Beam domain channel state information generation method, device, equipment and medium

Abstract

The invention discloses a beam domain channel state information generation method, apparatus and device, and a medium. The method comprises the steps of obtaining a geographic position coordinate of a target user terminal from a preset base station; inputting the geographic position coordinates into a conditional diffusion model, and outputting target beam domain channel state information of the target user terminal; wherein the conditional diffusion model is generated based on different geographic position coordinates and beam domain channel state information training corresponding to pilot signals respectively sent according to the geographic position coordinates. By utilizing the method, the beam domain channel state information can be directly predicted and generated according to the geographic position coordinate of the target user terminal through the trained conditional diffusion model, and the user terminal does not need to transmit a real pilot frequency at the geographic position coordinate of the target user terminal to carry out channel estimation to obtain the beam domain channel state information; the step of channel estimation is omitted, and the accuracy and efficiency of perception auxiliary communication are remarkably improved.

Description

Technical Field

[0001] The embodiments of the present invention relate to the field of communication technologies, and in particular, to a method, apparatus, device, and medium for generating beam-domain channel state information. Background Art

[0002] Advances in wireless technologies, such as the post-fifth-generation (B5G) and sixth-generation (6G) systems, are expected to support various sensing-based applications, such as indoor positioning, Wi-Fi sensing, and radar sensing. With the base station (BS) using massive multiple-input multiple-output and orthogonal frequency-division multiplexing technologies, high-resolution wireless signals can be transmitted, enabling high-precision sensing. The base station can implement a wireless positioning function while enhancing the service capabilities of user terminals (UTs), processing wireless signals carrying transmission data and location information, and seamless integration of sensing and communication has become a natural choice.

[0003] How to utilize the location information and sensing data obtained through sensors or base stations becomes crucial. Digital Twin (DT) is a technology for modeling physical processes through digital technologies. By analyzing the historical location information and sensing data of user terminals, identifying problems and optimizing by observing trends in the digital twin, predictive analysis can be provided for subsequent accurate decision-making. Digital twin will play a key role in the operation and infrastructure management of 6G ISAC networks, providing important support for the construction of future communication networks. In a 6G system, obtaining the beam-domain channel state information (CSI-B) of user terminals is crucial for subsequent communication performance. Through channel twin, instead of relying solely on obtaining channel information in a communication system, partial channel information can be obtained in advance using location information. How to establish a digital twin of the channel to utilize location information to obtain partial channel information in advance has become a challenge. Summary of the Invention

[0004] The embodiments of the present invention provide a method, apparatus, device, and storage medium for generating beam-domain channel state information, which can generate beam-domain channel state information according to the locations of multiple user terminals without pilots, with excellent performance, and significantly improve the accuracy and efficiency of sensing-assisted communication.

[0005] In a first aspect, the embodiments of the present invention provide a method for generating beam-domain channel state information, including:

[0006] Obtaining the geographical location coordinates of a target user terminal from a preset base station;

[0007] Input the geographical location coordinates into a conditional diffusion model to output the target beam domain channel state information of the target user terminal; wherein, the conditional diffusion model is trained based on different geographical location coordinates and the beam domain channel state information corresponding to the pilot signals respectively sent according to each geographical location coordinate.

[0008] In a second aspect, an embodiment of the present invention further provides a device for generating beam domain channel state information, and the device includes:

[0009] An acquisition module, configured to acquire the geographical location coordinates of a target user terminal from a preset base station;

[0010] A beam domain channel state information determination module, configured to input the geographical location coordinates into a conditional diffusion model to output the target beam domain channel state information of the target user terminal; wherein, the conditional diffusion model is trained based on different geographical location coordinates and the beam domain channel state information corresponding to the pilot signals respectively sent according to each geographical location coordinate.

[0011] In a third aspect, an embodiment of the present disclosure further provides an electronic device, and the electronic device includes:

[0012] One or more processors;

[0013] A storage device, configured to store one or more programs,

[0014] When the one or more programs are executed by the one or more processors, the one or more processors implement the method for generating beam domain channel state information provided by the embodiment of the present disclosure.

[0015] In a fourth aspect, an embodiment of the present disclosure further provides a storage medium containing computer-executable instructions, and the computer-executable instructions are used to execute the method for generating beam domain channel state information provided by the embodiment of the present disclosure when executed by a computer processor.

[0016] In a fifth aspect, an embodiment of the present disclosure further provides a computer program product, and the computer program product includes a computer program, and the computer program implements the method for generating beam domain channel state information provided by the embodiment of the present disclosure when executed by a processor.

[0017] The present invention discloses a method, device, equipment and storage medium for generating beam domain channel state information. This method can directly predict and generate beam domain channel state information according to the geographical location coordinates of a target user terminal through a trained conditional diffusion model, without obtaining the beam domain channel state information by transmitting real pilot signals from the user terminal for channel estimation at the geographical location coordinates of the target user terminal, saving the steps of channel estimation and significantly improving the accuracy and efficiency of sensing-assisted communication. Description of the Drawings

[0018] In conjunction with the accompanying drawings and with reference to the following specific embodiments, the above and other features, advantages, and aspects of the various embodiments of the present disclosure will become more apparent. Throughout the drawings, the same or similar reference numerals denote the same or similar elements. It should be understood that the drawings are schematic and that the original and the elements are not necessarily drawn to scale.

[0019] Figure 1 It is a flowchart of a method for generating beam domain channel state information provided by an embodiment of the present disclosure;

[0020] Figure 2 It is a flowchart of a method for training a conditional diffusion model provided by an embodiment of the present disclosure;

[0021] Figure 3 It is a flowchart of another method for training a conditional diffusion model provided by an embodiment of the present disclosure;

[0022] Figure 4 It is an example diagram of a U-shaped neural network architecture provided by an embodiment of the present disclosure;

[0023] Figure 5 It is an example diagram of a method for generating beam domain channel state information provided by an embodiment of the present disclosure;

[0024] Figure 6 It is a schematic two-dimensional plane diagram of a simulation provided by an embodiment of the present disclosure;

[0025] Figure 7 It is an example diagram of a simulation result provided by an embodiment of the present disclosure;

[0026] Figure 8 It is a schematic structural diagram of a device for generating beam domain channel state information provided by an embodiment of the present disclosure;

[0027] Figure 9 It is a schematic structural diagram of an electronic device provided by an embodiment of the present disclosure. Detailed Description of the Embodiments

[0028] The embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although some embodiments of the present disclosure are shown in the drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as limited to the embodiments set forth herein. On the contrary, these embodiments are provided to more thoroughly and completely understand the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not used to limit the protection scope of the present disclosure.

[0029] It should be understood that the various steps described in the method embodiments of the present disclosure can be executed in different orders and / or in parallel. In addition, the method embodiments may include additional steps and / or omit the steps shown. The scope of the present disclosure is not limited in this regard.

[0030] As used herein, the term "including" and its variations are open-ended, that is, "including but not limited to". The term "based on" means "at least partially based on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.

[0031] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order of the functions performed by these devices, modules or units or their interdependent relationships.

[0032] It should be noted that the modifications of "one" and "multiple" mentioned in the present disclosure are illustrative rather than restrictive. Those skilled in the art should understand that unless otherwise clearly stated in the context, it should be understood as "one or more".

[0033] The names of the messages or information exchanged between multiple devices in the embodiments of the present disclosure are only for illustrative purposes and are not used to limit the scope of these messages or information.

[0034] It can be understood that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, usage scopes, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.

[0035] For example, when responding to receiving an active request from a user, a prompt message is sent to the user to clearly prompt the user that the operation requested by the user will require obtaining and using the user's personal information. Thus, the user can autonomously choose whether to provide personal information to software or hardware such as an electronic device, an application program, a server or a storage medium that performs the operations of the technical solutions of the present disclosure according to the prompt message.

[0036] As an optional but non-limiting implementation manner, the manner of sending a prompt message to the user in response to receiving an active request from the user may be, for example, in the form of a pop-up window, and the prompt message may be presented in text in the pop-up window. In addition, the pop-up window may also carry a selection control for the user to choose "agree" or "disagree" to provide personal information to the electronic device.

[0037] It is understandable that the above-mentioned notification and the process of obtaining user authorization are only illustrative and do not limit the implementation manner of the present disclosure. Other manners that comply with relevant laws and regulations can also be applied to the implementation manner of the present disclosure.

[0038] It is understandable that the data involved in the present technical solution (including but not limited to the data itself, the acquisition or use of the data) should comply with the requirements of corresponding laws, regulations and related provisions.

[0039] Massive Multiple Input Multiple Output (MIMO) and Orthogonal Frequency Division Multiplexing (OFDM) transmission technologies are key components of the fifth-generation (5G) wireless cellular system. They improve the spectral efficiency and also bring better sensing and positioning capabilities to the system. In addition, 5G is expected to achieve a high spatial density of base stations in urban scenarios to support high-speed communication and high-precision positioning services.

[0040] The massive MIMO channel response has been used in geometry-based and fingerprint-based positioning methods. The geometry-based method relies on the estimation of Delay Of Arrival (DoA) and Angle of Arrival (AoA), which requires a Line of Sight (LoS) path and precise synchronization between the Base Station (BS) and the User Terminals (UT). Currently, different types of channel characteristics can be extracted as fingerprints related to the location of the user terminal, that is, beam domain channel state information. Currently, the beam domain channel state information only depends on obtaining the beam domain channel state information in the communication system. How to achieve obtaining the beam domain channel state information in advance using location information will provide new opportunities for the physical layer design of communication.

[0041] Figure 1 It is a flowchart of a method for generating beam domain channel state information provided by an embodiment of the present disclosure. The embodiment of the present disclosure is applicable to the situation of providing a solution to the problem of generating beam domain channel state information through the location of a user terminal without using pilots. This method can be executed by a device for generating beam domain channel state information, and this device can be implemented in the form of software and / or hardware. Optionally, it can be implemented by a base station or a host computer of the base station.

[0042] As Figure 1 shown, a method for generating beam domain channel state information provided by an embodiment of the present disclosure may specifically include the following steps:

[0043] S110. Obtain the geographical location coordinates of a target user terminal from a preset base station.

[0044] In this embodiment, the preset base station may be a device with the geographical location coordinates of the target user terminal. The preset base station can directly measure the geographical location of the target user terminal, or it can be obtained by the target user terminal reporting. In some inventive embodiments, the preset base station may be a base station in a large-scale multiple-input multiple-output orthogonal frequency division multiplexing system with signal transceiver, data processing, and positioning calculation functions. Among them, the preset base station can be configured with a large-scale uniform linear array antenna, and the antenna spacing is half a wavelength. The antenna array with a half-wavelength spacing is used to form a highly directional beam, improving the signal coverage range and intensity, and being able to more accurately distinguish signals in different directions. The target user terminal can be located within the coverage range of the preset base station and communicate with the preset base station. The preset base station can measure the geographical location coordinates of the target user terminal. The geographical location coordinates can include longitude and latitude coordinates, projection coordinates, local coordinates, system coordinates, etc.

[0045] Specifically, the ways to obtain the geographical location coordinates of the target user terminal from the preset base station can include: for example, connecting to the management interface of the preset base station and sending a query request to obtain the required geographical location coordinates of the target user terminal; for another example, accessing the log file storage location (such as local storage or remote server) in the preset base station to read and parse the log file to extract the required geographical location coordinates of the target user terminal; further, it is also possible to obtain the access permission and authentication information of the application programming interface (API) of the preset base station, call the API interface, send a data request, and receive and parse the geographical location coordinates of the target user terminal returned by the API. Any way that can obtain the geographical location coordinates of the target user terminal from the preset base station other than the above methods is acceptable.

[0046] S120. Input the geographical location coordinates into the conditional diffusion model to output the target beam domain channel state information of the target user terminal, where the conditional diffusion model is trained based on different geographical location coordinates and the beam domain channel state information corresponding to the pilot signals respectively sent according to each geographical location coordinate.

[0047] In this embodiment, the conditional diffusion model may be a generative model based on the diffusion process, guiding the generation process of the beam domain channel state information by using the geographical location coordinates as the conditional information in the conditional diffusion model. The conditional diffusion model is trained based on different geographical location coordinates and the beam domain channel state information corresponding to the pilot signals respectively sent according to each geographical location coordinate. Among them, the pilot signal is a known signal sent by the terminal device at each geographical location coordinate and is used for beam domain channel estimation.

[0048] The training process of the conditional diffusion model includes two stages: forward diffusion and reverse inference. The forward diffusion stage is a process of gradually adding noise to the input training data to transform the training data into a Gaussian noise distribution. The forward diffusion stage can be implemented through Markov chains, stochastic differential equations, generative adversarial networks, etc. The reverse inference stage starts from the Gaussian noise distribution and gradually denoises through a neural network to generate beam domain channel state information. The geographical location coordinates can be introduced as conditional information in the reverse inference to guide the generation process of the beam domain channel state information. The beam domain channel state information can be information about the propagation characteristics of signals in specific beam directions in a wireless communication system. The propagation characteristics can include the signal intensity and phase change in specific beam directions, the multipath interference generated by the signal, and the time difference and angle range generated when the signal arrives at the receiving end through different paths.

[0049] Specifically, the geographical location coordinates are input into the conditional diffusion model as conditional information. The conditional diffusion model generates Gaussian noise data, which is used as the starting state of the generation process. Then, the noise is gradually removed through multiple iterations, and finally, the target beam domain channel state information of the target user terminal is output. Among them, the method of generating Gaussian noise data can include: for example, randomly sampling from a preset multivariate Gaussian distribution with a mean of 0 and a variance of a preset value to generate Gaussian noise with the same dimension as the target beam domain channel state information; or directly generating it through a database for scientific computing, and the database can include a numerical computing library or a scientific computing library, etc.

[0050] The technical solution of the embodiment of the present invention can directly predict and generate beam domain channel state information according to the geographical location coordinates of the target user terminal through the trained conditional diffusion model, without obtaining the beam domain channel state information by transmitting real pilots from the user terminal at the geographical location coordinates of the target user terminal for channel estimation, saving the steps of channel estimation and significantly improving the accuracy and efficiency of perception-assisted communication.

[0051] On the basis of the above embodiment, a method for generating beam domain channel state information further includes the process of training the conditional diffusion model, as Figure 2 shown, the training process of the conditional diffusion model includes the following steps:

[0052] S210. Obtain a set of reference points including multiple different geographical location coordinates and the corresponding beam domain channel state information.

[0053] In this embodiment, the reference point can be a reference point or a reference position for positioning. Multiple reference points with different geographical location coordinates are located within the communication area corresponding to the preset base station, and each reference point records the corresponding position coordinates. The beam domain channel state information corresponding to the reference point can be a known pilot signal transmitted by the terminal device at the position of the reference point, and the preset base station extracts the beam domain channel state information corresponding to each geographical location coordinate based on the received pilot signal.

[0054] Specifically, it is obtained including reference points that record multiple different geographical location coordinates and the beam domain channel state information determined corresponding to the reference points. The set formed by all the beam domain channel state information is used as the set of beam domain channel state information.

[0055] Exemplarily, by controlling a self-guided vehicle to travel within the communication area corresponding to the preset base station, reference points recording spatial coordinates are established on the preset path of the vehicle. The vehicle is controlled to send a detection signal (pilot information) to the base station at the position of the reference point, and the preset base station captures the detection signal and determines the beam domain channel state information related to the positions of these reference points.

[0056] S220. Add incremental noise to the set of beam domain channel state information based on the Markov chain to obtain the first set of beam domain channel state information.

[0057] In this embodiment, the Markov chain can describe the transition process from one state to another through the state space and transition probability. Different noise states of the data are used as the states of the Markov chain, and the rules and iteration times for transitioning from one noise state to the next are defined. The incremental noise can be noise gradually introduced as the number of processing steps increases. The intensity of the introduced noise can be controlled by a variance regulator, a noise scheduling value, or by adjusting the noise parameters. The incremental noise introduced in this embodiment can be Gaussian noise, uniform noise, or Laplace noise. The first set of beam domain channel state information can be a data set in which the beam domain channel state information is finally transformed into a Gaussian noise distribution after adding incremental noise through the Markov chain. Among them, the dimensions of the first set of beam domain channel state information, the set of beam domain channel state information, and the incremental noise are the same, and the dimension is the number of beam domain channel state information in the set of beam domain channel state information. The first set of beam domain channel state information is obtained when the preset number of iterations is reached.

[0058] Specifically, through a Markov chain with a predefined number of iterations, the set of beam domain channel state information is used as the initial state. In each iteration, incremental noise is introduced into the initial state, and the data state is updated. The iterative process is repeated until the preset number of iterations is reached, and the data when the preset number of iterations is reached is used as the final first set of beam domain channel state information.

[0059] S230. Use the first beam domain channel state information set and the geographical location coordinates of the reference point as a training and testing set to train the initial neural network, and use the trained neural network as a conditional diffusion model.

[0060] In this embodiment, the initial neural network can be a U-shaped neural network-based model specifically designed for the diffusion model. The training and testing set can be a data set used to train the model and evaluate the performance of the trained model. The training and testing set can include a training data set and a testing data set. Training the initial neural network with the training and testing set is the reverse inference stage of the conditional diffusion model. During the reverse inference process, the learning rate is dynamically adjusted by setting a fixed learning rate or by an adaptive learning rate (such as the adaptive moment estimation optimizer) strategy. The neural network during the reverse inference process can be a U-shaped neural network-based model or a residual neural network model specifically designed for the diffusion model.

[0061] Specifically, use the training data set to train the initial neural network, and then use the testing data set to perform a performance test on the above-trained neural network. If the test result of the performance test meets the expectation, use the trained neural network as the conditional diffusion model; otherwise, retrain and test the initial neural network until the test result of the performance test meets the expectation and then stop.

[0062] Figure 3 The flowchart of another training method for the conditional diffusion model provided by the embodiments of the present disclosure is further optimized and extended based on the above training process of the conditional diffusion model, such as Figure 3 shown, and specifically includes the following steps:

[0063] S310. Obtain a set of reference points including multiple different geographical location coordinates and the beam domain channel state information corresponding to the reference points.

[0064] Exemplarily, reference points with multiple different geographical location coordinates are located within the area corresponding to a preset communication system, which can be a large-scale MIMO-OFDM system. Each preset base station is equipped with a uniform linear array (ULA) containing N antennas, and there are K single-antenna UTs randomly distributed in the target positioning area. The uplink transmission link of the preset communication system is a dual-beam channel model, and the beam-domain channel state information corresponding to the reference point can be extracted from the above dual-beam channel model. The dual-beam channel model can be used to represent the uplink channel estimation result of the preset base station using a first matrix, a beam-domain channel matrix, and a second matrix; among them, the beam-domain channel matrix is used to determine the beam-domain channel state information, the first matrix includes the sampling direction vector in the spatial domain, and the second matrix includes the sampling direction vector in the frequency domain; the elements in the beam-domain channel matrix are used to represent the fading coefficients of the spatial beam and the frequency beam, the sampling direction vector in the spatial domain corresponds to the direction cosine of the arrival angle of each path between the preset base station and the target user terminal, and the sampling direction vector in the frequency domain corresponds to the arrival delay of each path between the preset base station and the target user terminal.

[0065] Exemplarily, for convenience, cos(θ)=Θ represents the direction cosine of the arrival angle θ of the path signal, and the direction vector a(Θ) of the direction cosine of the arrival angle θ of the path signal is defined in the spatial domain as:

[0066]

[0067] where T represents the transpose, j represents the complex number, and N c represents the number of subcarriers of the OFDM system, Θ represents the direction cosine of the arrival angle θ of the path signal, and N g represents the number of cyclic prefixes (CP), T s represents the system sampling time interval, δ represents the interval between adjacent subcarriers, and the calculation formula is as follows: Therefore, the direction vector b(τ) corresponding to the arrival delay DoAτ in the frequency domain is:

[0068]

[0069] where Θ k,p，l represents the direction cosine of the p-th path between the k-th UT and the l-th BS, where the value range of k is (0, M1), the value range of l is (0, M2), M1 represents the maximum number of user terminals that the preset communication system can accommodate, M2 represents the number of base stations in the preset communication system, j represents the complex number, and N c represents the number of subcarriers of the OFDM system, T represents the transpose, δ represents the interval between adjacent subcarriers, and τ k,pDenote the DoA of the p-th path between the k-th UT and the l-th BS, β k,l,p Denote the fading coefficient of the p-th path between the k-th UT and the l-th BS, β k,l,p Satisfy the complex Gaussian distribution, that is Denote the variance of the complex Gaussian distribution. Assume that the maximum arrival delay spread τ of the k-th UT k,max Is less than the duration of the cyclic prefix, that is τ k,max ≤N g T s . The spatial-frequency domain channel matrix (the uplink channel estimation result of the preset base station) H between the k-th UT and the l-th BS k,l Can be expressed as

[0070]

[0071] Where, f c Denotes the carrier frequency, τ k,p Denote the DoA of the p-th path between the k-th UT and the l-th BS, β k,l,p Denote the fading coefficient of the p-th path between the k-th UT and the l-th BS, a(Θ k,p ) Denotes the direction vector of the direction cosine of the signal arrival angle θ of the p-th path between the k-th UT and the l-th BS, b(τ k,p ) Denotes the direction vector corresponding to the arrival delay DoA τ in the frequency domain of the signal of the p-th path between the k-th UT and the l-th BS, T denotes the transpose, P k Denote the p-th path between the k-th UT and the l-th BS.

[0072] Correspondingly, the element h of the spatial-frequency domain channel matrix when the path between the k-th UT and the l-th BS is at the direction cosine of the signal arrival angle θ and the direction vector corresponding to the arrival delay DoA τ in the frequency domain k,l (Θ, τ) is defined as:

[0073]

[0074] Define the i-th sampling set And the j-th sampling set Where, the sampling set The number of samplings is N a , the sampling set The number of samplings is N d , Θ denotes the direction cosine, τ denotes the arrival delay, δ() denotes the impulse response function, β k,l,p Denote the fading coefficient of the p-th path between the k-th UT and the l-th BS, Θ k,p，l Denote the direction cosine of the p-th path between the k-th UT and the l-th BS, fc Denotes the carrier frequency, τ k,p Denotes the DoA of the p-th path between the k-th UT and the l-th BS.

[0075] To ensure quantization accuracy, N a ≥ N c , N d ≥ N g , Evenly spaced in (-1, 1], τ j In (0, N g T s . N c Denotes the number of subcarriers in the OFDM system, N g Denotes the number of cyclic prefixes. When N a and N d Are large enough, (1) can be approximately expressed by sampling the steering vectors a(Θ i ), b(τ j ) as

[0076]

[0077] Where, N a Denotes the number of samples in the sampling set , N d Denotes the number of samples in the sampling set . Denotes that a(Θ i ) represents the spatial domain sampling steering vector of the direction cosine of the i-th sample, and b(τ j ) represents the frequency domain sampling steering vector of the direction cosine of the j-th sample. N g Denotes the number of cyclic prefixes, T s Denotes the system sampling time interval, δ represents the interval between adjacent subcarriers, Denotes the beam domain channel matrix between the k-th UT and the l-th BS.

[0078] In addition, define the first matrix A and the second matrix B as:

[0079]

[0080] Then (2) can be expressed by matrix multiplication as

[0081] H k,l = AG k,l B T (3)

[0082] (3) can be called a double-beam channel model. a(Θ) represents the direction vector of the direction cosine of the arrival angle θ of the path signal in the spatial domain, b(τ) represents the direction vector corresponding to the arrival delay DoA τ in the frequency domain, and G kl is expressed as the beam-domain channel matrix between the k-th UT and the l-th BS, represents the dimension as N c ×N a of the channel matrix elements, with the dimension of N p ×N d representing the channel matrix elements, and g k,t (Θ i ,τ j ) represents the (i,j)-th element of the beam-domain channel matrix. The i-th row elements of the beam-domain channel matrix represent the fading coefficients in space, and the j-th column elements of the beam-domain channel matrix represent the fading coefficients of the frequency beam. Each sampled spatial / frequency direction vector corresponds to a physical beam in the spatial / frequency domain. When N a = N c , N d = N g , A is a Discrete Fourier Transform (DFT) matrix, and B is a submatrix of the first N g columns of the DFT matrix.

[0083] The wavenumber-domain channel state information Γ of the direction cosine and DoA of the p-th path between the k-th UT and the l-th BS k (Θ k,p ,τ k,p ) can be expressed as

[0084] Let Then define Ω k to represent the energy coupling matrix, representing the N a ×N d -dimensional real channel matrix elements, h k,l (Θ,τ) represents the element of the spatial-frequency domain channel matrix when the direction cosine of the arrival angle θ of the path between the k-th UT and the l-th BS and the direction vector corresponding to the arrival delay DoA τ in the frequency domain. Θ k,p，l represents the direction cosine of the p-th path between the k-th UT and the l-th BS, τ k,p represents the DoA of the p-th path between the k-th UT and the l-th BS, E represents the expectation, * represents the conjugate, and δ() represents the impulse response function.

[0085] The (i,j)-th element [Ω k，l i,j of the energy coupling matrix is:​

[0086]

[0087] G k,l The elements of

[0088]

[0089] wherein, represents expectation, * represents conjugate, δ() represents the impulse response function, and G k,l is expressed as the beam-domain channel matrix between the k-th UT and the l-th BS, represents the channel power of the path between the k-th UT and the l-th BS, represents the i-th sampling set and represents the j-th sampling set, and δ() represents the impulse response function. (4) represents the beam-domain channel matrix, and G k,l The different elements of are uncorrelated. Ω k，l reflects the channel information of power, delay, and angle of each path related to the scatterers between the k-th UT and the l-th BS. Ω k，l is called "beam-domain channel state information", indicating that it captures the time-invariant characteristics of the channel.

[0090] S320. Obtain the preset number of iterations.

[0091] In this embodiment, the number of iterations determines the fineness of noise addition. The more the number of iterations, the finer the noise addition process, but the higher the computational complexity. Usually, the number of iterations is set according to actual requirements and application scenarios.

[0092] S330. Iterate with the set of beam-domain channel state information as the initial state.

[0093] In this embodiment, the initial state can be the state before the start of iteration. Taking the set of beam-domain channel state information as the initial state, start the iterative process of noise addition.

[0094] S340. In each iteration, calculate the variance value of the current noise and generate incremental Gaussian noise, and add the incremental Gaussian noise to the set of beam-domain channel state information.

[0095] In this embodiment, the variance value determines the intensity of the noise, and the intensity of the introduced noise can be controlled by a variance regulator, a noise scheduling value, or by adjusting the variance value. The dimension of the incremental Gaussian noise is equal to the number of beam-domain channel state information in the set.

[0096] Specifically, according to the current iteration number, determine the variance value of the current noise, and generate Gaussian noise with the same dimension as the beam domain channel state information set. Among them, the mean of the noise is 0, and the variance is the current variance value. Add the generated incremental Gaussian noise to the beam domain channel state information set to update the state of the set.

[0097] S350. When the number of iterations reaches the preset number of iterations, stop the iteration, and use the final state at the stop of the iteration as the first beam domain channel state information set.

[0098] Specifically, when the number of iterations reaches the preset value, stop the noise addition process, and use the beam domain channel state information set at the stop of the iteration as the first beam domain channel state information set.

[0099] S360. Divide the training and test set into a training set and a test set according to the preset data set division ratio.

[0100] Specifically, divide the original data set into a training set and a test set according to the preset ratio. The training set is used for model training, and the test set is used for model performance evaluation. Random splitting, stratified splitting and other methods can also be used to ensure the distribution consistency of the training set and the test set. Exemplarily, the splitting ratio is 70:30 or 80:20.

[0101] S370. Combine the geographical location coordinates of the reference point as conditional information with the first beam domain channel state information set as the model input.

[0102] In this embodiment, the geographical location coordinates of the reference point are used to provide spatial information and enhance the context awareness ability of the model. Concatenate or embed the geographical location coordinates with the beam domain channel state information set to form a conditional input for the training and inference of the conditional diffusion model.

[0103] S380. Use relative entropy as the loss function, and optimize the model parameters of the initial neural network with the goal of minimizing relative entropy to obtain the basic conditional diffusion model.

[0104] In this embodiment, relative entropy is used to measure the difference between the model prediction distribution and the true distribution, and it is a common loss function for conditional diffusion models. The basic conditional diffusion model can be a basic model that iterates multiple times on the training set to gradually optimize the model parameters.

[0105] Specifically, through gradient descent or other optimization algorithms, minimize relative entropy, adjust the model parameters of the initial neural network, iterate multiple times on the training set, and gradually optimize the model parameters to obtain the basic conditional diffusion model.

[0106] S390. Use the test set to perform a performance test on the basic conditional diffusion model. If the test result meets the preset model test pass condition, then use the basic conditional diffusion model as the final conditional diffusion model; otherwise, retrain the initial neural network.

[0107] Specifically, use the test set to evaluate the performance of the basic conditional diffusion model, such as prediction accuracy, generalization ability, and computational efficiency, etc. The preset test pass conditions include but are not limited to prediction error thresholds, convergence speed, and resource consumption, etc. If the test result meets the pass condition, then use the basic conditional diffusion model as the final model; otherwise, retrain the initial neural network until the condition is met.

[0108] Exemplarily, the forward process of the conditional diffusion model includes:

[0109] According to the state transition probability formula of the Markov chain as follows:

[0110]

[0111] Among them, represents the distribution, and the transition probability density function q(Ω t |Ω t-1 ) represents a pre-designed Gaussian distribution, Ω t is the first beam domain channel state information set, and the variance scheduler represents the set of noises added for all iterations from the first iteration to the last iteration, satisfying 0 < β1 < β2 < … < β D < 1. In each iteration, the variance scheduler is used to control the intensity of the added noise, which is a function that changes with the number of iterations. According to the current number of iterations, substitute it into the variance scheduler to determine the variance value of the noise added in this iteration, and generate incremental Gaussian noise based on the variance value. Among them, D represents the preset number of iterations, t represents the number of the current iteration, that is, the time step, I G is a G-dimensional identity matrix, and G is the number of beam domain channel state information in the set. Use the set of beam domain channel state information as the initial state Ω0

[0112] According to the Markov chain derivation, the expression for obtaining Ω t from Ω0 can be obtained:

[0113] When t ≥ 2,

[0114]

[0115] Among them:

[0116]

[0117] Set α0 = 1 to ensure consistency with the case at t = 1.

[0118] Among them, the set of beam domain channel state information is used as the initial state Ω0, and the transition probability density function q(Ω t |Ω t-1 ) represents a pre-designed Gaussian distribution, q(Ω t-1 |Ω t , Ω0) represents the probability of the transition probability density function from the initial state to the (t - 1)-th to the t-th time, q(Ω t-1 |Ω0) represents the transition probability density function from the initial state to the (t - 1)-th time, q(Ω t |Ω0) represents the transition probability density function from the initial state to the t-th time, Ω t is the first beam domain channel state information set, represents the mean value from the initial state Ω0 to the final state Ω t at the t-th iteration, α represents the noise scheduling value, and the variance scheduler represents the set of noises added for all iterations from the first iteration to the last iteration, satisfying 0 < β1 < β2 < … < β D < 1. In each iteration, the current variance value is calculated according to the variance scheduler and incremental Gaussian noise is generated. Among them, D represents the preset number of iterations, t represents the number of the current iteration, i.e., the time step, and I G is an identity matrix with a dimension of G, which is the number of beam domain channel state information in the set.

[0119] The reverse process of the conditional diffusion model includes:

[0120] The reverse process is constructed as a reverse Markov chain starting with Gaussian noise. The geographical location coordinates p of the reference point are used as the conditional information of the reverse Markov chain, and the relative entropy is used as the loss function. The model parameters of the initial neural network are optimized with the goal of minimizing the relative entropy. The loss function is:

[0121]

[0122] Among them, represents the expected value of the squared error between the true value Ω0 of the model and the output value ε σ of the model, represents the square of the Euclidean norm, ε σ is the output value of the model, t represents the time step, I G is an identity matrix with a dimension of G, which is the number of beam domain channel state information in the set.

[0123] Figure 4An example diagram of the U-shaped neural network architecture provided by the embodiments of the present disclosure is as follows. Figure 4 As shown, the U-shaped neural network architecture of the embodiment includes an input layer, an encoder, a decoder, a residual network module, and an output layer. Among them, the input layer includes three inputs: the first beam domain channel state information set, location condition information, and time step. At the same time, a self-attention mechanism is introduced between the residual network modules, and the normalization layer in the residual network module is replaced by group normalization. The principle of group normalization is to divide the channels (C) of the input features into a fixed number of groups (G), each group contains C / G channels, and the mean and variance are calculated independently within the group for normalization.

[0124] In this embodiment, the preset data set segmentation ratio can be a preset ratio, which is set according to the actual situation and is not specifically limited in this embodiment. The geographical location coordinates of the reference point are combined with the first beam domain channel state information set as the model input. The relative entropy is used as the loss function, and the model parameters of the initial neural network are optimized with the goal of minimizing the relative entropy to obtain the basic conditional diffusion model. The performance of the basic conditional diffusion model is tested using the test set. If the test results of the performance test meet the preset model test passing conditions, the basic conditional diffusion model is used as the final conditional diffusion model; otherwise, the initial neural network is retrained.

[0125] Exemplarily, Figure 5 An example diagram of the method for generating beam domain channel state information provided by the embodiments of the present disclosure; the process of generating beam domain channel state information is divided into an offline and an online mode. In the offline mode, that is, Figure 5 in the training stage, the location coordinates and beam domain channel state information of the reference points with known locations are recorded to establish a discrete fingerprint database, which is used to train the conditional diffusion model; in the online mode, the base station can obtain the geographical location coordinates of the user, and using the conditional diffusion model, the base station can generate the beam domain channel state information under specific location conditions to establish a continuous fingerprint database.

[0126] The embodiments of the present invention use a geometry-based model to simulate the large-scale MIMO-OFDM wireless transmission environment and consider a two-dimensional propagation scenario. The base station is equipped with a uniform linear array. Figure 6 A schematic diagram of the simulated two-dimensional plane provided by the embodiments of the present disclosure. In Figure 6A two - dimensional plane schematic diagram is given, showing the simulation configuration. The coordinates (X, Y) of this plane correspond to the X - axis and the Y - axis. It is assumed that the base station is located at the origin of coordinates, and the uniform linear array it is equipped with is parallel to the Y - axis. The target area is a square area centered at (200, 0) meters with each side length of 50 meters. In the training phase, the target area is discretized along its geometric spatial dimensions, that is, along the X - axis and the Y - axis. The position coordinates of each grid point are recorded as the true position of the reference point, and the beam - domain channel state information of the reference point is obtained through the transmission channel model. Thus, a training data set of the reference point can be obtained. In the online mode, 500 randomly distributed user terminals are generated within the entire target area as a set of test user terminals to evaluate the performance of the conditional diffusion model in generating beam - domain channel state information. Since the accurate beam - domain channel state information is known, the normalized mean squared error (NMSE, Normalized Mean Squared Error) between the generated beam - domain channel state information and the accurate beam - domain channel state information is used to evaluate the accuracy.

[0127] Continuing from the above, the batch size of the training data used in the simulation is 32, the gradient optimizer is selected as the Adam optimizer, and the learning rate is set to 0.0002 to obtain good performance. Figure 7 This is an example diagram of the simulation results provided by the embodiments of the present disclosure; as Figure 7 shown, the comparison of the simulation results of the normalized mean squared error of different - position beam - domain channel state information methods. By comparing the performance lines of the generated results and the number of pilots T p equal to 3 and 10, it can be observed that increasing the number of pilots T p can improve the performance of the existing pilot - based methods. For example, when the number of pilots is T p = 10, it can estimate 24 user terminals simultaneously and achieve an NMSE of 8.28. In contrast, the proposed generation method does not require pilot information and can generate the beam - domain channel state information of 100 user terminals only using the known position coordinates. In addition, as the training progresses, the performance gradually improves and exceeds the baseline method, and the NMSE gradually increases from - 2.92 to - 9.387. The simulation results prove the accuracy and effectiveness of this method. Note that all the above - mentioned training processes are carried out offline. In the generation phase, the saved network parameters can be selected according to needs without increasing the computational complexity of the generation phase. By comparing the performance lines of the generated results and the interpolation results, it can be seen that when the data in the training set of this method exceeds 3000, an NMSE of 6.97 can be achieved, exceeding the NMSE of the existing interpolation methods. Although the existing interpolation methods are natural and convenient, for higher precision, the proposed generation method can achieve more satisfactory results.

[0128] Figure 8The embodiment of the present invention also provides a schematic structural diagram of a method and device for generating beam domain channel state information, as shown in Figure 8 FIG. 424, including an acquisition module 410 and a beam domain channel state information determination module 420.

[0129] The acquisition module 410 is configured to obtain the geographical location coordinates of the target user terminal from a preset base station;

[0130] The beam domain channel state information determination module 420 is configured to input the geographical location coordinates into a conditional diffusion model and output the target beam domain channel state information of the target user terminal; wherein, the conditional diffusion model is trained and generated based on different geographical location coordinates and the beam domain channel state information corresponding to the pilot signals respectively sent according to each geographical location coordinate.

[0131] For the technical solution provided by the embodiment of the present disclosure, using this method: without the need for pilots, it is possible to simultaneously generate beam domain channel state information from the positions of multiple user terminals, with excellent performance, significantly improving the accuracy and efficiency of sensing-assisted communication.

[0132] Further, the beam domain channel state information determination module 420 may be configured to:

[0133] Obtain a set of reference points with multiple different geographical location coordinates and the beam domain channel state information corresponding to the reference points;

[0134] Add incremental noise to the set of beam domain channel state information based on a Markov chain to obtain a first set of beam domain channel state information;

[0135] Use the first set of beam domain channel state information and the geographical location coordinates of the reference points as a training and test set to train an initial neural network, and use the trained neural network as the conditional diffusion model.

[0136] Further, the beam domain channel state information determination module 420 may also be configured to:

[0137] Obtain a predefined variance scheduler and a preset number of iterations;

[0138] Iterate with the set of beam domain channel state information as the initial state;

[0139] In each iteration, calculate the current variance value according to the variance scheduler and generate incremental Gaussian noise, and add the incremental Gaussian noise to the set of beam domain channel state information; the dimension of the incremental Gaussian noise is equal to the number of beam domain channel state information in the set;

[0140] Stop iterating when the preset number of iterations is reached, and obtain the first beam domain channel state information set.

[0141] Further, the beam domain channel state information determination module 420 can also be used to:

[0142] Divide the training and testing set into a training set and a testing set according to a preset data set splitting ratio;

[0143] Combine the geographical location coordinates of the reference point as conditional information with the first beam domain channel state information set as the model input;

[0144] Use relative entropy as the loss function, and optimize the model parameters of the initial neural network with the goal of minimizing the relative entropy to obtain a basic conditional diffusion model;

[0145] Perform performance testing on the basic conditional diffusion model using the testing set. If the test result of the performance testing meets the preset model testing passing condition, then use the basic conditional diffusion model as the final conditional diffusion model, otherwise retrain the initial neural network.

[0146] Further, the beam domain channel state information determination module 420 can also be used to:

[0147] The beam domain channel state information characterizes the channel information of the power, delay, and angle of each path related to the scatterers between the target user terminal and the preset base station.

[0148] Further, the beam domain channel state information determination module 420 can also be used to:

[0149] The conditional diffusion model includes a model based on a U-shaped neural network.

[0150] The above device can execute the methods provided in all the foregoing embodiments of the present invention, and has the corresponding functional modules and beneficial effects for executing the above methods. Technical details not described in detail in this embodiment can be seen in the methods provided in all the foregoing embodiments of the present invention.

[0151] Figure 9FIG. 0 shows a schematic structural diagram of an electronic device 10 that can be used to implement an embodiment of the present invention. The electronic device is intended to represent various forms of digital computers, such as, for example, laptop computers, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframe computers, and other suitable computers. The electronic device can also represent various forms of mobile devices, such as, for example, personal digital assistants, cellular phones, smart phones, wearable devices (such as helmets, glasses, watches, etc.) and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely examples and are not intended to limit the implementation of the present invention described and / or claimed herein.

[0152] As Figure 9 shown, the electronic device 10 includes at least one processor 11, and a memory communicatively connected to the at least one processor 11, such as a read-only memory (ROM) 12, a random access memory (RAM) 13, etc., wherein the memory stores a computer program executable by the at least one processor. The processor 11 can perform various appropriate actions and processes according to the computer program stored in the read-only memory (ROM) 12 or the computer program loaded from the storage unit 18 into the random access memory (RAM) 13. In the RAM 13, various programs and data required for the operation of the electronic device 10 can also be stored. The processor 11, the ROM 12, and the RAM 13 are connected to each other via a bus 14. An input / output (I / O) interface 15 is also connected to the bus 14.

[0153] Multiple components in the electronic device 10 are connected to the I / O interface 15, including: an input unit 16, such as a keyboard, a mouse, etc.; an output unit 17, such as various types of displays, speakers, etc.; a storage unit 18, such as a magnetic disk, an optical disk, etc.; and a communication unit 19, such as a network card, a modem, a wireless communication transceiver, etc. The communication unit 19 allows the electronic device 10 to exchange information / data with other devices via a computer network such as the Internet and / or various telecommunication networks.

[0154] The processor 11 can be various general-purpose and / or special-purpose processing components with processing and computing capabilities. Some examples of the processor 11 include, but are not limited to, a central processing unit (CPU), a graphics processing unit (GPU), various dedicated artificial intelligence (AI) computing chips, various processors running machine learning model algorithms, a digital signal processor (DSP), and any suitable processor, controller, microcontroller, etc. The processor 11 executes the various methods and processes described above, such as the beam domain channel state information generation method.

[0155] In some embodiments, the method for generating beam domain channel state information may be implemented as a computer program tangibly embodied in a computer-readable storage medium, such as storage unit 18. In some embodiments, part or all of the computer program may be loaded and / or installed onto the electronic device 10 via the ROM 12 and / or the communication unit 19. When the computer program is loaded into the RAM 13 and executed by the processor 11, one or more steps of the method for generating beam domain channel state information described above may be performed. Alternatively, in other embodiments, the processor 11 may be configured to perform the method for generating beam domain channel state information by any other suitable means (e.g., by means of firmware).

[0156] The various embodiments of the systems and techniques described above in this document may be implemented in digital electronic circuitry, integrated circuit systems, field programmable gate arrays (FPGA), application specific integrated circuits (ASIC), application specific standard products (ASSP), systems on a chip (SOC), complex programmable logic devices (CPLD), computer hardware, firmware, software, and / or combinations thereof. These various embodiments may include: being implemented in one or more computer programs executable and / or interpretable on a programmable system including at least one programmable processor, which may be a special-purpose or general-purpose programmable processor that receives data and instructions from a storage system, at least one input device, and at least one output device, and transmits the data and instructions to the storage system, the at least one input device, and the at least one output device.

[0157] The computer program for implementing the method of the present invention may be written in any combination of one or more programming languages. These computer programs may be provided to a processor of a general-purpose computer, a special-purpose computer, or other programmable data processing device, such that when the computer program is executed by the processor, the functions / operations specified in the flowchart and / or block diagram are implemented. The computer program may be executed entirely on the machine, partially on the machine, as a stand-alone software package partially on the machine and partially on a remote machine, or entirely on a remote machine or server.

[0158] In the context of the present invention, a computer-readable storage medium can be a tangible medium that can contain or store a computer program for use by or in connection with an instruction execution system, apparatus, or device. The computer-readable storage medium can include, but is not limited to, electronic, magnetic, optical, electromagnetic, infrared, or semiconductor systems, apparatus, or devices, or any suitable combination of the foregoing. Alternatively, the computer-readable storage medium can be a machine-readable signal medium. More specific examples of the machine-readable storage medium would include an electrical connection based on one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.

[0159] In order to provide interaction with a user, the systems and techniques described herein can be implemented on an electronic device having: a display device (e.g., a CRT (cathode ray tube) or LCD (liquid crystal display) monitor) for displaying information to the user; and a keyboard and a pointing device (e.g., a mouse or a trackball) by which the user can provide input to the electronic device. Other kinds of devices can also be used to provide interaction with the user; for example, the feedback provided to the user can be any form of sensory feedback (e.g., visual feedback, auditory feedback, or tactile feedback); and input from the user can be received in any form (including acoustic input, speech input, or tactile input).

[0160] The systems and techniques described herein can be implemented in a computing system that includes a back-end component (e.g., as a data server), or a computing system that includes a middleware component (e.g., an application server), or a computing system that includes a front-end component (e.g., a user computer having a graphical user interface or a web browser through which the user can interact with an implementation of the systems and techniques described herein), or a computing system that includes any combination of such back-end, middleware, or front-end components. The components of the system can be interconnected to each other by any form or medium of digital data communication (e.g., a communication network). Examples of communication networks include: local area network (LAN), wide area network (WAN), blockchain network, and the Internet.

[0161] A computing system may include a client and a server. The client and the server are generally far from each other and usually interact via a communication network. The client-server relationship is created by computer programs that run on respective computers and have a client-server relationship with each other. The server can be a cloud server, also known as a cloud computing server or a cloud host, which is a host product in the cloud computing service system, solving the defects of difficult management and weak business scalability existing in traditional physical hosts and VPS services.

[0162] It should be understood that various forms of the processes shown above can be used, with steps reordered, added, or deleted. For example, the steps recited in the present invention can be executed in parallel, sequentially, or in a different order, as long as the desired results of the technical solution of the present invention can be achieved, and no limitation is made herein.

[0163] The above specific embodiments do not constitute a limitation on the protection scope of the present invention. Those skilled in the art should understand that various modifications, combinations, sub-combinations, and substitutions can be made according to design requirements and other factors. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention shall be included within the protection scope of the present invention.

Claims

1. A method for generating beam domain channel state information, characterized in that: include: Acquire the geographic location coordinates of the target user terminal from a preset base station; The geographic location coordinates are input into a conditional diffusion model, and the target beam domain channel state information of the target user terminal is output; wherein the conditional diffusion model is generated based on different geographic location coordinates and the beam domain channel state information corresponding to the pilot signal sent respectively according to each of the geographic location coordinates.

2. The method according to claim 1, characterized in that The training process of the conditional diffusion model includes: Acquire a set of reference points including coordinates of a plurality of different geographical locations and the beam domain channel state information corresponding to the reference points; Adding incremental noise to the set of beam-domain channel state information based on a Markov chain to obtain a first set of beam-domain channel state information; The first beam domain channel state information set and the geographical location coordinates of the reference point are used as a training test set to train the initial neural network, and the trained neural network is used as the conditional diffusion model.

3. The method according to claim 2, characterized in that The adding incremental noise to the set of beam domain channel state information based on the Markov chain to obtain a first set of beam domain channel state information includes: Get the preset number of iterations; Iterate using the set of beam domain channel state information as an initial state; In each iteration, the variance value of the current Gaussian noise is calculated and an incremental Gaussian noise is generated, and the incremental Gaussian noise is added to the set of the beam-domain channel state information; the dimension of the incremental Gaussian noise is equal to the number of the beam-domain channel state information in the set; When the number of iterations of the set of beam domain channel state information reaches the preset number of iterations, the iteration is stopped, and the final state of the set of beam domain channel state information when the iteration is stopped is used as the first set of beam domain channel state information.

4. The method according to claim 2, characterized in that: The step of using the first beam domain channel state information set and the geographical location coordinates of the reference point as a training test set to train the initial neural network, and using the trained neural network as the conditional diffusion model, includes: Dividing the training test set into a training set and a test set according to a preset data set segmentation ratio; combining the geographical location coordinates of the reference point as condition information with the first beam domain channel state information set as model input; Taking relative entropy as a loss function, optimizing the model parameters of the initial neural network with the goal of minimizing the relative entropy, and obtaining a basic conditional diffusion model; The basic conditional diffusion model is performance tested using the test set. If the test result meets the preset model test pass condition, the basic conditional diffusion model is used as the final conditional diffusion model, otherwise the initial neural network is retrained.

5. The method according to claim 1, characterized in that The beam domain channel state information represents the channel information of power, delay and angle of each path related to the scatterer between the target user terminal and the preset base station.

6. The method according to claim 1, characterized in that The conditional diffusion model includes a U-type neural network based model or a residual based neural network model.

7. A beam domain channel state information generating device, characterized in that: include: An acquisition module, used to acquire the geographic location coordinates of the target user terminal from a preset base station; A beam domain channel state information determination module is used to input the geographic location coordinates into a conditional diffusion model and output the target beam domain channel state information of the target user terminal; wherein the conditional diffusion model is generated based on different geographic location coordinates and the beam domain channel state information corresponding to the pilot signal sent respectively according to each of the geographic location coordinates.

8. An electronic device, characterized in that: The electronic device comprises: at least one processor; and a memory communicatively connected to the at least one processor; wherein, The memory stores a computer program executable by the at least one processor, and the computer program is executed by the at least one processor so that the at least one processor can execute the beam domain channel state information generation method according to any one of claims 1 to 6.

9. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores computer instructions, and the computer instructions are used to enable a processor to implement the beam-domain channel state information generation method according to any one of claims 1 to 6 when executed.

10. A computer program product, characterized in that The computer program product comprises a computer program, and when the computer program is executed by a processor, the computer program implements the beam domain channel state information generation method according to any one of claims 1-6.

Images