Channel prediction method and device

By obtaining the position information of the points to be measured and reference points and historical channel status information, and using the direction of motion for channel prediction, the problems of low channel prediction accuracy and high data acquisition difficulty in the prior art are solved, and the problem of low channel prediction accuracy and high data acquisition difficulty is realized in different scenarios high accuracy channel prediction.

CN118945013APending Publication Date: 2024-11-12HUAWEI TECH CO LTD
View PDF 0 Cites 2 Cited by

Patent Information

Application Number
CN202310521588.4
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2023-05-09
Publication Date
2024-11-12

AI Technical Summary

Technical Problem

In the prior art, the accuracy of channel prediction is not high, and the training data acquisition of the channel prediction model is difficult, especially when the user moves quickly or the prediction interval is large, the channel prediction performance has significant attenuation.

Method used

By obtaining the position information of the points to be measured and the reference points and the historical channel state information of the reference points, using the motion direction information to perform channel prediction, converting the channel state information into an angle delay channel matrix, and training of the neural network model is carried out based on this and reasoning reduces the difficulty of data acquisition and improves the accuracy of channel prediction.

Benefits of technology

It improves the accuracy of channel prediction and the feasibility of model training, and can maintain high prediction accuracy at different sampling densities and sequence sampling intervals, avoiding the damage to data acquisition difficulty and channel state parameter correlation.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118945013A_ABST
    Figure CN118945013A_ABST
Patent Text Reader

Abstract

The invention discloses a channel prediction method and device, relates to the technical field of communication, and is used for solving the problems that the accuracy of channel prediction is not high, and training data acquisition of a channel prediction model is difficult to realize. The method comprises the following steps: acquiring position information of a point to be measured, position information of a reference point and historical channel state information corresponding to the reference point; the channel state information of the point to be measured is obtained according to the historical channel state information of the reference point and the motion direction information between the point to be measured and the reference point, and the historical channel state information of the reference point comprises an angle time delay domain channel matrix corresponding to the reference point.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present application relates to the field of communication technology, and in particular to a channel prediction method and device. Background Art

[0002] In the field of wireless communications, channel state information (CSI) can be used to describe the channel characteristics of a wireless link, characterizing the joint characteristics of path loss, scattering, diffraction, fading or shadowing during the propagation of signals from transmission to reception. Therefore, the deployment of physical layer parameters or network resources in wireless communication networks is related to CSI, and accurate prediction of CSI plays an important role in improving network resource utilization.

[0003] At present, CSI prediction can be achieved through a channel prediction solution based on a long short-term memory (LSTM) network. Specifically, the CSI matrix collected at equal time intervals is used as the training data of the LSTM network for network training. The CSI matrix sampled at equal time intervals in a past sequence can be used as the input parameter of the LSTM network, and its output can be the instantaneous channel state information at the current moment.

[0004] However, the above scheme can only mine the data correlation between discrete time channel sequences. In the actual channel transmission process, the propagation of electromagnetic waves in space has special intrinsic data characteristics, which cannot be reflected in the above scheme. Therefore, the accuracy of channel prediction is not high. In addition, when the prediction interval is large or the user moves at a fast speed, the performance of channel prediction is greatly attenuated. For training data, the above scheme requires the collection of channel sampling data with equal time intervals to train the LSTM network. The sampling density and sampling spacing requirements are high, which is difficult to achieve in practical applications. Summary of the invention

[0005] The present application provides a channel prediction method and device, which solves the problems in the prior art that the accuracy of channel prediction is low and the training data collection of the channel prediction model is difficult to implement.

[0006] In order to achieve the above objectives, this application adopts the following technical solutions:

[0007] In a first aspect, a channel prediction method is provided, the method comprising: acquiring position information of a point to be measured, as well as position information of a reference point and historical channel state information corresponding to the reference point, wherein the historical channel state information of the reference point comprises an angle delay domain channel matrix corresponding to the reference point; predicting the channel state information of the point to be measured based on the historical channel state information of the reference point and the motion direction information between the point to be measured and the reference point, wherein the motion direction information is used to represent a predicted moving path of a user at the point to be measured to the reference point.

[0008] In the above implementation, the input data of the channel prediction is the position information of the point to be measured, the position information of the reference point and the historical channel state information corresponding to the reference point, and the sampling density and sampling interval of the input data are not limited, which reduces the difficulty of data collection and improves the feasibility of model training. In addition, by converting the complex matrix of the channel state information into an angle delay domain channel matrix and performing channel prediction based on the moving path of the moving direction of the point to be measured, the internal correlation of the data parameters affecting the channel state will not be destroyed, thereby improving the accuracy of the channel prediction, so that under different implementation scenario settings (such as different sampling densities, different sequence sampling intervals, etc.), there are obvious advantages in prediction accuracy.

[0009] In one embodiment, the channel state information of the point to be tested is predicted based on the historical channel state information of the reference point and the motion direction information between the point to be tested and the reference point, specifically including: inputting the historical channel state information of the reference point and the motion direction information between the point to be tested and the reference point into a first channel prediction model to obtain the channel state information of the point to be tested.

[0010] In one embodiment, the channel state information of the point to be tested is predicted based on the historical channel state information of the reference point and the movement direction information between the point to be tested and the reference point, specifically including: inputting the historical channel state information of the reference point and the movement direction information between the point to be tested and the reference point into a second channel prediction model to obtain gradient information, wherein the gradient information is used to indicate the change information of the channel state in the movement direction; integrating the gradient information according to the moving path of the point to be tested to obtain the channel state information of the point to be tested.

[0011] In the above implementation, the channel state information is converted into an angle delay domain channel matrix, and the gradient information of the spatial position is calculated based on the angle delay domain channel matrix, thereby performing model training and reasoning of the neural network to achieve channel prediction. Compared with the training data of the neural network in the prior art, the difficulty of data collection is reduced and the feasibility of model training is improved. In addition, by predicting the gradient information obtained by deriving the angle delay domain channel matrix, the internal correlation of the data parameters that affect the channel state will not be destroyed, thereby improving the accuracy of channel prediction.

[0012] In one embodiment, the method further includes: obtaining a velocity characteristic vector of the user based on the user's historical location information and / or historical channel state information; performing a Doppler elimination operation based on the channel state information of the point to be measured and the velocity characteristic vector of the user to obtain the mobile channel information of the point to be measured.

[0013] In the above implementation, the communication device can also estimate the user's speed characteristic vector based on the user's historical location information and / or historical channel state information, so that the mobile channel information can be further estimated based on the static channel state information and the user's speed characteristic vector, thereby improving the accuracy of channel prediction.

[0014] In one implementation, the gradient information includes first information D1 and second information D2, wherein the first information D1 satisfies:

[0015]

[0016] The second information D2 satisfies:

[0017] in, Used to represent the real part of the angle delay domain channel matrix, It is used to represent the imaginary part of the angle delay domain channel matrix, θ is used to represent the corresponding arrival angle in the angle delay domain channel matrix, τ is used to represent the corresponding delay in the angle delay domain channel matrix, d p It is used to indicate the total length of the channel transmission path, and m is used to indicate the location information.

[0018] In one embodiment, the method further includes: training the first channel prediction model or the second channel prediction model according to historical channel state information of multiple reference points and motion direction information of multiple motion paths corresponding to each reference point.

[0019] In one embodiment, obtaining the location information of the user's test point, the location information of the reference point, and the historical channel status information corresponding to the reference point specifically includes: predicting the location information of the user's test point based on the user's historical location information and / or the historical channel status information corresponding to the historical location information; determining the location information of the reference point and the historical channel status information of the reference point from a historical channel data set based on the location information of the test point, wherein the reference point is a detection point in the historical channel data set with the shortest distance from the test point, and the historical channel data set includes the location information of multiple detection points and the corresponding relationship between the historical channel status information.

[0020] In one embodiment, the location information of the point to be measured is obtained according to the historical location information of the user and / or the historical channel status information corresponding to the historical location information, specifically including: performing user location prediction according to the historical location information of the user to obtain the location feature vector of the point to be measured; or, performing user location prediction according to the historical channel status information corresponding to the historical location information of the user to obtain the location feature vector of the point to be measured; or, performing user location prediction according to the historical location information of adjacent users to obtain the location feature vector of the point to be measured; or, performing user location prediction according to the historical channel status information corresponding to the historical location information of adjacent users to obtain the location feature vector of the point to be measured.

[0021] In a second aspect, a channel prediction device is provided, which includes: a channel positioning module, used to obtain position information of a point to be measured, as well as position information of a reference point and historical channel state information corresponding to the reference point, wherein the historical channel state information of the reference point includes an angle delay domain channel matrix corresponding to the reference point; a channel prediction module, used to predict the channel state information of the point to be measured based on the historical channel state information of the reference point and movement direction information between the point to be measured and the reference point, wherein the movement direction information is used to represent the predicted moving path of a user at the point to be measured to the reference point.

[0022] In one implementation, the channel prediction module is specifically used to input the historical channel state information of the reference point and the motion direction information between the point to be measured and the reference point into a first channel prediction model to obtain the channel state information of the point to be measured.

[0023] In one embodiment, the channel prediction module is specifically used to input the historical channel state information of the reference point and the movement direction information between the point to be tested and the reference point into a second channel prediction model to obtain gradient information, wherein the gradient information is used to indicate the change information of the channel state in the movement direction; and the channel state information of the point to be tested is obtained according to the moving path of the point to be tested and the gradient information.

[0024] In one embodiment, the channel prediction module is also used to obtain the user's velocity characteristic vector based on the user's historical location information and / or historical channel state information; perform Doppler elimination operation based on the channel state information of the point to be measured and the user's velocity characteristic vector to obtain the mobile channel information of the point to be measured.

[0025] In one implementation, the gradient information includes first information D1 and second information D2, wherein the first information D1 satisfies:

[0026]

[0027] The second information D2 satisfies:

[0028] in, Used to represent the real part of the angle delay domain channel matrix, It is used to represent the imaginary part of the angle delay domain channel matrix, θ is used to represent the corresponding arrival angle in the angle delay domain channel matrix, τ is used to represent the corresponding delay in the angle delay domain channel matrix, d p It is used to indicate the total length of the channel transmission path, and m is used to indicate the location information.

[0029] In one embodiment, the channel prediction module is further used to train the first channel prediction model or the second channel prediction model according to historical channel state information of multiple reference points and motion direction information of multiple motion paths corresponding to each reference point.

[0030] In one embodiment, the channel positioning module is specifically used to: predict the location information of the user's test point based on the user's historical location information, and / or the historical channel status information corresponding to the historical location information; determine the location information of the reference point and the historical channel status information of the reference point from the historical channel data set based on the location information of the test point, wherein the reference point is the detection point in the historical channel data set with the shortest distance from the test point, and the historical channel data set includes the location information of multiple detection points and the corresponding relationship between the historical channel status information.

[0031] In one embodiment, the channel positioning module is specifically used to: predict the user position based on the historical position information of the user to obtain the position feature vector of the point to be measured; or, predict the user position based on the historical channel state information corresponding to the historical position information of the user to obtain the position feature vector of the point to be measured; or, predict the user position based on the historical position information of adjacent users to obtain the position feature vector of the point to be measured; or, predict the user position based on the historical channel state information corresponding to the historical position information of adjacent users to obtain the position feature vector of the point to be measured.

[0032] According to a third aspect, a device is provided, comprising: a processor, the processor being coupled to a memory, the memory being used to store programs or instructions, and when the programs or instructions are executed by the processor, the device executes a method as described in any one of the above-mentioned first aspects.

[0033] In a fourth aspect, a chip is provided, comprising: a processor, the processor being coupled to a memory, the memory being used to store programs or instructions, and when the programs or instructions are executed by the processor, the chip executes a method as described in any one of the above-mentioned first aspects.

[0034] In a fifth aspect, a computer-readable storage medium is provided, on which a computer program or instruction is stored, and when the computer program or instruction is executed, the computer executes the method as described in any one of the above-mentioned first aspects.

[0035] According to a sixth aspect, a computer program product is provided, wherein the computer program product includes a computer program code, and when the computer program code is executed on a computer, the computer implements the method as described in any one of the above-mentioned first aspects.

[0036] It can be understood that any communication device, computer-readable storage medium, computer program product or chip provided in the above-mentioned second to sixth aspects can be implemented by the corresponding method provided above. Therefore, the beneficial effects that can be achieved can refer to the beneficial effects in the corresponding method provided above, and will not be repeated here. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 A schematic diagram of a flow chart of a channel prediction method;

[0038] Figure 2 It is a flowchart of a channel prediction scheme of an LSTM network;

[0039] Figure 3 A system architecture diagram of a communication system provided in an embodiment of the present application;

[0040] Figure 4A system architecture diagram of a communication device provided in an embodiment of the present application;

[0041] Figure 5 A schematic diagram of a flow chart of a channel prediction method provided in an embodiment of the present application;

[0042] Figure 6 A schematic diagram of the structure of another communication device provided in an embodiment of the present application;

[0043] Figure 7 A schematic diagram of an implementation of a channel prediction model provided in an embodiment of the present application;

[0044] Figure 8 A schematic diagram of the structure of a distributed communication system for implementing channel prediction provided in an embodiment of the present application;

[0045] Fig. 9 A schematic diagram of the structure of another device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0046] Before introducing the embodiments of the present application, the related technologies involved in the embodiments of the present application are explained. It is understandable that these explanations are intended to make the embodiments of the present application easier to understand and should not be regarded as limiting the scope of protection required by the embodiments of the present application.

[0047] 1. Neural network: A nonlinear transformation with a large number of adjustable parameters that can be updated through training to complete various tasks. The training process usually calculates the loss function first and then uses the back gradient propagation algorithm to update the parameters.

[0048] 2. Convolutional neural network: Convolutional neural network is a feedforward neural network whose artificial neurons can respond to surrounding units within a part of the coverage area and performs well in large image processing.

[0049] 3. Model loss: The loss function is used to measure the degree of deviation between the prediction made by the data model and the true value or sample label, that is, the gap between the final output result of the model and the actual correct result (or the sample labeled result). The smaller the gap, the better the performance of the model. Conversely, the poorer the performance of the model. It is necessary to continue to iteratively optimize the parameters of the model to find suitable parameters to reduce the loss between the model output result and the actual correct result, so that the model has a similar effect to manual recognition. There are many ways to calculate loss, such as cross entropy loss or minimum mean square error loss.

[0050] 4. Channel: refers to the communication channel and the medium for signal transmission.

[0051] 5. Channel state information: refers to the channel properties of the communication link, which can be used to describe the attenuation factor of the signal on each transmission path, that is, the value of each element in the channel state matrix H, such as signal scattering, environmental attenuation or distance attenuation.

[0052] Exemplarily, Y=HX+N, where Y is the signal vector of the receiving end, X is the signal vector of the transmitting end, H represents the channel state matrix, and N represents Gaussian white noise.

[0053] In one embodiment, the network device can determine the physical layer parameters and schemes deployed in the wireless communication network based on the CSI. For example, the physical layer needs to adopt a low-order modulation scheme when the channel is poor and a high-order modulation scheme when the channel is reliable, so as to obtain better bit error performance and improve information transmission efficiency. In addition, CSI also has a great impact on resource allocation and interference coordination in wireless communication systems. However, due to the complex external environment such as scattering and diffraction in the communication link, the channel modeling usually contains many unknown parameters. If it is necessary to accurately estimate them, it will lead to unbearable signaling overhead. Therefore, efficient and accurate estimation of CSI in a wireless environment plays a vital role in improving network resource utilization.

[0054] Currently, CSI can be obtained through two methods: one is model-based channel estimation, in which the channel model is usually defined by a statistical function containing a certain number of unknown parameters. If the parameters used to represent the channel are insufficient, the accuracy of the CSI estimation is difficult to guarantee. On the contrary, when there are many channel parameters, a large number of pilots are required for estimation, which may lead to high signaling overhead and processing delay. Another CSI acquisition method is called a deterministic channel model, such as a ray tracing algorithm, which obtains CSI by calculating the physical properties of the propagation of electromagnetic waves in a specific real environment without any channel measurement. Due to the analysis framework based on electromagnetic theory, this method takes into account the field characteristics and propagation path of the antenna, and can achieve good accuracy, but this type of method requires a complete understanding of the real physical propagation environment, which is usually difficult to obtain in actual systems.

[0055] 6. Static channel database: It is a channel prediction scheme in wireless communication systems. This scheme takes into account that the user's moving distance scale can be considered very small compared to the service area scale of the base station during the prediction process, and the scatterers (such as buildings) in the environment that play a major role in electromagnetic wave reflection are static, so it is considered that the relevant large-scale parameters in the channel prediction over a period of time are static, including path loss, arrival angle, number of paths, and delay.

[0056] In a communication implementation scenario, the base station usually serves a fixed area, and the number and position of scatterers in the area are usually fixed. The wireless channel is mainly determined by the scatterers distributed therein and the corresponding multipath propagation behaviors of electromagnetic waves such as specular reflection, diffraction, diffusion, and blocking. Although the channel characteristics will inevitably change due to the small-scale perturbations of the transmission medium, considering that the scatterers in the area do not change, the long-term multipath component of the channel can be regarded as quasi-static. Therefore, using a set of static CSI samples randomly measured in a fixed area, the CSI at any position in the same fixed area can be predicted. In addition, since the communication process is always carried out in the service area of ​​the base station, a large number of channel measurements at corresponding positions are generated, and these data can be potentially used, resulting in a significant reduction in signal overhead and measurement costs. Therefore, channel prediction is proposed as a way to significantly save pilot overhead. Its main idea is to use the channel measurement data over a period of time in the past and mine the implicit spatiotemporal correlations therein to predict the channel at the current moment.

[0057] For example, first, the channel prediction device can collect the measured CSI data and extract the multipath components (MPC) of the channel parameters through the Multiple Signal Classification (MUSIC) algorithm to obtain an MPC data set. Then, the channel prediction device can extract the average large-scale channel parameters of each sub-region through a clustering algorithm to establish a static channel database, such as Figure 1 In practical applications, it can be assumed that the large-scale parameters of the channel are known, so the unknown channel can be estimated by a small number of pilots to achieve the purpose of saving pilots.

[0058] However, in the above scheme, as the user position changes, the large-scale parameters of the channel will inevitably change, which still has a significant impact on the actual channel state, so this scheme will lead to a large channel acquisition error. In addition, this scheme still requires a certain pilot to be responsible for the acquisition of channel parameters, which still brings a lot of signaling overhead.

[0059] 7. Long Short-Term Memory (LSTM) network channel prediction solution: Use the collected CSI sequence data as the training data of the general LSTM network to train the network. For example, the system architecture of the LSTM network channel prediction can be as follows: Figure 2As shown in the figure, its input is set to the CSI matrix sampled at equal time intervals in the past sequence, denoted as: {H[t], H[t+1], ... H[t+D-1]}; its output is set to the instantaneous channel at the current moment, denoted as H[t+D]. Among them, the matrix H[t] is used to indicate the channel state information CSI corresponding to a certain moment t, and the matrix H[t] is a collection of channel information for each subcarrier.

[0060] The above-mentioned channel prediction method based on LSTM network can only mine the data correlation between discrete time channel sequences, while the channel comes from the propagation of actual electromagnetic waves in space and has special intrinsic data characteristics, and the above-mentioned LSTM network scheme cannot reflect this data characteristic, resulting in a low accuracy rate of actual channel prediction. In addition, when the prediction interval is large or the user moves at a fast speed, the performance is greatly attenuated. On the other hand, when users are at different moving speeds, the change characteristics of the channel often change to a certain extent, and this single model is difficult to adapt to different complex changes, so there is a certain performance bottleneck. In addition, for training data, the above-mentioned scheme needs to collect channel sampling data with equal time intervals for network training. In actual applications, base stations can often only obtain a small amount of scattered information, which makes it difficult to complete the training task of the network.

[0061] Based on the above problems, the present application provides a channel prediction method based on a spatial channel gradient network (SCGNet) of a neural network ordinary differential equation (ODE). Specifically, the neural network structure and model training and reasoning methods are designed based on channel characteristics, and the channel state information is predicted by combining the neural network ODE and position information, thereby improving the accuracy and feasibility of channel prediction.

[0062] In the following, the terms "first" and "second" are used for descriptive purposes only and are not to be understood as indicating or implying relative importance or implicitly indicating the number of the indicated technical features. Thus, a feature defined as "first" or "second" may explicitly or implicitly include one or more of the features. In the description of this embodiment, unless otherwise specified, "plurality" means two or more.

[0063] It should be noted that, in this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations or descriptions. Any embodiment or design described as "exemplary" or "for example" in this application should not be interpreted as being more preferred or more advantageous than other embodiments or designs. Specifically, the use of words such as "exemplary" or "for example" is intended to present related concepts in a specific way.

[0064] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.

[0065] First, the implementation environment and application scenarios of the embodiments of the present application are briefly introduced.

[0066] The present application can be applied to the existing NR system, and can also be used in any other wireless communication system with similar structure and function. Figure 3 As shown, the communication system at least includes one or more terminal devices 301 and network devices 302.

[0067] The terminal device 301 involved in the embodiment of the present application may be a user equipment (UE), wherein the UE includes a handheld device, a vehicle-mounted device, a vehicle, a wearable device or a computing device with a wireless communication function. Exemplarily, the UE may be a mobile phone, a smart phone, a tablet computer or a computer with a wireless transceiver function. The terminal device may also be a wireless data card, a personal digital assistant (PDA) computer, a wireless modem, a handheld device (handset), a laptop computer, a machine type communication (MTC) terminal, a virtual reality (VR) terminal, an augmented reality (AR) terminal, a wireless terminal in industrial control, a wireless terminal in unmanned driving, a wireless terminal in telemedicine, a wireless terminal in a smart grid, a wireless terminal in a smart city, a wireless terminal in a smart home, etc.

[0068] In the embodiment of the present application, the device for realizing the function of the terminal may be a terminal; or it may be a device capable of supporting the terminal to realize the function, such as a chip system, which may be installed in the terminal. In the embodiment of the present application, the chip system may be composed of a chip, or may include a chip and other discrete devices. In the technical solution provided in the embodiment of the present application, the technical solution provided in the embodiment of the present application is described by taking the device for realizing the function of the terminal device as a UE as an example.

[0069] The network device 302 involved in the embodiment of the present application may include a base station (BS), which may be a device deployed in a wireless access network and capable of wirelessly communicating with a terminal.

[0070] Among them, the base station may have various forms, such as macro base stations, micro base stations, relay stations and access points, etc. Exemplarily, the base station involved in the embodiment of the present application may be a base station in 5G or a base station in LTE, wherein the base station in 5G may also be called a transmission reception point (TRP) or gNB.

[0071] In an embodiment of the present application, the device for implementing the function of the network device may be a network device; or it may be a device that can support the network device to implement the function, such as a chip system, which may be installed in the network device.

[0072] In the technical solution provided in the embodiments of the present application, the apparatus for realizing the functions of the network device is a network device, and the network device is a base station as an example to describe the technical solution provided in the embodiments of the present application.

[0073] The technical solution provided in the embodiment of the present application can be applied to wireless communication between network equipment and terminal equipment. In the embodiment of the present application, the term "wireless communication" can also be referred to as "communication", and the term "communication" can also be described as "data transmission", "information transmission" or "transmission".

[0074] It should be noted that Figure 3 This is just an exemplary framework diagram. Figure 3 The number of network element nodes included in the Figure 3 In addition to the functional nodes shown, other nodes may also be included, such as core network equipment, gateway equipment, application servers, etc., without limitation. The access network equipment communicates with the core network equipment through a wired network or a wireless network, such as through an NG interface.

[0075] In the specific implementation, Figure 3 The network elements shown, such as terminal equipment and network equipment, can be used Figure 4 The structure shown or includes Figure 4 Parts shown. Figure 4 A schematic diagram of the structure of a communication device 400 provided in an embodiment of the present application. When the communication device 400 has the function of the terminal device described in the embodiment of the present application, the communication device 400 may be a terminal device or a chip or system on chip in the terminal device. When the communication device 400 has the function of the network device described in the embodiment of the present application, the communication device 400 may be a network device or a chip or system on chip in the network device.

[0076] like Figure 4 As shown, the communication device 400 may include a processor 401, a communication line 402, and a communication interface 403. Further, the communication device 400 may also include a memory 404. The processor 401, the memory 404, and the communication interface 403 may be connected via a communication line 402.

[0077] The processor 401 may be a central processing unit (CPU), a graphics processing unit (Graphics Processing Unit), a general-purpose processor, a network processor (NP), a digital signal processor (DSP), a microprocessor, a microcontroller, a programmable logic device, or any combination thereof. The processor 401 may also be other devices with processing functions, such as circuits, devices, or software modules.

[0078] The communication line 402 is used to transmit information between the components included in the communication device 400.

[0079] The communication interface 403 is used to communicate with other devices or other communication networks. The other communication networks may be Ethernet, Radio Access Network (RAN), Wireless Local Area Networks (WLAN), etc. The communication interface 403 may be an interface circuit, a pin, a radio frequency module, a transceiver, or any device capable of achieving communication.

[0080] The memory 404 is used to store instructions, where the instructions may be computer programs.

[0081] Among them, the memory 404 can be a read-only memory (ROM) or other types of static storage devices that can store static information and / or instructions, or a random access memory (RAM) or other types of dynamic storage devices that can store information and / or instructions, or an electrically erasable programmable read-only memory (EEPROM), a compact disc (CD-ROM) or other optical disc storage, optical disc storage, magnetic disk storage medium or other magnetic storage device, and the optical disc storage includes a compressed optical disc, a laser disc, an optical disc, a digital versatile disc, or a Blu-ray disc, etc.

[0082] It should be noted that the memory 404 can exist independently of the processor 401, or can be integrated with the processor 401. The memory 404 can be used to store instructions or program codes or some data, etc. The memory 404 can be located in the communication device 400, or can be located outside the communication device 400, without limitation. The processor 401 is used to execute the instructions stored in the memory 404 to implement the method provided in the following embodiments of the present application.

[0083] In one example, the processor 401 may include one or more CPUs, such as Figure 4 CPU0 and CPU1 in.

[0084] As an optional implementation, the communication device 400 includes multiple processors, for example, Figure 4 In addition to the processor 401, a processor 407 may also be included.

[0085] As an optional implementation, the communication device 400 further includes an output device 405 and an input device 406. Exemplarily, the input device 406 is a device such as a keyboard, a mouse, a microphone or a joystick, and the output device 405 is a device such as a display screen and a speaker.

[0086] It should be noted that the communication device 400 can be a wearable device, a desktop computer, a portable computer, a network server, a mobile phone, a tablet computer, a wireless terminal, an embedded device, a chip system or a Figure 4 In addition, Figure 4 The structure shown in the figure does not constitute a limitation on the communication device, except Figure 4 In addition to the components shown, the communication device may include more or fewer components than shown, or combine certain components, or arrange the components differently.

[0087] In the embodiment of the present application, the chip system may be composed of a chip, or may include a chip and other discrete devices.

[0088] The following describes in detail an implementation method for predicting channel state information based on location information provided by the present application in conjunction with the accompanying drawings.

[0089] like Figure 5 As shown, the present application provides a channel prediction method, which may include the following steps.

[0090] 501: The communication device obtains the position information of the user's test point, the position information of the reference point, and the historical channel state information corresponding to the reference point.

[0091] The communication device refers to a channel prediction device that can implement some or all of the steps in the channel prediction method provided in this application. For example, the communication device can be the aforementioned Figure 3 The terminal device described in the specification may be a network device, or may be a communication system consisting of at least one terminal device and a network device, and this application does not make any specific limitations on this.

[0092] Exemplarily, the channel prediction device can be implemented through distributed deployment, or can also be implemented through centralized deployment. For example, in a centralized deployment mode, the execution functions of steps 501 and 502 can be deployed on the same communication device, such as a network device, to implement channel prediction. For another example, in a distributed deployment mode, the execution functions of steps 501 and 502 can be deployed on different communication devices, respectively. For example, if a terminal is used to execute step 501 and a network device is used to execute step 502, the terminal can send the acquired location information of the user's test point, the location information of the reference point, and the historical channel state information corresponding to the reference point to the network device, and then the network device executes step 502 to obtain a channel prediction result. The possible deployment methods will be introduced in conjunction with specific implementation methods later, and will not be repeated here.

[0093] Among them, the channel prediction method provided by the present application is based on the prediction of position information, and at least two position information are required. Among them, the point to be measured can be used to indicate the user position corresponding to the moment of the channel to be predicted. For example, the moment of the channel to be predicted can be the current moment, or the next time unit (such as the next moment), then the point to be measured refers to the position of the user at the current moment or the next moment. The reference point is a spatial position different from the point to be measured. A set of data pairs is formed by the position information of the reference point and the point to be measured to perform channel prediction.

[0094] In one embodiment, the reference point may be a neighboring point (or detection point) closest to the point to be measured in a database of historical position information. That is, the present application can estimate the user's movement direction information through two relatively close position information, and can predict the channel state information of the point to be measured based on the user's movement direction information and the historical channel state information of the reference point, that is, to achieve channel state prediction at the current moment or the next moment. The movement direction information can be used to indicate the predicted movement path of the user at the point to be measured to the reference point.

[0095] In one implementation, in the above step 501, the communication device obtains the location information of the user's test point, the location information of the reference point, and the historical channel state information corresponding to the reference point, which may specifically include the following steps:

[0096] First, the communication device can predict the location information of the user's point to be measured based on the user's historical location information and / or historical channel state information.

[0097] Then, the communication device can determine the location information of the reference point and the historical channel state information of the reference point from the historical channel data set according to the location information of the point to be measured. The reference point is the detection point (such as a neighboring point) with the smallest distance from the point to be measured in the historical channel data set, and the historical channel data set includes a large amount of user location information and the historical channel state information corresponding to the location information. Thus, the communication device can obtain the historical channel state information corresponding to the reference point by querying the historical channel data set according to the location information of the reference point.

[0098] Furthermore, the communication device may obtain the movement direction information between the point to be measured and the reference point according to the position information of the point to be measured and the reference point, so as to execute the following step 502 .

[0099] 502: The communication device obtains the channel state information of the point to be measured according to the historical channel state information of the reference point and the moving direction information between the point to be measured and the reference point.

[0100] The historical channel state information of the reference point includes the angle delay domain channel matrix corresponding to the reference point. That is, in the embodiment of the present application, the channel prediction device can predict the channel state information of the point to be measured based on the angle delay domain channel matrix corresponding to the reference point.

[0101] In one implementation, channel prediction may be implemented through a data model of a neural network, and the above step 502 may specifically include:

[0102] Step 1: The communication device inputs the historical channel state information of the reference point and the motion direction information between the point to be measured and the reference point into the channel prediction model to obtain gradient information, wherein the gradient information is used to indicate the change information of the channel state in the motion direction.

[0103] Step 2: The communication device obtains the channel state information of the point to be measured according to the moving path and gradient information of the point to be measured.

[0104] The inverse operation of the derivative operation may be an integral budget, and therefore, the gradient information may be subjected to an integral operation based on the moving path to obtain a static channel prediction result.

[0105] For example, Wherein, G(x′) is the channel state information of the neighboring point x′, G(x) is the channel state information of the point x to be measured, and s is the path length from the neighboring point x′ to the point x to be measured.

[0106] like Figure 6As shown in the system architecture, the communication device (i.e., the channel prediction device / system) may include a channel prediction module and a channel positioning module. Among them, the channel positioning module can be used to estimate the position information of the point to be measured, which can be implemented by the data module of the neural network or other operation functions, and the present application does not make specific limitations on this. The channel prediction module is used to characterize the gradient operation of the channel state to the spatial position through the data model of the neural network, and to implement the channel operation. Among them, the operation function of the gradient information in the channel prediction module can be implemented by the channel prediction model.

[0107] Furthermore, Figure 6 The communication device shown may also include a historical channel data set for collecting historical location information, historical channel state information and other data, thereby providing training data and part of the input data for the neural network.

[0108] Next, by introducing the relevant technologies of channel state information, the principle and specific implementation method of the channel prediction model designed in this application are explained.

[0109] Assume that the base station is equipped with a uniform linear array (ULA), the antenna spacing is half a wavelength, and the orthogonal frequency division multiplexing (OFDM) modulation method is used. Among them, the base station can be equipped with N t Antennas, using N c subcarriers. The terminal device can be equipped with a single omnidirectional antenna. For each subcarrier, the static channel frequency response (CFR) can be expressed as:

[0110]

[0111] Where l is the subcarrier index, N pat h represents the total number of propagation paths, α p represents the path loss of the pth path, θ p represents the angle of arrival, d p is the total length of the propagation path, λ l is the wavelength of the corresponding subcarrier. e(θ) represents the steering vector of ULA:

[0112]

[0113] Where d is the distance between adjacent antennas and λ is the wavelength. Therefore, CFR can also be expressed as:

[0114] H=[h[1],h[2],...,h[N c ]]

[0115] As shown below, the angle delay domain channel matrix corresponding to the CSI can be expressed as:

[0116] G=V H HF,

[0117] in,

[0118] For the elements in the angle delay domain channel matrix corresponding to the arrival angle θ and the delay τ, the real and imaginary parts of the angle delay domain channel matrix can be expressed as:

[0119]

[0120] Among them, α τ,θ It's about p The inverse proportional function of ξ p Related to the electromagnetic properties of the material, Alternatively, in a practical system, different carrier wavelengths can be considered approximately equal, so ρ k d p ≈ρd p .

[0121] Based on the above-mentioned representation of the real and imaginary parts of the angle delay domain channel matrix, an embodiment of the present application provides a channel prediction scheme by calculating the change information of the channel state relative to the direction of motion, which can also be called the gradient information of the channel relative to the spatial position, and then performing integration operations based on the gradient information to achieve channel prediction.

[0122] In one implementation, the real part and the imaginary part of the angle delay domain channel matrix may be derivated with respect to the spatial position information to obtain gradient information, that is, the derivative of the channel relative to the spatial position, wherein the gradient information is used to indicate the change information of the channel state in the motion direction.

[0123] Among them, the real part of the angle delay domain channel matrix is ​​derived from the spatial position information to obtain:

[0124]

[0125] By taking the derivative of the imaginary part of the angle delay domain channel matrix with respect to the spatial position information, we can obtain:

[0126]

[0127] It can be obtained that the gradient information includes the first information D1 and the second information D2, and the first information D1 satisfies:

[0128]

[0129] The second information D2 satisfies:

[0130]

[0131] in, Used to represent the real part of the angle delay domain channel matrix, It is used to represent the imaginary part of the angle delay domain channel matrix, θ is used to represent the corresponding arrival angle in the angle delay domain channel matrix, τ is used to represent the corresponding delay in the angle delay domain channel matrix, d p It is used to indicate the total length of the channel transmission path, and m is used to indicate the location information.

[0132] It should be noted that, through the above content, the CSI matrix is ​​converted into an angle delay domain channel matrix, and after calculating the gradient information of the real and imaginary parts of the angle delay domain channel matrix for the spatial position in the above manner, it can be seen that the mathematical expression of the above gradient information has obvious advantages for designing the data model of the neural network. For the aforementioned existing channel prediction implementation scheme, the original complex channel matrix is ​​split into real and imaginary parts; and because the data model of the neural network can only process real numbers, it is necessary to splice the data processed by the real and imaginary parts together to obtain the result. However, this data processing method greatly destroys the internal correlation of the input data. In the expression of the above gradient information of the present application, the expression of the gradient information D1 or D2 is composed of the weighted sum of the real and imaginary parts of the angle delay domain channel matrix, respectively. Therefore, splitting the original angle delay domain channel matrix into real and imaginary parts through the above calculation process does not destroy the internal correlation of the data, but also helps network training and learning.

[0133] In addition, unlike the existing channel prediction scheme that needs to restore high-dimensional complex data (complex channel matrix) as output, the data model provided by the embodiment of the present application only needs to learn a simple mapping from high-dimensional complex data to low-dimensional simple data. Therefore, the parameter scale and complexity of the neural network can be significantly reduced. In addition, for the propagation path, the delay of the path and the change of the propagation angle can be ignored because the sampling interval is usually not long enough. Therefore, in the channel change process between two points, the index of the matrix element representing a specific propagation path is fixed, which ensures the robustness of the implementation method of the present application.

[0134] In summary, according to the expressions of the first information D1 and the second information D2 included in the aforementioned gradient information, a data model of a neural network, such as an SCGNet network, can be designed. In the embodiments of the present application, it can also be called a channel prediction model.

[0135] In one embodiment, Figure 7As shown, a possible implementation of a channel prediction model is shown, which can be composed of two parts: a scattering learning network and a directional embedding network.

[0136] Among them, the scattering learning network can adopt a fully connected structure, taking the angle delay domain channel matrix after CSI dimension transformation as input, and the output is a function of the total length of the channel transmission path The direction embedding network can take the motion direction information (such as the direction vector) between the test point and the reference point as input to obtain because Depending only on the direction of integration, you can choose to use sine and cosine values ​​as input parameters instead of angle values. The above input method can be regarded as a special case of positional encoding, which is often used in network structures such as transformers.

[0137] In addition, since the output of the directional embedding network is of dimension N t ×N c ×2 vector, so we can get t ×2 input simple replication N c times to balance the dimensions of input and output, thereby improving learning efficiency. Then, the final outputs of D1 and D2 can be directly obtained by simply adding and subtracting the output results of the scattering learning network using two products.

[0138] Based on the above-mentioned neural network structure, in one embodiment, the model training process of the channel prediction model may include: the communication device trains the data model according to the historical channel state information of multiple reference points and the motion direction information of multiple motion paths corresponding to each reference point to obtain a channel prediction model, and the channel prediction model is used to implement channel prediction.

[0139] Exemplarily, the training process of the channel prediction model may include the following steps.

[0140] Step 1: The communication device may construct a training data set, or obtain a training data set by collecting channel detection data.

[0141] The training data set may include the location information of multiple reference points, the location information of the z nearest neighboring points corresponding to each reference point, and the channel state information corresponding to each reference point or neighboring point. Thus, for each reference point, z data pairs including the reference point and the neighboring point can be obtained. Furthermore, the motion direction information of each data pair, such as the direction vector, can be obtained. For example, for each reference point, the reference point can be used as the starting point and the neighboring point as the end point; or, the neighboring point can be used as the starting point and the reference point as the end point, and the direction vector can be obtained by calculation based on the location information of the reference point and the neighboring point.

[0142] Step 2: The communication device may input the direction vector of each data pair and the channel state information of the neighboring points into the data model for model training.

[0143] Among them, since the input data of the channel prediction model of the present application includes the real part and the imaginary part of the angle delay domain channel matrix, the CSI matrix needs to be converted into the angle delay domain channel matrix before the input data model to obtain the real part G real and the imaginary part G imag .

[0144] The output of the data model can be set as the derivative of the channel state information of the reference point to the spatial position, that is, the aforementioned gradient information, including the first information D1 and the second information D2. By comparing the output of the data model with the derivative of the channel state information of the detected reference point to the spatial position, the loss function loss is calculated, and the model training is performed by repeatedly iteratively optimizing the parameters of the model; until the loss between the channel prediction output result of the model and the actual result is small enough to meet the preset standard, the model training is determined to be completed.

[0145] It should be noted that, as can be seen from the foregoing, the channel positioning module can also be implemented by a neural network, and the model training process can be implemented with reference to relevant algorithms or technologies. For example, the initial data model can be trained by collecting a large amount of user location information and / or historical channel state information as training data, and the output of the data model can be set as the estimated result of the reference point location information. This application does not make specific limitations on this.

[0146] Furthermore, after the channel prediction model is trained, the inference process of the channel prediction can be implemented. Exemplarily, the inference process of the channel prediction model can include the following steps.

[0147] Step 1: For the point x to be measured, find the known reference point closest to the position of the point x to be measured in the historical channel data set as the neighboring point x′.

[0148] Step 2: Based on the position information of the reference point x′ and the point to be measured x, calculate the movement direction information between the point to be measured and the reference point, for example, the direction vector.

[0149] Step 3: Input the channel state information of the reference point x′ (ie, the real and imaginary values ​​of the angle delay domain channel matrix) and the motion direction information obtained in step 2 into the channel prediction model, and the channel prediction model outputs the gradient information.

[0150] Through the inference process of the above channel prediction model, the derivative of the predicted channel can be obtained, so the above integral operation needs to be performed to obtain the final predicted channel result. Further, the following steps can be included.

[0151] Step 4: According to the moving path of the point to be measured, the gradient information is integrated to obtain the predicted channel, that is, the channel state information of the point to be measured.

[0152] It should be noted that the above embodiments of the present application Figure 7 This is only an example and only one possible design of a channel prediction model (such as corresponding to the second channel prediction model in the claims). A variety of different model designs and flexible deformations can be obtained through the above-mentioned channel prediction principles and schemes, and this application does not make any specific limitations on this.

[0153] In addition, the aforementioned step 502 includes a possible specific implementation step 2: the communication device obtains the channel state information of the point to be measured according to the moving path and gradient information of the point to be measured. In a possible implementation, this operation process can also be designed in the model. For example, the above Figure 7 The model shown can also include a layer of integration operation, which is used to implement the integration operation of the gradient information on the moving path to obtain the real and imaginary values ​​of the angle delay domain channel matrix, so that the output data corresponding to the deformed channel prediction model (such as corresponding to the first channel prediction model in the claim) is the angle delay domain channel matrix corresponding to the predicted reference point.

[0154] In one implementation, the above step 501 may specifically include: the communication device may be used to predict the location information of the user's point to be measured according to the user's historical location information and / or historical channel state information.

[0155] Specifically, the location information can be represented by a location feature vector. In one embodiment, the communication device can estimate the location feature vector of the user's point to be measured based on the user's historical location information. Alternatively, the communication device can estimate the location feature vector of the user's point to be measured based on the historical channel state information corresponding to the user's historical location information. Alternatively, the communication device can estimate the location feature vector of the user's point to be measured based on the historical location information of other users adjacent to the user. Alternatively, the communication device can obtain the location feature vector of the user's point to be measured based on the historical channel state information corresponding to the historical location information of other users adjacent to the user.

[0156] Further optionally, the communication device may also estimate the user's velocity characteristic vector based on the user's historical location information and / or historical channel state information, wherein the user's velocity characteristic vector may be used to estimate mobile channel information (or dynamic channel information).

[0157] In the above implementation manner, the channel state information predicted by the communication device is static channel information. Furthermore, the communication device can also obtain mobile channel information according to the speed characteristic information of the user.

[0158] In one embodiment, Figure 5 Said method may further comprise the following steps.

[0159] 503: The communication device performs Doppler elimination operation according to the channel state information of the point to be measured and the speed characteristic vector of the user to obtain the mobile channel information of the point to be measured.

[0160] Exemplarily, the static channel state information of the point to be tested is The speed feature vector of the user is V[t n+1 ], the communication device can be based on the predicted static channel state information With the velocity characteristic vector V[t n+1 ] Combined with Doppler elimination operation, such as The dynamic channel state information G[t n+1 ]. Among them, the operation It refers to the Hadamard product in matrix operations, or the basic product, E(V[t n+1 ]) is the Doppler elimination matrix.

[0161] In one embodiment, the communication device can obtain the location information of the point to be measured by the aforementioned Figure 6 Optionally, the channel positioning module and the channel prediction module can be deployed on the same communication device to achieve centralized deployment; or, the channel positioning module and the channel prediction module can be deployed on the same communication device to achieve distributed deployment, such as Figure 8 shown.

[0162] For example, Figure 8 The diagram shows an architecture of a channel prediction system. The entire system may include n nodes and 1 shared service node. The function of the channel prediction module may be deployed on the shared service node, so that the shared service node may be used to implement the channel pre-estimation function; the channel positioning module may be deployed on nodes 1, 2, ..., and n, so that n nodes may be used to implement the location information estimation function of the point to be measured. In this way, a distributed layout of data processing may be realized, the feasibility and comprehensiveness of obtaining historical data of users may be improved, and resource overheads such as data processing capability and power consumption of a certain node may be reduced, and the flexibility of scheme implementation may be improved.

[0163] Optionally, the shared service node may be a network device, such as a base station. Node 1, node 2, ... and node n may be multiple terminal devices.

[0164] In one implementation, each node can exchange historical location information and / or historical channel state information of the user. For example, node 2 can obtain historical channel state information, such as {H[t1], H[t2], …, H[t n ]} is sent to neighboring node 1 and node 3, etc. Thus, the distributed node can obtain the location information of the user's test point according to the user's historical location information and / or historical channel state information obtained from the neighboring nodes; further, the user's speed feature vector can also be obtained.

[0165] Combination Figure 8 As shown, the exemplary interactive flow and implementation process of the distributed deployed channel prediction system may specifically include the following steps.

[0166] Step 1: Obtain a channel prediction module (including a channel prediction model) through model training, and deploy the channel prediction module on a shared service node.

[0167] Step 2: Through model training, a channel positioning module (including a channel positioning model) is obtained, and the channel positioning module is deployed on node 1, node 2, ... and node n.

[0168] Step 3: Node i obtains the historical channel state information of the user, such as {H[t1], H[t2], …, H[t n ]}.

[0169] Here, i may be a positive integer, specifically one or more of {1, 2, ..., n}.

[0170] Step 4: Node i sends the historical channel state information to neighbor node j.

[0171] Here, j may be a positive integer, which is different from i, and may specifically be one or more of {1, 2, ..., n}.

[0172] Step 5: Node i or node j obtains the historical channel state information {H[t1], H[t2], …, H[t n ]} Input the channel positioning module to obtain the position feature vector X[t n+1 ], and the user's velocity feature vector V[t n+1 ].

[0173] Step 6: Node i or node j converts the position feature vector X[t n+1 ] is sent to the shared service node.

[0174] Optionally, node i or node j can also convert the user's speed feature vector V[t n+1 ] is also sent to the shared service node.

[0175] Step 7: The shared service node converts the position feature vector X[t n+1 ] Input the channel prediction module to obtain the static channel state information of the user reference point

[0176] Optionally, the following step 8 may also be included, or the following steps 9-10 may also be included.

[0177] Step 8: Shared service nodes use static channel status information With the velocity characteristic vector V[t n+1 ], and obtain the dynamic channel state information G[t n+1 ].

[0178] Step 9: The shared service node will predict the static channel state information Sent to node i or node j.

[0179] Step 10: Node i or node j receives static channel state information and the user's velocity feature vector V[t n+1 ], and obtain the dynamic channel state information G[t n+1 ].

[0180] The above-mentioned implementation mode of the present application realizes channel prediction by converting the channel state information into an angle delay domain channel matrix, and calculating the gradient information of the spatial position based on the angle delay domain channel matrix, thereby training and reasoning the neural network. In the above-mentioned implementation mode, compared with the training data of the neural network in the prior art, the difficulty of data collection is reduced and the feasibility of model training is improved. In addition, by predicting the gradient information obtained by deriving the angle delay domain channel matrix, the internal correlation of the data parameters affecting the channel state will not be destroyed, thereby improving the accuracy of channel prediction, so that under different implementation scenario settings (such as different sampling densities, different sequence sampling intervals, etc.), there are obvious advantages in prediction accuracy.

[0181] In addition, based on the aforementioned implementation, the present application also provides a channel prediction device, such as Fig. 9 As shown, the apparatus 900 may include a channel positioning module 901 and a channel prediction module 902 .

[0182] Among them, the channel positioning module 901 is used to obtain the position information of the point to be measured, as well as the position information of the reference point and the historical channel state information corresponding to the reference point, wherein the historical channel state information of the reference point includes the angle delay domain channel matrix corresponding to the reference point.

[0183] The channel prediction module 902 is used to predict the channel state information of the point to be measured based on the historical channel state information of the reference point and the movement direction information between the point to be measured and the reference point, wherein the movement direction information is used to represent the predicted movement path of the user at the point to be measured to the reference point.

[0184] In one implementation, the channel prediction module 902 is specifically used to input the historical channel state information of the reference point and the motion direction information between the point to be measured and the reference point into a first channel prediction model to obtain the channel state information of the point to be measured.

[0185] In one embodiment, the channel prediction module 902 is specifically used to input the historical channel state information of the reference point and the movement direction information between the point to be tested and the reference point into a second channel prediction model to obtain gradient information, wherein the gradient information is used to indicate the change information of the channel state in the movement direction; and the channel state information of the point to be tested is obtained according to the moving path of the point to be tested and the gradient information.

[0186] In one embodiment, the channel prediction module 902 is also used to obtain the user's velocity characteristic vector based on the user's historical location information and / or historical channel state information; perform Doppler elimination operation based on the channel state information of the point to be measured and the user's velocity characteristic vector to obtain the mobile channel information of the point to be measured.

[0187] In one implementation, the gradient information includes first information D1 and second information D2, wherein the first information D1 satisfies:

[0188]

[0189] The second information D2 satisfies:

[0190] in, Used to represent the real part of the angle delay domain channel matrix, It is used to represent the imaginary part of the angle delay domain channel matrix, θ is used to represent the corresponding arrival angle in the angle delay domain channel matrix, τ is used to represent the corresponding delay in the angle delay domain channel matrix, d p It is used to indicate the total length of the channel transmission path, and m is used to indicate the location information.

[0191] In one implementation, the channel prediction module 902 is further configured to train the first channel prediction model or the second channel prediction model according to historical channel state information of multiple reference points and motion direction information of multiple motion paths corresponding to each reference point.

[0192] In one embodiment, the channel positioning module 901 is specifically used to: predict the location information of the user's test point based on the user's historical location information, and / or the historical channel status information corresponding to the historical location information; determine the location information of the reference point and the historical channel status information of the reference point from the historical channel data set based on the location information of the test point, wherein the reference point is the detection point in the historical channel data set with the shortest distance from the test point, and the historical channel data set includes the location information of multiple detection points and the corresponding relationship between the historical channel status information.

[0193] In one embodiment, the channel positioning module 901 is specifically used to: predict the user position based on the historical position information of the user to obtain the position feature vector of the point to be measured; or, predict the user position based on the historical channel state information corresponding to the historical position information of the user to obtain the position feature vector of the point to be measured; or, predict the user position based on the historical position information of adjacent users to obtain the position feature vector of the point to be measured; or, predict the user position based on the historical channel state information corresponding to the historical position information of adjacent users to obtain the position feature vector of the point to be measured.

[0194] Specifically, the communication device 900 can implement the functions of the communication device in the above possible implementation modes. For details, please refer to the detailed descriptions in the above-mentioned method examples, which will not be repeated here.

[0195] Understandable, combined Figure 4 As shown, when the above device is an electronic device, the above communication device may further include a transceiver module, which may be a transceiver, such as including an antenna and a radio frequency circuit, etc. The channel positioning module 901 and the channel prediction module 902 may be implemented by a processing module, such as a processor, such as a baseband chip, etc.

[0196] It should be noted that the specific execution process and embodiments in the above-mentioned device can refer to the steps and related descriptions executed by the communication device in the above-mentioned method embodiment, and the technical problems solved and the technical effects brought about can also refer to the contents described in the above-mentioned embodiments, which will not be repeated here.

[0197] In this embodiment, the communication device is presented in the form of dividing various functional modules in an integrated manner. The "module" here may refer to a specific circuit, a processor and memory that executes one or more software or firmware programs, an integrated logic circuit, and / or other devices that can provide the above functions. In a simple embodiment, those skilled in the art can imagine that the communication device can be used as described above. Figure 4 The form shown.

[0198] For example, Fig. 9The function / implementation process of the channel prediction module 902 in Figure 4 The processor 401 in the embodiment calls the computer program instructions stored in the memory 404 to implement.

[0199] In some embodiments, Figure 4 The processor 401 in the device 400 can call the computer-executable instructions stored in the memory 404 so that the device 400 can execute the operations implemented by the communication device in the above-mentioned various method embodiments, thereby realizing the above-mentioned various possible implementation methods of the present application.

[0200] The communication device in each of the above-mentioned device embodiments may completely correspond to the terminal in the method embodiment, and the corresponding steps are performed by the corresponding modules or units. For example, when the device is implemented in the form of a chip, the transceiver module may be an interface circuit of the chip for receiving signals from other chips or devices. The above-mentioned transceiver module for sending or receiving is an interface circuit of the device, which is used to send signals to other devices. For example, when the device is implemented in the form of a chip, the transceiver module may be an interface circuit for sending signals to other chips or devices.

[0201] In an exemplary embodiment, a computer-readable storage medium including instructions or a computer program product is also provided, and the instructions can be executed by the processor 401 of the communication device 400 to complete the method of the above embodiment. Therefore, the technical effects that can be obtained can refer to the above method embodiment, and will not be repeated here.

[0202] The present application also provides a computer program product, which includes instructions. When the instructions are executed, the computer can respectively perform operations of a terminal device or a network device corresponding to the above method.

[0203] The embodiment of the present application also provides a system chip, which includes: a processing unit and a communication unit, the processing unit, for example, may be a processor, and the communication unit, for example, may be an input / output interface, a pin or a circuit, etc. The processing unit may execute computer instructions so that the communication device used by the chip performs the operations of the terminal device and the network device in the method provided in the embodiment of the present application.

[0204] Optionally, any one of the communication devices provided in the above-mentioned embodiments of the present application may include the system chip.

[0205] Optionally, the computer instructions are stored in a storage unit.

[0206] An embodiment of the present application also provides a communication system, which may include: any one of the terminals and network devices in the above-mentioned implementation modes.

[0207] In the above embodiments, it can be implemented in whole or in part by software, hardware, firmware or any combination thereof. When implemented using a software program, it can be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions. When the computer program instructions are loaded and executed on a computer, the process or function according to the embodiment of the present application is generated in whole or in part. The computer can be a general-purpose computer, a special-purpose computer, a computer network, or other programmable devices.

[0208] Those skilled in the art will readily appreciate other embodiments of the present application after considering the specification and practicing the invention disclosed herein. The present application is intended to cover any modification, use or adaptation of the present application, which follows the general principles of the present application and includes common knowledge or customary technical means in the art that are not disclosed in the present application.

[0209] Finally, it should be noted that the above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any changes or substitutions within the technical scope disclosed in the present application should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A channel prediction method, characterized in that: The method comprises: Acquire the position information of the point to be measured, the position information of the reference point and the historical channel state information corresponding to the reference point, wherein the historical channel state information of the reference point includes the angle delay domain channel matrix corresponding to the reference point; The channel state information of the point to be measured is predicted based on the historical channel state information of the reference point and the movement direction information between the point to be measured and the reference point, wherein the movement direction information is used to represent the predicted movement path of a user at the point to be measured to the reference point.

2. The method according to claim 1, characterized in that Predicting the channel state information of the point to be measured according to the historical channel state information of the reference point and the motion direction information between the point to be measured and the reference point specifically includes: The historical channel state information of the reference point and the motion direction information between the point to be measured and the reference point are input into a first channel prediction model to obtain the channel state information of the point to be measured.

3. The method according to claim 1, characterized in that Predicting the channel state information of the point to be measured according to the historical channel state information of the reference point and the motion direction information between the point to be measured and the reference point specifically includes: Inputting historical channel state information of the reference point and information of the motion direction between the point to be measured and the reference point into a second channel prediction model to obtain gradient information, wherein the gradient information is used to indicate change information of the channel state in the motion direction; The gradient information is integrated according to the moving path of the point to be measured to obtain the channel state information of the point to be measured.

4. The method according to any one of claims 1 to 3, characterized in that: The method further comprises: Obtaining a velocity feature vector of the user according to historical location information and / or historical channel state information of the user; A Doppler elimination operation is performed according to the channel state information of the point to be measured and the speed characteristic vector of the user to obtain the mobile channel information of the point to be measured.

5. The method according to claim 3, characterized in that: The gradient information includes first information D1 and second information D2, wherein the first information D1 satisfies: The second information D2 satisfies: in, Used to represent the real part of the angle delay domain channel matrix, It is used to represent the imaginary part of the angle delay domain channel matrix, θ is used to represent the corresponding arrival angle in the angle delay domain channel matrix, τ is used to represent the corresponding delay in the angle delay domain channel matrix, d p It is used to indicate the total length of the channel transmission path, and m is used to indicate the location information.

6. The method according to claim 2 or 3, characterized in that: The method further comprises: The first channel prediction model or the second channel prediction model is trained according to the historical channel state information of multiple reference points and the motion direction information of multiple motion paths corresponding to each reference point.

7. The method according to any one of claims 1 to 6, characterized in that: The acquiring the position information of the user's test point, the position information of the reference point, and the historical channel state information corresponding to the reference point specifically includes: Predicting the location information of the user's point to be measured based on the user's historical location information and / or the historical channel state information corresponding to the historical location information; According to the position information of the point to be measured, the position information of the reference point and the historical channel status information of the reference point are determined from the historical channel data set, wherein the reference point is the detection point in the historical channel data set with the shortest distance from the point to be measured, and the historical channel data set includes the corresponding relationship between the position information of multiple detection points and the historical channel status information.

8. The method according to claim 7, characterized in that The obtaining the location information of the point to be measured according to the historical location information of the user and / or the historical channel state information corresponding to the historical location information specifically includes: Perform user location prediction based on the user's historical location information to obtain a location feature vector of the point to be measured; or Predict the user's position based on the historical channel state information corresponding to the historical position information of the user to obtain the position feature vector of the point to be measured; or Predict the user's location based on the historical location information of the adjacent users to obtain the location feature vector of the point to be measured; or The user position is predicted according to the historical channel state information corresponding to the historical position information of the adjacent users to obtain the position feature vector of the point to be measured.

9. A channel prediction device, characterized in that: The device comprises: A channel positioning module, used to obtain the position information of the point to be measured, the position information of the reference point and the historical channel state information corresponding to the reference point, wherein the historical channel state information of the reference point includes the angle delay domain channel matrix corresponding to the reference point; A channel prediction module is used to predict the channel state information of the point to be measured based on the historical channel state information of the reference point and the movement direction information between the point to be measured and the reference point, wherein the movement direction information is used to represent the predicted movement path of a user at the point to be measured to the reference point.

10. The device according to claim 9, characterized in that The channel prediction module is specifically used to input the historical channel state information of the reference point and the motion direction information between the point to be measured and the reference point into a first channel prediction model to obtain the channel state information of the point to be measured.

11. The device according to claim 9, characterized in that The channel prediction module is specifically used to input the historical channel state information of the reference point and the motion direction information between the point to be measured and the reference point into a second channel prediction model to obtain gradient information, wherein the gradient information is used to indicate the change information of the channel state in the motion direction; The channel state information of the point to be measured is obtained according to the moving path of the point to be measured and the gradient information.

12. The device according to any one of claims 9 to 11, characterized in that: The channel prediction module is also used to: Obtaining a velocity feature vector of the user according to historical location information and / or historical channel state information of the user; A Doppler elimination operation is performed according to the channel state information of the point to be measured and the speed characteristic vector of the user to obtain the mobile channel information of the point to be measured.

13. The device according to claim 11, characterized in that The gradient information includes first information D1 and second information D2, wherein the first information D1 satisfies: The second information D2 satisfies: in, Used to represent the real part of the angle delay domain channel matrix, It is used to represent the imaginary part of the angle delay domain channel matrix, θ is used to represent the corresponding arrival angle in the angle delay domain channel matrix, τ is used to represent the corresponding delay in the angle delay domain channel matrix, d p It is used to indicate the total length of the channel transmission path, and m is used to indicate the location information.

14. The device according to claim 10 or 11, characterized in that The channel prediction module is also used to train the first channel prediction model or the second channel prediction model according to the historical channel state information of multiple reference points and the motion direction information of multiple motion paths corresponding to each reference point.

15. The device according to any one of claims 9 to 14, characterized in that: The channel positioning module is specifically used for: Predicting the location information of the user's point to be measured based on the user's historical location information and / or the historical channel state information corresponding to the historical location information; According to the position information of the point to be measured, the position information of the reference point and the historical channel status information of the reference point are determined from the historical channel data set, wherein the reference point is the detection point in the historical channel data set with the shortest distance from the point to be measured, and the historical channel data set includes the corresponding relationship between the position information of multiple detection points and the historical channel status information.

16. The device according to claim 15, characterized in that The channel positioning module is specifically used for: Perform user location prediction based on the user's historical location information to obtain a location feature vector of the point to be measured; or Predict the user's position based on the historical channel state information corresponding to the historical position information of the user to obtain the position feature vector of the point to be measured; or Predict the user's location based on the historical location information of the adjacent users to obtain the location feature vector of the point to be measured; or The user position is predicted according to the historical channel state information corresponding to the historical position information of the adjacent users to obtain the position feature vector of the point to be measured.

17. A device, characterized in that: include: A processor, wherein the processor is coupled to a memory, wherein the memory is used to store programs or instructions, and when the programs or instructions are executed by the processor, the device executes the method according to any one of claims 1 to 8.

18. A chip, characterized in that: include: A processor, wherein the processor is coupled to a memory, wherein the memory is used to store programs or instructions, and when the programs or instructions are executed by the processor, the chip executes the method as claimed in any one of claims 1 to 8.

19. A computer-readable storage medium having a computer program or instruction stored thereon, characterized in that: When the computer program or instructions are executed, the computer is caused to perform the method according to any one of claims 1 to 8.

20. A computer program product, comprising computer program code, characterized in that: When the computer program code is executed on a computer, the computer is enabled to implement the method according to any one of claims 1 to 8.

Citation Information

Cited By

  • Wireless communication system signal sending and receiving method based on AI model

    CN121690447A

  • A method for transmitting and receiving a signal of a wireless communication system based on an AI model

    CN121690447B