capture mode or the image capture mode. Process the image data of still pictures or videos obtained by cameras (such as cameras). The display unit 7006 may include a display panel 7061, which may be configured in the form of a liquid crystal display, an organic light emitting diode, or the like. The user input unit 7007 includes a touch panel 7071 and at least one of other input devices 7072 . Touch panel 7071, also called touch screen. The touch panel 7071 may include two parts: a touch detection device and a touch controller. Other input devices 7072 may include but are not limited to physical keyboards, function keys (such as volume control keys, switch keys, etc.), trackballs, mice, and joysticks, which will not be described again here.

In this embodiment of the present application, after receiving downlink data from the network side device, the radio frequency unit 7001 can transmit it to the processor 7010 for processing; in addition, the radio frequency unit 7001 can send uplink data to the network side device. Generally, the radio frequency unit 7001 includes, but is not limited to, an antenna, amplifier, transceiver, coupler, low noise amplifier, duplexer, etc.

Memory 7009 may be used to store software programs or instructions as well as various data. The memory 7009 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 for at least one function (such as a sound playback function, Image playback function, etc.) etc. Additionally, memory 7009 may include volatile memory or nonvolatile memory, or memory 7009 may include both volatile and nonvolatile memory. Among them, 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), electrically removable memory. Erase programmable read-only memory (Electrically EPROM, EEPROM) or flash 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 link dynamic random access memory (Synch link DRAM) , SLDRAM) and direct memory bus random access memory (Direct Rambus RAM, DRRAM). Memory 709 in embodiments of the present application includes, but is not limited to, these and any other suitable types of memory.

The processor 7010 may include one or more processing units; optionally, the processor 7010 integrates an application processor and a modem processor, where the application processor mainly handles operations related to the operating system, user interface, application programs, etc., Modem processors mainly process wireless communication signals, such as baseband processors. It can be understood that the above modem processor may not be integrated into the processor 7010.

The processor 7010 is configured to determine a target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel; and is configured to predict the first target channel based on the target model. Each model is associated with an auxiliary parameter in the auxiliary parameter set, each auxiliary parameter is mapped to the Doppler frequency of a channel, and the first parameter is mapped to the Doppler frequency of the first target channel.

In a possible implementation, each of the above auxiliary parameters includes at least one of the following: the Doppler frequency of a channel, the first autocorrelation function of a channel, and the time standard deviation of the first autocorrelation function.

In a possible implementation, the above-mentioned first parameter includes at least one of the following: the Doppler frequency of the first target channel, the second Doppler frequency of the first target channel The time standard deviation of the autocorrelation function and the second autocorrelation function.

In a possible implementation, the above-mentioned processor 7010 is specifically configured to predict the first target channel based on the target model and the first channel response; wherein the first channel response is the first signal received or estimated by the receiving end communication device. Channel response of the target channel.

In a possible implementation, the above-mentioned processor 7010 is specifically configured to perform the matching on the first channel when the matching degree between the auxiliary parameter associated with the target model and the first parameter is less than or equal to the first matching threshold. The preprocessing corresponding to the target comparison result obtains the second channel response, and the target comparison result is the comparison result between the auxiliary parameters associated with the target model and the first parameter; the wireless communication device performs on the first target channel based on the target model and the second channel response. predict.

In a possible implementation, the above-mentioned processor 7010 is further configured to determine the target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel, based on the auxiliary parameters associated with the target model. Target model is trained.

In a possible implementation, the auxiliary parameters associated with the target model are mapped to the Doppler frequency of the second target channel; the above processor 7010 is specifically configured to generate a channel based on the auxiliary parameters associated with the target model and the second target channel. In response, the target model is trained.

In a possible implementation, the above-mentioned processor 7010 is also used to perform performance evaluation on the target model after training the target model based on the auxiliary parameters associated with the target model and the channel response of the second target channel; and If the performance evaluation results of the model do not meet the performance requirements, update or fine-tune the model parameters of the target model;

In a possible implementation, the above-mentioned processor 7010 is further configured to, before determining the target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel, based on the target information, determine the first target channel Perform parameter estimation to obtain the first parameter;

In the wireless communication device provided by the embodiment of the present application, when it is required to predict the first target channel, since the channel prediction device can determine the target model based on the first parameter mapped to the Doppler frequency of the first target channel, This ensures that the auxiliary parameters associated with the target model are adapted to the Doppler frequency of the first target channel. For example, the auxiliary parameters used in training the target model (mapping with the Doppler frequency of a channel) match the first parameters, That is, a target model whose Doppler frequency characteristics match the Doppler frequency characteristics of the first target channel to be predicted can be used to predict the first target channel. Therefore, compared with the use of a generalized learning model in related technologies to predict the channel For prediction solutions, the channel prediction method provided by the embodiments of this application can improve the accuracy of channel prediction.

Optionally, in this embodiment of the present application, the above-mentioned wireless communication device is a network-side device as an example. Figure 14 is the second schematic diagram of the hardware structure of a wireless communication device that implements the embodiment of the present application.

As shown in Figure 14, the network side device 8000 includes: an antenna 8001, a radio frequency device 8002, a baseband device 8003, a processor 8004 and a memory 8005. Antenna 8001 is connected to radio frequency device 8002. In the uplink direction, the radio frequency device 8002 receives information through the antenna 8001 and sends the received information to the baseband device 8003 for processing. In the downlink direction, the baseband device 8003 processes the information to be sent and sends it to the radio frequency device 8002. The radio frequency device 8002 processes the received information and sends it out through the antenna 8001.

The method performed by the wireless communication device in the above embodiment can be implemented in the baseband device 8003, which includes a baseband processor.

The baseband device 8003 may include, for example, at least one baseband board on which multiple chips are disposed, as shown in FIG. The program performs the operations of the wireless communication device shown in the above method embodiment.

The network side device may also include a network interface 8006, which is, for example, a common public radio interface (CPRI).

Specifically, the network side device 8000 in this embodiment of the present invention also includes: instructions or programs stored in the memory 8005 and executable on the processor 8004. The processor 8004 calls the instructions or programs in the memory 8005 to execute the above-mentioned modules. method and achieve the same technical effect. To avoid repetition, we will not repeat it here.

Embodiments of the present application also provide a readable storage medium. Programs or instructions are stored on the readable storage medium. When the program or instructions are executed by a processor, each process of the above channel prediction method embodiment is implemented, and the same can be achieved. The technical effects will not be repeated here to avoid repetition.

Wherein, the processor is the processor in the wireless communication device described in the above embodiment. The readable storage medium includes computer readable storage media, such as computer read-only memory ROM, random access memory RAM, magnetic disk or optical disk, etc.

An 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. 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. To avoid duplication, it will not be described again here.

It should be understood that the chips mentioned in the embodiments of this application may also be called system-on-chip, system-on-a-chip, system-on-chip or system-on-chip, etc.

Embodiments of the present application further provide a computer program/program product. The computer program/program product is stored in a storage medium. The computer program/program product is executed by at least one processor to implement the above channel prediction method embodiment. Each process can achieve the same technical effect. To avoid repetition, we will not go into details here.

It should be noted that, in this document, the terms "comprising", "comprising" or any other variations thereof are intended to cover a non-exclusive inclusion, such that a process, method, article or device that includes a series of elements not only includes those elements, It also includes other elements not expressly listed or inherent in the process, method, article or apparatus. Without further limitation, an element defined by the statement "comprises a..." does not exclude the presence of additional identical elements in a process, method, article, or device that includes that element. In addition, it should be pointed out that 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, but may also include performing functions in a substantially simultaneous manner or in reverse order according to the functions involved. Functions may be performed, for example, the methods described may be performed in an order different from that described, and various steps may be added, omitted, or combined. Additionally, features described with reference to certain examples may be combined in other examples.

Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus the necessary general hardware platform. Of course, it can also be implemented by hardware, but in many cases the former is better. implementation. Based on this understanding, the technical solution of the present application can be embodied in the form of a computer software product that is essentially or contributes to the existing technology. The computer software product is stored in a storage medium (such as ROM/RAM, disk , CD), including several instructions to cause a terminal (which can be a mobile phone, computer, server, air conditioner, or network device, etc.) to execute the methods described in various embodiments of this application.

The embodiments of the present application have been described above in conjunction with the accompanying drawings. However, the present application is not limited to the above-mentioned specific implementations. The above-mentioned specific implementations are only illustrative and not restrictive. Those of ordinary skill in the art will Inspired by this application, many forms can be made without departing from the purpose of this application and the scope protected by the claims, all of which fall within the protection of this application.

Claims

A channel prediction method, including:

The wireless communication device determines the target model from the at least one model based on the auxiliary parameter set and the first parameter of the first target channel;

The wireless communication device predicts the first target channel based on the target model;

Wherein, each model is associated with an auxiliary parameter in the auxiliary parameter set, each auxiliary parameter is mapped to the Doppler frequency of a channel, and the first parameter is mapped to the Doppler frequency of the first target channel. Phase mapping.
The method of claim 1, wherein each auxiliary parameter includes at least one of the following:

The Doppler frequency of the one channel, the first autocorrelation function of the one channel, and the time standard deviation of the first autocorrelation function.
The method according to claim 1 or 2, wherein the first parameter includes at least one of the following:

The Doppler frequency of the first target channel, the second autocorrelation function of the first target channel, and the time standard deviation of the second autocorrelation function.
The method of claim 1, wherein the auxiliary parameters associated with the target model match the first parameters.
The method of claim 4, wherein the second auxiliary parameter is the auxiliary parameter in the auxiliary parameter set that has the greatest matching degree with the first parameter.
The method of claim 1, wherein the wireless communication device predicts the first target channel based on the target model, including:

The wireless communication device predicts the first target channel based on the target model and the first channel response;

Wherein, the first channel response is a channel response of the first target channel received or estimated by the receiving end communication device.
The method of claim 6, wherein the wireless communication device predicts the first target channel based on the target model and the first channel response, including:

When the matching degree between the auxiliary parameter associated with the target model and the first parameter is less than or equal to a first matching threshold, the wireless communication device performs a predetermined process corresponding to the target comparison result on the first channel. Process to obtain a second channel response, and the target comparison result is a comparison result between the auxiliary parameter associated with the target model and the first parameter;

The wireless communication device predicts the first target channel based on the target model and the second channel response.
The method according to claim 7, wherein the target comparison result is a ratio between an auxiliary parameter associated with the target model and the first parameter.
The method of claim 1, wherein before the wireless communication device determines the target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel, the method further includes:

The wireless communication device trains the target model based on the auxiliary parameters associated with the target model.
The method of claim 9, wherein the auxiliary parameters associated with the target model are mapped to the Doppler frequency of the second target channel;

The wireless communication device trains the target model based on the auxiliary parameters associated with the target model, including:

The wireless communication device trains the target model based on the auxiliary parameters associated with the target model and the channel response of the second target channel.
The method according to claim 9 or 10, wherein after the wireless communication device trains the target model based on the auxiliary parameters associated with the target model and the channel response of the second target channel, the method Also includes:

The wireless communication device performs performance evaluation on the target model;

The wireless communication device updates or fine-tunes the model parameters of the target model when the performance evaluation result of the target model does not meet the performance requirements;

Wherein, the model parameters of the target model include at least one of the following: a time domain sampling range of the target model, a prediction interval of the target model, and a sampling interval of the target model.
The method of claim 1, wherein before the wireless communication device determines the target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel, the method further includes:

The wireless communication device performs parameter estimation on the first target channel based on target information to obtain the first parameter;

Wherein, the target information includes at least one of the following: a reference signal on the first target channel, a third channel response of the first target channel estimated or received by the receiving end communication device.
The method of claim 2, wherein the Doppler frequency of the channel is determined based on the angle of arrival of the reference signal on the channel.
The method according to claim 1, wherein the auxiliary parameters associated with each model are agreed by high-level configuration or protocol.
A channel prediction device, the device includes: a determination module and a prediction module;

The determining module is configured to determine a target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel;

The prediction module is configured to predict the first target channel based on the target model determined by the determination module;

Wherein, each model is associated with an auxiliary parameter in the auxiliary parameter set, each auxiliary parameter is mapped to the Doppler frequency of a channel, and the first parameter is mapped to the Doppler frequency of the first target channel. Phase mapping.
The device of claim 15, wherein each auxiliary parameter includes at least one of the following:

The Doppler frequency of the one channel, the first autocorrelation function of the one channel, and the time standard deviation of the first autocorrelation function.
The device according to claim 15 or 16, wherein the first parameter includes at least one of the following:

The Doppler frequency of the first target channel, the second autocorrelation function of the first target channel, and the time standard deviation of the second autocorrelation function.
The apparatus of claim 15, wherein an auxiliary parameter associated with the target model matches the first parameter.
The device according to claim 18, wherein the second auxiliary parameter is the auxiliary parameter with the greatest matching degree between the auxiliary parameter set and the first parameter.
The device of claim 15, wherein:

The prediction module is specifically configured to predict the first target channel based on the target model and the first channel response;

Wherein, the first channel response is a channel response of the first target channel received or estimated by the receiving end communication device.
The device according to claim 20, wherein the prediction module is specifically configured to: when the matching degree between the auxiliary parameters associated with the target model and the first parameter is less than or equal to a first matching threshold, Perform preprocessing corresponding to the target comparison result on the first channel to obtain a second channel response, where the target comparison result is the comparison result between the auxiliary parameters associated with the target model and the first parameter;

The wireless communication device predicts the first target channel based on the target model and the second channel response.
The device according to claim 21, wherein the target comparison result is a ratio between an auxiliary parameter associated with the target model and the first parameter.
The device of claim 15, wherein the device further includes a training module;

The trained module is configured to, before the determining module determines a target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel, based on the auxiliary parameters associated with the target model, evaluate the target model. Conduct training.
The apparatus of claim 23, wherein the auxiliary parameters associated with the target model are mapped to the Doppler frequency of the second target channel;

The training module is specifically configured to train the target model based on the auxiliary parameters associated with the target model and the channel response of the second target channel.
The device according to claim 23 or 24, wherein the device further includes an evaluation module and an update module;

The evaluation module is configured to perform performance evaluation on the target model after the training module trains the target model based on the auxiliary parameters associated with the target model and the channel response of the second target channel;

The updated module is used to update or fine-tune the model parameters of the target model when the performance evaluation results of the target model do not meet the performance requirements;

Wherein, the model parameters of the target model include at least one of the following: a time domain sampling range of the target model, a prediction interval of the target model, and a sampling interval of the target model.
The device of claim 15, wherein the device further comprises an estimation module;

The estimating module is configured to perform parameter estimation on the first target channel based on target information before the determining module determines a target model from at least one model based on the auxiliary parameter set and the first parameter of the first target channel. , obtain the first parameter;

Wherein, the target information includes at least one of the following: a reference signal on the first target channel, a third channel response of the first target channel estimated or received by the receiving end communication device.
The apparatus of claim 16, wherein the Doppler frequency of the channel is determined based on the angle of arrival of the reference signal on the channel.
The apparatus according to claim 15, wherein the auxiliary parameters associated with each model are agreed upon by a high-level configuration or protocol.
A wireless communication device, including a processor and a memory, the memory stores a program or instructions that can be run on the processor, and when the program or instructions are executed by the processor, any one of claims 1 to 14 is implemented. The steps of the channel prediction method described in one item.
A readable storage medium on which a program or instructions are stored, and when the program or instructions are executed by a processor, the steps of the channel prediction method according to any one of claims 1 to 14 are implemented.
A computer software product executed by at least one processor to implement the channel prediction method according to any one of claims 1 to 14.
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 program programs or instructions to implement the method as described in any one of claims 1 to 14 Channel prediction methods.