Multi-modal wireless environment prediction method and device, equipment and storage medium

Abstract

The embodiment of the invention provides a multi-mode wireless environment prediction method and device, electronic equipment and a storage medium, and belongs to the technical field of wireless communication. The method comprises the following steps: determining spatial position parameters of a plurality of sampling points on a signal channel corresponding to a receiving position of at least one receiver in a target area; mapping the spatial position parameters corresponding to the sampling points from the same signal channel to obtain spatial signal features, and rendering the spatial signal features to obtain discrete angle power spectrums corresponding to the sampling points; according to a pre-obtained array coefficient and the discrete angle power spectrum corresponding to each sampling point, determining the predicted reference signal power of the signal channel; and performing signal modeling on the target area based on the predicted reference signal power corresponding to all the signal channels to obtain a wireless channel modeling result of the target area. According to the method and the device, the accuracy of the wireless environment prediction result can be improved under the condition of shortage of required measured data.

Description

Technical Field

[0001] The present application relates to the field of wireless communication technology, and in particular to a multimodal wireless environment prediction method, device, electronic device and storage medium. Background Art

[0002] Channels are the medium for information transmission in communication systems, and channels include wired channels and wireless channels. Among them, studying the propagation characteristics of wireless channels and establishing corresponding wireless channel models can help relevant personnel better understand the transmission performance of communication systems, and then predict multimodal wireless environments, so as to assist in optimizing communication systems in the future.

[0003] In the related art, deterministic channel modeling is achieved based on the channel measurement data obtained by a specific measuring instrument and the geographical image of the target area where the communication system is located. However, such a channel modeling method requires a large amount of accurate measured data as support. In the case of a shortage of the required measured data, the accuracy of the wireless environment prediction results in the related art is poor. Summary of the invention

[0004] The main purpose of the embodiments of the present application is to propose a multimodal wireless environment prediction method, device, electronic device and storage medium, aiming to improve the accuracy of wireless environment prediction results when the required measured data is scarce.

[0005] To achieve the above object, a first aspect of an embodiment of the present application proposes a multimodal wireless environment prediction method, the method comprising:

[0006] Acquire an array coefficient of a target base station located in a target area and a receiving position of at least one receiver in the target area;

[0007] Determine at least one signal channel formed between the receiving position and the target area, and determine spatial position parameters of a plurality of sampling points on the signal channel;

[0008] The spatial position parameters are input into a pre-trained wireless channel model, the spatial position parameters corresponding to each sampling point from the same signal channel are mapped to obtain spatial signal features, and the spatial signal features are rendered to obtain the discrete angle power spectrum corresponding to each sampling point;

[0009] Determine the predicted reference signal power of the signal channel according to the array coefficient and the discrete angle power spectrum corresponding to each sampling point;

[0010] Signal modeling of the target area is performed based on the predicted reference signal powers corresponding to all signal channels to obtain a wireless channel modeling result of the target area.

[0011] In some embodiments, the target area includes a plurality of spatial grids;

[0012] Determine at least one signal channel formed between the receiving position and the target area, and determine the spatial position parameters of multiple sampling points on the signal channel, including:

[0013] Obtain the corresponding two-dimensional image of the target area, and determine the camera position and direction when taking the two-dimensional image;

[0014] Selecting one of the plurality of spatial grids as a target spatial grid, determining at least one signal channel formed between the target spatial grid and a receiving position, wherein the signal channel includes a plurality of sampling points;

[0015] According to the camera posture and direction, the three-dimensional coordinate information and observation direction information of each sampling point on each signal channel are determined to obtain the spatial position parameters.

[0016] In some embodiments, the spatial position parameters corresponding to each sampling point from the same signal channel are mapped to obtain spatial signal features, including:

[0017] Obtain point cloud data of the target area;

[0018] Taking point cloud data as signal constraints, the three-dimensional coordinate information and observation direction information are mapped based on a preset encoding function to obtain spatial signal features, where the spatial signal features include discrete radiation signal values ​​and volume density values.

[0019] In some embodiments, rendering processing is performed on the spatial signal features to obtain a discrete angle power spectrum corresponding to each sampling point, including:

[0020] For each sampling point, obtain the interval distance value between the current sampling point and the neighboring sampling point;

[0021] According to the interval distance value and the volume density value, the transmittance value and the opacity value of the sampling point are calculated;

[0022] Determine a signal weight converted from a volume density value based on a product of a transmittance value and an opacity value;

[0023] Based on the product of the discrete radiation signal value and the signal weight, the discrete angle power spectrum of the sampling point is obtained.

[0024] In some embodiments, the wireless channel model is trained by the following steps, comprising:

[0025] Acquire a sample array coefficient of a sample base station located in the sample area and a sample receiving position where at least one sample receiver in the sample area is located;

[0026] Determine at least one sample signal channel formed between the sample receiving position and the sample area, and determine sample spatial position parameters of a plurality of sample sampling points on the sample signal channel;

[0027] The sample space position parameters are input into the initial wireless channel model, the sample space position parameters corresponding to each sample sampling point from the same sample signal channel are mapped to obtain the sample space signal characteristics, and the sample space signal characteristics are rendered to obtain the sample discrete angle power spectrum corresponding to each sample sampling point;

[0028] Determine the sample reference signal power of the sample signal channel according to the sample array coefficient and the sample discrete angle power spectrum corresponding to each sample sampling point;

[0029] A model loss value is obtained according to the sample reference signal power calculation, and the parameters of the initial wireless channel model are adjusted according to the model loss value to obtain a trained wireless channel model.

[0030] In some embodiments, the model loss value includes a wireless signal loss value and a point cloud depth loss value;

[0031] The model loss value is calculated based on the sample reference signal power, including:

[0032] Obtain verification reference signal power and verification point cloud depth value of the sample area;

[0033] Calculate the wireless signal loss value according to the difference between the sample reference signal power and the verification reference signal power;

[0034] Determine a sample point cloud depth value based on a plurality of sample sampling points from the same sample signal channel;

[0035] The point cloud depth loss value is calculated based on the difference between the sample point cloud depth value and the verification point cloud depth value.

[0036] In some embodiments, a model loss value is obtained according to the sample reference signal power calculation, and a parameter of the initial wireless channel model is adjusted according to the model loss value to obtain a trained wireless channel model, including:

[0037] Superimpose the wireless signal loss value and the point cloud depth loss value to obtain the total loss value;

[0038] The parameters of the initial wireless channel model are jointly adjusted according to the total loss value to obtain a trained wireless channel model.

[0039] To achieve the above-mentioned purpose, a second aspect of an embodiment of the present application proposes a multimodal wireless environment prediction device, the device comprising:

[0040] An acquisition module, used to acquire an array coefficient of a target base station located in a target area and a receiving position of at least one receiver in the target area;

[0041] A sampling point determination module, used to determine at least one signal channel formed between the receiving position and the target area, and to determine spatial position parameters of a plurality of sampling points on the signal channel;

[0042] A rendering module is used to input the spatial position parameters into a pre-trained wireless channel model, map the spatial position parameters corresponding to each sampling point from the same signal channel to obtain spatial signal features, and render the spatial signal features to obtain a discrete angle power spectrum corresponding to each sampling point;

[0043] A prediction module, used to determine the predicted reference signal power of the signal channel according to the array coefficient and the discrete angle power spectrum corresponding to each sampling point;

[0044] The target result is used to perform signal modeling on the target area based on the predicted reference signal power corresponding to all signal channels to obtain the wireless channel modeling result of the target area.

[0045] To achieve the above-mentioned purpose, the third aspect of an embodiment of the present application proposes an electronic device, the electronic device comprising a memory and a processor, the memory storing a computer program, and the processor implementing the method of the above-mentioned first aspect when executing the computer program.

[0046] To achieve the above objectives, the fourth aspect of an embodiment of the present application proposes a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the method of the first aspect is implemented.

[0047] The multimodal wireless environment prediction method, device, electronic device and storage medium proposed in the present application obtain the array coefficient of the target base station located in the target area, and the receiving position of at least one receiver in the target area; determine at least one signal channel formed between the receiving position and the target area, and determine the spatial position parameters of multiple sampling points on the signal channel; input the spatial position parameters into a pre-trained wireless channel model, map the corresponding spatial position parameters of each sampling point from the same signal channel to obtain the spatial signal characteristics, and render the spatial signal characteristics to obtain the corresponding discrete angle power spectrum of each sampling point; determine the predicted reference signal power of the signal channel according to the array coefficient and the corresponding discrete angle power spectrum of each sampling point; perform signal modeling on the target area based on the corresponding predicted reference signal power of all signal channels to obtain the wireless channel modeling result of the target area. The present application can improve the accuracy of the channel modeling result when the required measured data is in short supply. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 This is a schematic diagram of an optional application scenario of the multimodal wireless environment prediction device provided in an embodiment of the present application;

[0049] Figure 2 is an optional flowchart of the multimodal wireless environment prediction method provided in an embodiment of the present application;

[0050] Figure 3 yes Figure 2 An optional implementation flowchart of step 102 in ;

[0051] Figure 4 It is an optional multi-layer perceptron schematic diagram of the multimodal wireless environment prediction method provided in the embodiment of the present application;

[0052] Figure 5 yes Figure 2 An optional implementation flowchart of step 103 in ;

[0053] Figure 6 yes Figure 2 Another optional implementation flowchart of step 103 in;

[0054] Figure 7 is another optional flow chart of the multimodal wireless environment prediction method provided in an embodiment of the present application;

[0055] Figure 8 yes Figure 7 An optional implementation flowchart of step 505 in ;

[0056] Fig. 9 yes Figure 7 Another optional implementation flowchart of step 505 in;

[0057] Fig.10 is a schematic diagram of a target area of ​​the multimodal wireless environment prediction method provided in an embodiment of the present application;

[0058] Fig.11 It is a target area data collection schematic diagram of the multi-modal wireless environment prediction method provided in an embodiment of the present application;

[0059] Fig.12 This is an optional flow chart of the multimodal wireless environment prediction device provided in an embodiment of the present application;

[0060] Fig.13 It is a schematic diagram of the hardware structure of the electronic device provided in the embodiment of the present application. DETAILED DESCRIPTION

[0061] In order to make the purpose, technical solution and advantages of the present application more clearly understood, the present application is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present application and are not used to limit the present application.

[0062] It should be noted that, although the functional modules are divided in the device schematic diagram and the logical order is shown in the flowchart, in some cases, the steps shown or described may be performed in a different order than the module division in the device or the order in the flowchart. The terms "first", "second", etc. in the specification, claims and the above drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence.

[0063] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art to which this application belongs. The terms used herein are only for the purpose of describing the embodiments of this application and are not intended to limit this application.

[0064] Channel is the medium for information transmission in communication system, including wired channel and wireless channel. Among them, studying the propagation characteristics of wireless channel and establishing the corresponding wireless channel model can help relevant personnel better understand the transmission performance of communication system, so as to assist in the subsequent optimization of communication system.

[0065] In the related art, deterministic channel modeling is achieved based on the channel measurement data obtained by a specific measuring instrument and the geographical image of the target area where the communication system is located. However, such a channel modeling method requires a large amount of accurate measured data as support. In the case of a shortage of the required measured data, the accuracy of the channel modeling results in the related art is poor.

[0066] Based on this, the embodiments of the present application provide a multimodal wireless environment prediction method, device, electronic device and storage medium, which aim to improve the accuracy of wireless environment prediction results when the required measured data is scarce.

[0067] The multimodal wireless environment prediction method, device, electronic device and storage medium provided in the embodiments of the present application are specifically described through the following embodiments. First of all, it should be noted that the multimodal wireless environment prediction method in the embodiments of the present application relates to the field of wireless communication technology, which can be applied to the terminal, can also be applied to the server side, and can also be software running in the terminal or the server side. In some embodiments, the terminal can be a smart phone, a tablet computer, a laptop computer, a desktop computer, etc.; the server side can be configured as an independent physical server, or a server cluster or distributed system composed of multiple physical servers, and can also be configured as a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms; the software can be an application that implements the wireless environment prediction method, etc., but is not limited to the above forms.

[0068] In addition, the multimodal wireless environment prediction method provided by the embodiment of the present application can also be used in many general or special computer system environments or configurations. For example: personal computers, server computers, handheld devices or portable devices, tablet devices, multi-processor systems, microprocessor-based systems, set-top boxes, programmable consumer electronic devices, network PCs, minicomputers, mainframe computers, distributed computing environments including any of the above systems or devices, etc. The present application can be described in the general context of computer-executable instructions executed by a computer, such as program modules. Generally, program modules include routines, programs, objects, components, data structures, etc. that perform specific tasks or implement specific abstract data types. The present application can also be practiced in distributed computing environments, in which tasks are performed by remote processing devices connected through a communication network. In a distributed computing environment, program modules can be located in local and remote computer storage media including storage devices.

[0069] It should also be noted that in the embodiments of the present application, when it comes to information related to user characteristics such as user basic information or user identity, the user's permission or consent will be obtained first, and the collection, use and processing of these data will comply with relevant laws, regulations and standards. In addition, when the embodiments of the present application need to obtain sensitive personal information of the user, the user's separate permission or separate consent will be obtained first. After the user's separate permission or separate consent is clearly obtained, the necessary data for the normal operation of the embodiments of the present application will be obtained. For example, the multimodal wireless environment prediction device of the embodiment of the present application obtains the consent of the relevant personnel in the target area before obtaining the wireless channel modeling result of the target area using the multimodal wireless environment prediction method proposed in the embodiment of the present application, otherwise the array coefficient of the target base station located in the target area and other data used to obtain the wireless channel modeling result cannot be obtained. That is, all relevant data obtained by the embodiments of the present application are authorized data after obtaining consent, and will not be repeated here.

[0070] Next, the application scenario of the multimodal wireless environment prediction device in the embodiment of the present application is described, such as Figure 1 As shown, Figure 1 The present invention provides a schematic diagram of an optional application scenario of a multi-modal wireless environment prediction device. A target base station (BS) is provided in the target area, where a downlink multi-input multi-output wireless system (Multi-User Multiple-Input Multiple-Output, MIMO) is deployed. The target base station is equipped with a set of N T antenna array, and the target base station uses a codebook The reference signal is sent by using different beams in the target area to scan the entire target area, thereby achieving the propagation of wireless signals in the target area. However, the target area is usually distributed with many and cluttered buildings, such as Figure 1 In the target area, in addition to the target base station, there are also building complexes a and building complexes b. The buildings in the building complexes will block the propagation path of the wireless signal, resulting in signal attenuation.

[0071] In response to this situation, relevant technologies need to complete wireless channel modeling based on a large amount of accurate measured data. The measured data obtained usually include but are not limited to accurate geographic images and channel measurement data obtained by high-precision measuring instruments to measure the target area. It is understandable that once the relevant technology cannot obtain the required measured data, the wireless channel modeling work cannot be carried out.

[0072] In comparison, the multimodal wireless environment prediction device in the embodiment of the present application can process a small amount of spatial position parameters based on a pre-trained wireless channel model without a large amount of measured data as support, thereby obtaining a predicted reference signal power. In this way, the dependence of wireless channel modeling on actual environmental information is reduced, while ensuring the accuracy of the wireless channel modeling results corresponding to the output target area.

[0073] After understanding the example application scenarios of the multimodal wireless environment prediction device proposed in the present application, the multimodal wireless environment prediction method proposed in the embodiment of the present application is described next.

[0074] In the embodiment of the present application, the multimodal wireless environment prediction device will be described from the perspective of the multimodal wireless environment prediction device, wherein the multimodal wireless environment prediction device can be integrated into a computer device, such as a server. Figure 2 As shown, Figure 2 is an optional flowchart of the multimodal wireless environment prediction method provided in an embodiment of the present application. Figure 2 The method may include but is not limited to the following steps 101 to 105. When the multi-modal wireless environment prediction device executes the multi-modal wireless environment prediction method, the specific process is as follows. It should be noted that the present embodiment Figure 2 The order of step 101 to step 105 is not specifically limited, and the order of steps can be adjusted or some steps can be reduced or added according to actual needs.

[0075] Step 101: Obtain array coefficients of a target base station located in a target area and a receiving position of at least one receiver in the target area.

[0076] The following is a detailed description of step 101.

[0077] In some embodiments, the target area is divided into L spatial grids, and the receiver can be set in any spatial grid of the target area. Specifically, the receiver (User Equipment, UE) of the lth spatial grid can measure the channel impulse response at the tth time, and report the beam-based reference signal power (Reference Signal Receiving Power, RSRP) information for realizing wireless channel modeling such as geo-statistical channel modeling tasks, which includes the incident angle from the target base station to the receiver, channel gain, and other key parameters used to reflect the strength of the wireless signal in the target area.

[0078] Among them, the receiver refers to an electronic device used to receive analog or digital AM, FM, or digital broadcast signals from radio frequencies and convert them into corresponding audio, video or digital signals. Exemplarily, the receiver in the embodiment of the present application includes but is not limited to mobile phones, computers, tablets, etc.; and, in order to help improve the accuracy of the final wireless environment prediction results, the embodiment of the present application can also obtain multiple receiving positions of the same receiver, for example, the receiver can be held and moved along a preset trajectory in the target area; or, the embodiment of the present application can set receivers at multiple locations in the target area and obtain the receiving positions of the receivers at each location; it can be set according to actual conditions, and the embodiment of the present application does not limit this.

[0079] The target base station may be adaptively set according to different target areas to which it is applied, and the embodiments of the present application do not limit this. The following are some specific examples:

[0080] Macro Site: Macro base stations are usually used in wide outdoor target areas. They have large power and coverage and can provide wireless coverage over a large area. They are usually installed on towers or high-rise buildings and are suitable for open areas such as cities, suburbs, and villages.

[0081] Micro Site: Micro Site is small in size and is usually used in indoor or densely populated target areas, such as commercial centers and office buildings. Its coverage is relatively small, but it can provide higher signal quality and capacity.

[0082] Pico Site: Pico Site is smaller than micro base station and is usually used to provide coverage in local areas, such as shopping malls, hotels, stadiums, etc.

[0083] Femto Site: Femto Site is mainly used for home users. Its size is similar to that of a home router. It can provide wireless coverage inside the home to improve the indoor signal quality and data rate of home users.

[0084] Furthermore, in order to determine the predicted reference signal power in the target area, so as to obtain the corresponding wireless channel modeling result of the target area according to the predicted reference signal power, it is first necessary to obtain the array coefficient of the target base station located in the target area, and the steps for obtaining the array coefficient are as follows:

[0085] ① Obtain the antenna steering vector of the target base station;

[0086] ② Determine the array coefficient of the target base station based on the antenna steering vector and the preset beam vector.

[0087] Among them, an antenna array is deployed in the target base station. The antenna array refers to a large number of antennas integrated in the target base station. These antennas work together to achieve specific wireless communication functions, especially in the fifth generation mobile communication technology (5th Generation Mobile Communication Technology, referred to as 5G). This antenna array is usually called a large-scale antenna array. The placement of each antenna in the antenna array, such as the direction of the antenna, has an impact on the signal transmitted by the target base station. Based on this, the wireless channel modeling result finally output by the embodiment of the present application helps optimize the antenna array (such as adjusting the antenna direction) by feeding back the signal transmission situation in the target area, so that the wireless signal transmitted by the target base station can be better propagated in the target area.

[0088] Furthermore, the angle space in the target area is used to define the directional range of the target base station antenna. The three-dimensional angle space of the target base station is divided, and the free space is evenly discretized into NV pitch angles and NH azimuth angles to obtain N spatial angles. N = NV*NH, N is the number of spatial angles, NV is the number of pitch angles, and NH is the number of azimuth angles. The tilt angle of incidence (tilt AoD) of the i-th pitch angle is θ i , the azimuth angle of incidence (azimuth AoD) of the jth azimuth is φ j It should be noted that the number of actual propagation paths from the target base station to the receiver is less than N. Antenna array steering vectors are used to describe the radiation or reception capabilities of antenna arrays in different directions. Antenna steering vectors are used to describe the radiation or reception capabilities of antenna arrays in different directions, including multiple steering vectors in different directions. The antenna steering vector is a steering matrix, denoted as S, and the steering vector in the nth direction (spatial angle) is denoted as s(θ n ,φ n ), so the antenna steering vector S is expressed as follows <1> :

[0089] S=[s(θ1,φ1),…,s(θ N ,φ N )] T <1>

[0090] Wherein, T represents transposition. In the formulas listed in the embodiments of the present application, the same parameters have the same meanings, and the repeated parts are not repeated.

[0091] Furthermore, the preset beam vector includes beamforming vectors of M beams, and the preset beamforming vector is expressed as W = [w1, ..., w M], where represents the beamforming vector of the i-th beam. The measurement matrix is ​​calculated according to the antenna array steering vector S and the preset beamforming vector W, that is, the array coefficient of the target base station is determined. The specific calculation method of the array coefficient is as follows: <2> As shown:

[0092] Φ=(|W H S| 2 ) <2>

[0093] Where Φ is the array coefficient of the target base station; H represents the conjugate transpose operation; |·| 2 It is taking the absolute value of the element and squaring it.

[0094] Step 102: determine at least one signal channel formed between the receiving position and the target area, and determine the spatial position parameters of multiple sampling points on the signal channel.

[0095] The following is a detailed description of step 102.

[0096] In some embodiments, since wireless signals are affected by factors such as reflection, refraction and scattering during propagation, multiple signal channels are usually formed between the receiving position of the receiver and the target area; at least one signal channel is selected from the multiple signal channels, and each signal channel has multiple sampling points, and the spatial position parameters of each sampling point are determined, so that the spatial position parameters can be input into a pre-trained wireless channel model, and the corresponding discrete angle power spectrum of each sampling point can be obtained, thereby helping to determine the predicted reference signal power of the signal channel where the sampling point is located.

[0097] The following is a detailed description of how to determine the spatial position parameters of each sampling point.

[0098] In some embodiments, Figure 3 As shown, Figure 3 yes Figure 2 An optional implementation flowchart of step 102 in the embodiment of the present invention is to determine at least one signal channel formed between the receiving position and the target area, and determine the spatial position parameters of multiple sampling points on the signal channel, including the following steps 201 to 203:

[0099] Step 201 , obtaining a two-dimensional image corresponding to the target area, and determining the camera position and direction when shooting the two-dimensional image.

[0100] Step 202: select one of the multiple space grids as a target space grid, determine at least one signal channel formed between the target space grid and the receiving position, and the signal channel includes multiple sampling points.

[0101] Step 203, according to the camera posture and camera direction, determine the three-dimensional coordinate information and observation direction information of each sampling point on each signal channel to obtain the spatial position parameters.

[0102] Steps 201 to 203 are described in detail below.

[0103] In some embodiments, the target area includes multiple spatial grids. The receiver can form a signal channel with any spatial grid, so one of the multiple spatial grids is selected as the target spatial grid, and at least one signal channel formed between the receiving position of the receiver and the target spatial grid, as well as multiple sampling points on the signal channel are determined. The sampling points can be randomly selected, or can be obtained by sampling at the same distance on the signal channel. The embodiment of the present application does not limit the selection method of the sampling points.

[0104] Furthermore, the spatial position parameters of each sampling point are determined based on Neural Radiance Fields (NeRF). Among them, Neural Radiance Fields are used to generate high-quality 3D reconstruction models. Specifically, Neural Radiance Fields uses deep learning technology to extract the geometric shape and texture information of the target object from images of at least one perspective, and then uses these several shapes and texture information to generate a continuous 3D radiation field, so that a highly realistic 3D model can be presented at any angle and distance. It is understandable that the 3D model includes physical obstacles such as buildings, metal objects (shielding objects), and reflectors.

[0105] Furthermore, in order to generate the corresponding three-dimensional model in the target area based on the sampling points, it is first necessary to determine the spatial position parameters corresponding to each sampling point that meets the neural radiation field input. Specifically, at least one two-dimensional image of the target area and the camera parameters when shooting the two-dimensional image are obtained. The camera parameters include but are not limited to camera posture, camera direction, etc.; combined with the camera parameters, while ensuring that the sampling points in the two-dimensional images taken from different perspectives are consistent in the three-dimensional space, the spatial position parameters of the sampling points are obtained. The spatial position parameters include three-dimensional coordinate information and observation direction information. The three-dimensional coordinate information is denoted as (x, y, z), and the observation direction information is denoted as The spatial location parameters are used as input to the pre-trained wireless channel model of the subsequent network.

[0106] Step 103, input the spatial position parameters into a pre-trained wireless channel model, map the spatial position parameters corresponding to each sampling point from the same signal channel to obtain spatial signal features, and render the spatial signal features to obtain a discrete angle power spectrum corresponding to each sampling point.

[0107] The following is a detailed description of step 103.

[0108] In some embodiments, the pre-trained wireless channel model maps the input spatial position parameters to obtain the spatial signal features corresponding to the spatial position parameters of each sampling point on each signal channel. Specifically, the neural radiation field set in the wireless channel model is set with an eight-layer multilayer perceptron (MLP), and the hidden layer dimension of the MLP is set as follows Figure 4 As shown, Figure 4 It is an optional multi-layer perceptron schematic diagram of the multimodal wireless environment prediction method provided in the embodiment of the present application. It should be noted that the number of network layers of the MLP can be set according to actual conditions, and the embodiment of the present application does not limit this.

[0109] The processing of step 103 involves the extraction of multimodal features in signal processing, specifically including but not limited to power spectrum, phase spectrum, power spectrum density, etc., which are common multimodal features in signal processing.

[0110] It is understandable that, without the need for measured channel data as support, the wireless channel model maps the spatial position parameters of the sampling points that can be easily obtained to obtain the spatial signal characteristics of the response signal strength of each sampling point on each signal channel. Compared with the traditional multimodal wireless environment prediction method, the multimodal wireless environment prediction method proposed in the embodiment of the present application does not require detailed physical modeling and parameter measurement of the channel environment in the target area, thereby reducing the computational resource cost of determining the wireless channel modeling results in the target area, while also based on the high-accuracy three-dimensional model restoration of the neural radiation field, by improving the accuracy of the wireless channel modeling results, thereby improving the accuracy of the wireless environment prediction.

[0111] In some embodiments, Figure 5 As shown, Figure 5 yes Figure 2 In an optional implementation flowchart of step 103 in the embodiment, the spatial position parameters corresponding to each sampling point from the same signal channel are mapped to obtain spatial signal features, including the following steps 301 to 302:

[0112] Step 301, obtaining point cloud data of the target area.

[0113] Step 302, using the point cloud data as a signal constraint condition, mapping the three-dimensional coordinate information and the observation direction information based on a preset encoding function to obtain spatial signal features, wherein the spatial signal features include discrete radiation signal values ​​and volume density values.

[0114] Steps 301 to 302 are described in detail below.

[0115] In some embodiments, point cloud data of the target area is obtained so that the neural radiation field can map the spatial position parameters based on the point cloud data as a signal constraint condition; wherein, point cloud data refers to a set of vectors in a three-dimensional coordinate system, each point contains three-dimensional coordinates, and some may contain color information (RGB) or reflection intensity information (Intensity); the mapping processing based on the point cloud data as a signal constraint condition can further ensure the accuracy of the input three-dimensional coordinate information and observation direction information, and then can process and obtain high-accuracy spatial signal features based on accurate spatial position parameters.

[0116] Among them, point cloud data is often used in combination with other types of sensor data (such as images, depth maps, radar signals, etc.) to provide a more comprehensive multimodal environment understanding, thereby providing precise spatial location information, as well as rich color and texture information.

[0117] Furthermore, the spatial signal characteristics include the volume density value σ of the sampling point and the discrete radiation signal value, and the discrete radiation signal value further includes the radiation signal amplitude A and the signal phase ψ. The spatial signal characteristics are denoted as (σ, A, ψ).

[0118] Furthermore, based on the point cloud data as the signal constraint, the wireless channel model uses the multi-layer perceptron deployed in it to perform position encoding on each input spatial position parameter. L (·) operation, γ L (·) is a transition from R to R 2M The mapping is formally expressed as follows: <3> The encoding function shown maps the input spatial position parameters:

[0119] Y L (p) = (sin(2 0 πp), cos(2 0 πp),...,sin(2 M-1 πp), cos(2 M-1 πp)) <3>

[0120] Furthermore, in practical applications, wireless channel models usually output real and imaginary data (I, Q) related to the radiated signal amplitude A and signal phase m, rather than directly outputting the radiated signal amplitude A and signal phase m. The reason is that the signal phase is modulo 2π and is not differentiable at some points. Based on this, the following formula is used <4> and <5> Convert (I, Q) into radiation signal amplitude A and signal phase mountain:

[0121]

[0122] In some embodiments, Figure 6 As shown, Figure 6 yes Figure 2 Another optional implementation flowchart of step 103 in the above embodiment renders the spatial signal features to obtain a discrete angle power spectrum corresponding to each sampling point, including the following steps 401 to 404:

[0123] Step 401: for each sampling point, obtain the interval distance value between the current sampling point and the neighboring sampling point.

[0124] Step 402: Calculate the transmittance value and opacity value of the sampling point according to the interval distance value and the volume density value.

[0125] Step 403: Determine the signal weight converted from the volume density value based on the product of the transmittance value and the opacity value.

[0126] Step 404: Obtain a discrete angle power spectrum of the sampling point based on the product of the discrete radiation signal value and the signal weight.

[0127] Steps 401 to 404 are described in detail below.

[0128] In some embodiments, by the formula <6> and <7> Calculate the transmittance value T of the sampling point rp and opacity value α rp :

[0129]

[0130] Where rp represents the pth sampling point on the rth signal channel; σ p is the volume density value of the current sampling point p calculated by the neural radiation field, δ p is the interval distance between the current sampling point p and the neighbor sampling point (p+1); exp(·) represents the natural exponential function.

[0131] Furthermore, by the following formula <8> Calculate the signal weight w rp :

[0132] W rp =T rp α rp <8>

[0133] Furthermore, by the following formula <9> and <10> , based on the product of the discrete radiation signal value and the signal weight, the discrete angle power spectrum S of the sampling point is calculated r :

[0134]

[0135]

[0136] Among them, e is a natural constant.

[0137] Step 104: Determine the predicted reference signal power of the signal channel according to the array coefficient and the discrete angle power spectrum corresponding to each sampling point.

[0138] Step 104 is described in detail below.

[0139] In some embodiments, by the formula <11> and <12> Determine the predicted reference signal power RSRP of the signal channel:

[0140] Δ(r)=|S r | 2 <11>

[0141] RSRP≈φΔ(r) <12>

[0142] Among them, based on the discrete angular power spectra of multiple sampling points from the same signal channel, the channel angular power spectrum Δ(r) of the corresponding signal channel is determined; and then according to the product of the array coefficient and the channel angular power spectrum, the predicted reference signal power RSRP of the signal channel where the multiple sampling points are located is determined.

[0143] Step 105 , performing signal modeling on the target area based on the predicted reference signal powers corresponding to all signal channels, and obtaining a wireless channel modeling result of the target area.

[0144] The following is a detailed description of step 105.

[0145] In some embodiments, after obtaining the corresponding predicted reference signal power of each signal channel, a signal modeling operation is performed on the target area based on multiple predicted reference signal powers to obtain a wireless channel modeling result of the target area. It can be understood that the predicted reference signal power better reflects the signal strength of each signal channel, and thus a wireless channel modeling result with higher accuracy can be obtained to achieve multi-modal wireless environment prediction. Specifically, through the following formula <13> and <14> Determine the predicted reference signal power RSRP of the target area final :

[0146] X=[Δ(r1),…,Δ(r N )] T <13>

[0147] RSRP final ≈φX <14>

[0148] In some embodiments, Figure 7 As shown, Figure 7This is another optional flow chart of the multimodal wireless environment prediction method provided in an embodiment of the present application. The wireless channel model is obtained by training through the following steps 501 to 505:

[0149] Step 501: Obtain sample array coefficients of a sample base station located in a sample area, and a sample receiving position of at least one sample receiver in the sample area.

[0150] Step 502: determine at least one sample signal channel formed between the sample receiving position and the sample area, and determine sample space position parameters of a plurality of sample sampling points on the sample signal channel.

[0151] Step 503, input the sample space position parameters into the initial wireless channel model, map the sample space position parameters corresponding to each sample sampling point from the same sample signal channel to obtain the sample space signal characteristics, and render the sample space signal characteristics to obtain the sample discrete angle power spectrum corresponding to each sample sampling point.

[0152] Step 504: Determine the sample reference signal power of the sample signal channel according to the sample array coefficient and the sample discrete angle power spectrum corresponding to each sample sampling point.

[0153] Step 505 , obtain a model loss value based on the sample reference signal power calculation, and adjust the parameters of the initial wireless channel model based on the model loss value to obtain a trained wireless channel model.

[0154] Steps 501 to 505 are described below.

[0155] In some embodiments, before the wireless channel model is formally put into practical application, the wireless channel model needs to be trained to improve the ability of the wireless channel model to process input data and the accuracy of the output results. The acquired relevant sample data is input into the initial wireless channel model, and the specific implementation method of the initial wireless channel model processing the input sample data described in steps 501 to 504 is similar to steps 101 to 104, which will not be repeated here.

[0156] Further, the model loss value is calculated based on the sample reference power output by the initial wireless channel model, and the parameters of the initial wireless channel model are adjusted based on the model loss value to obtain the trained wireless channel model. The adjusted parameters include but are not limited to path loss parameters, shadow fading parameters, multipath effect parameters, Doppler effect parameters, and noise parameters, etc. The specific adjusted parameters can be set according to actual conditions, and the embodiments of the present application do not limit this.

[0157] In some embodiments, Figure 8 As shown, Figure 8 yes Figure 7 An optional implementation flowchart of step 505 in the embodiment of the present invention is to calculate the model loss value according to the sample reference signal power, including the following steps 601 to 604:

[0158] Step 601: Obtain verification reference signal power and verification point cloud depth value of the sample area.

[0159] Step 602: Calculate the wireless signal loss value according to the difference between the sample reference signal power and the verification reference signal power.

[0160] Step 603: Determine a sample point cloud depth value based on multiple sample sampling points from the same sample signal channel.

[0161] Step 604: Calculate the point cloud depth loss value according to the difference between the sample point cloud depth value and the verification point cloud depth value.

[0162] Steps 601 to 604 are described in detail below.

[0163] In some embodiments, the model loss value includes a wireless signal loss value and a point cloud depth loss value. The parameters of the initial wireless channel model are adjusted by the calculated wireless signal loss value and the point cloud depth loss value to obtain a trained wireless channel model.

[0164] Specifically, through the following formula <15> Calculate the wireless signal loss value L signal :

[0165] L signal =||RSRP test -φX||2 <15>

[0166] Among them, RSRP test To verify the reference signal power, φX is the sample reference signal power; ||RSRP test -φX||2 is the I2 norm function.

[0167] Furthermore, by the following formula <16> and <17> Calculate the point cloud depth loss value L depth :

[0168]

[0169] Among them, z is the depth value of the sample point cloud, To verify the point cloud depth value, p n and p f Respectively represent the nearest and farthest sampling points on a signal channel; w rprepresents the signal weight at the pth sampling point on the corresponding signal channel r, δ p-1 represents the distance between the p-th sampling point and the p-1-th sampling point; D represents the set of all signal channels; E is used to characterize the expected calculation.

[0170] In some embodiments, Fig. 9 As shown, Fig. 9 yes Figure 7 Another optional implementation flowchart of step 505 in the embodiment of the present invention is to calculate a model loss value according to the sample reference signal power, and adjust the parameters of the initial wireless channel model according to the model loss value to obtain a trained wireless channel model, including the following steps 701 to 702:

[0171] Step 701, superimpose the wireless signal loss value and the point cloud depth loss value to obtain a total loss value.

[0172] Step 702: jointly adjust the parameters of the initial wireless channel model according to the total loss value to obtain a trained wireless channel model.

[0173] Steps 701 to 702 are described in detail below.

[0174] In some embodiments, by the formula <18> The total loss value L is calculated:

[0175] L=L signal +L depth <18>

[0176] Among them, in addition to training the initial wireless channel model by the wireless signal loss value and the point cloud depth loss value respectively, the total loss value can be obtained by superimposing the wireless signal loss value and the point cloud depth loss value, and then the parameters of the initial wireless channel model are jointly adjusted according to the total loss value to obtain the trained wireless channel model. In order to improve the output accuracy of the trained wireless channel model in the embodiment of the present application while reducing the waste of energy resources.

[0177] Furthermore, the performance of the trained wireless channel model is verified:

[0178] like Fig.10 and Fig.11 As shown, Fig.10 is a schematic diagram of a target area of ​​the multimodal wireless environment prediction method provided in an embodiment of the present application, Fig.11This is a target area data collection schematic diagram of the multimodal wireless environment prediction method provided by the embodiment of the present application. For a certain target area, Ray Tracing is used to generate channels, and the horizontal angle and downtilt angle of the base station mechanical angle are adjusted by 20 degrees and 2 degrees respectively, and 3 decibels (dB) of noise are added. The traditional NeRF2, NeWRF, and LSCM method control groups are selected, and the test comparison results are as follows:

[0179]

[0180] It should be noted that when the four methods are used to process the same data set respectively, and the mean absolute error MAE and Top8-MAE parameters are considered (the smaller the better the performance), it can be found that the model proposed in the embodiment of the present application performs better. This shows that the model proposed in the embodiment of the present application not only implements the multimodal wireless environment prediction method, but also solves the problems of complex wireless signal propagation and data shortage, thereby achieving better performance.

[0181] like Fig.12 As shown, Fig.12 This is an optional flow chart of a multimodal wireless environment prediction device provided in an embodiment of the present application. The multimodal wireless environment prediction device includes the following modules 801 to 805:

[0182] The acquisition module 801 is used to acquire the array coefficient of the target base station located in the target area and the receiving position of at least one receiver in the target area.

[0183] The sampling point determination module 802 is used to determine at least one signal channel formed between the receiving position and the target area, and to determine the spatial position parameters of multiple sampling points on the signal channel.

[0184] The rendering module 803 is used to input the spatial position parameters into a pre-trained wireless channel model, map the spatial position parameters corresponding to each sampling point from the same signal channel to obtain spatial signal features, and render the spatial signal features to obtain a discrete angle power spectrum corresponding to each sampling point.

[0185] The prediction module 804 is used to determine the predicted reference signal power of the signal channel according to the array coefficient and the discrete angle power spectrum corresponding to each sampling point.

[0186] The target result 805 is used to perform signal modeling on the target area based on the predicted reference signal powers corresponding to all signal channels to obtain the wireless channel modeling result of the target area.

[0187] The multimodal wireless environment prediction method, device, electronic device and storage medium proposed in the present application obtain the array coefficient of the target base station located in the target area, and the receiving position of at least one receiver in the target area; determine at least one signal channel formed between the receiving position and the target area, and determine the spatial position parameters of multiple sampling points on the signal channel; input the spatial position parameters into a pre-trained wireless channel model, map the corresponding spatial position parameters of each sampling point from the same signal channel to obtain the spatial signal characteristics, and render the spatial signal characteristics to obtain the corresponding discrete angle power spectrum of each sampling point; determine the predicted reference signal power of the signal channel according to the array coefficient and the corresponding discrete angle power spectrum of each sampling point; perform signal modeling on the target area based on the corresponding predicted reference signal power of all signal channels to obtain the wireless channel modeling result of the target area. The present application can improve the accuracy of the channel modeling result when the required measured data is in short supply.

[0188] The specific implementation of the multimodal wireless environment prediction device is substantially the same as the specific implementation of the multimodal wireless environment prediction method described above, and will not be described in detail herein.

[0189] The embodiment of the present application also provides an electronic device, the electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the above multi-modal wireless environment prediction method when executing the computer program. The electronic device can be any smart terminal including a tablet computer, a car computer, etc.

[0190] like Fig.13 As shown, Fig.13 : is a schematic diagram of the hardware structure of an electronic device provided in an embodiment of the present application, the electronic device includes:

[0191] The processor 901 may be implemented by a general-purpose CPU (Central Processing Unit), a microprocessor, an application-specific integrated circuit (ASIC), or one or more integrated circuits, and is used to execute relevant programs to implement the technical solutions provided in the embodiments of the present application;

[0192] The memory 902 can be implemented in the form of a read-only memory (ROM), a static storage device, a dynamic storage device, or a random access memory (RAM). The memory 902 can store an operating system and other applications. When the technical solution provided in the embodiment of this specification is implemented by software or firmware, the relevant program code is stored in the memory 902, and the processor 901 calls and executes the multi-modal wireless environment prediction method of the embodiment of the present application;

[0193] Input / output interface 903, used to implement information input and output;

[0194] Communication interface 904, used to realize communication interaction between the device and other devices, which can be realized by wired mode (such as USB, network cable, etc.) or wireless mode (such as mobile network, WIFI, Bluetooth, etc.);

[0195] A bus 905 that transmits information between various components of the device (e.g., the processor 901, the memory 902, the input / output interface 903, and the communication interface 904);

[0196] The processor 901 , the memory 902 , the input / output interface 903 and the communication interface 904 are connected to each other in communication within the device via a bus 905 .

[0197] An embodiment of the present application further provides a computer-readable storage medium, which stores a computer program. When the computer program is executed by a processor, the multi-modal wireless environment prediction method is implemented.

[0198] The memory, as a non-transient computer-readable storage medium, can be used to store non-transient software programs and non-transient computer executable programs. In addition, the memory may include a high-speed random access memory, and may also include a non-transient memory, such as at least one disk storage device, a flash memory device, or other non-transient solid-state storage device. In some embodiments, the memory may optionally include a memory remotely disposed relative to the processor, and these remote memories may be connected to the processor via a network. Examples of the above-mentioned network include, but are not limited to, the Internet, an intranet, a local area network, a mobile communication network, and combinations thereof.

[0199] The embodiments described in the embodiments of the present application are intended to more clearly illustrate the technical solutions of the embodiments of the present application and do not constitute a limitation on the technical solutions provided in the embodiments of the present application. Those skilled in the art will appreciate that with the evolution of technology and the emergence of new application scenarios, the technical solutions provided in the embodiments of the present application are also applicable to similar technical problems.

[0200] Those skilled in the art will appreciate that the technical solutions shown in the figures do not constitute a limitation on the embodiments of the present application, and may include more or fewer steps than shown in the figures, or a combination of certain steps, or different steps.

[0201] The device embodiments described above are merely illustrative, and the units described as separate components may or may not be physically separated, that is, they may be located in one place or distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0202] Those skilled in the art will appreciate that all or some of the steps in the methods disclosed above, and the functional modules / units in the systems and devices may be implemented as software, firmware, hardware, or a suitable combination thereof.

[0203] The terms "first", "second", "third", "fourth", etc. (if any) in the specification of the present application and the above-mentioned drawings are used to distinguish similar objects, and are not necessarily used to describe a specific order or sequence. It should be understood that the data used in this way can be interchangeable where appropriate, so that the embodiments of the present application described herein can be implemented in an order other than those illustrated or described herein. In addition, the terms "including" and "having" and any of their variations are intended to cover non-exclusive inclusions, for example, a process, method, system, product or device comprising a series of steps or units is not necessarily limited to those steps or units clearly listed, but may include other steps or units that are not clearly listed or inherent to these processes, methods, products or devices.

[0204] It should be understood that in the present application, "at least one (item)" means one or more, and "plurality" means two or more. "And / or" is used to describe the association relationship of associated objects, indicating that three relationships may exist. For example, "A and / or B" can mean: only A exists, only B exists, and A and B exist at the same time, where A and B can be singular or plural. The character " / " generally indicates that the objects associated before and after are in an "or" relationship. "At least one of the following" or similar expressions refers to any combination of these items, including any combination of single or plural items. For example, at least one of a, b or c can mean: a, b, c, "a and b", "a and c", "b and c", or "a and b and c", where a, b, c can be single or multiple.

[0205] In the several embodiments provided in the present application, it should be understood that the disclosed devices and methods can be implemented in other ways. For example, the device embodiments described above are only schematic. For example, the division of the above units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0206] The units described above as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0207] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit. The above-mentioned integrated unit may be implemented in the form of hardware or in the form of software functional units.

[0208] If the integrated unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or all or part of the technical solution can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including multiple instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods of various embodiments of the present application. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (Read-Only Memory, referred to as ROM), random access memory (Random Access Memory, referred to as RAM), disk or optical disk and other media that can store programs.

[0209] The preferred embodiments of the present invention are described above with reference to the accompanying drawings, but the scope of the rights of the present invention is not limited thereto. Any modification, equivalent substitution and improvement made by a person skilled in the art without departing from the scope and essence of the present invention should be within the scope of the rights of the present invention.

Claims

1. A multimodal wireless environment prediction method, characterized in that: The method comprises: Acquire an array coefficient of a target base station located in a target area and a receiving position of at least one receiver in the target area; Determine at least one signal channel formed between the receiving position and the target area, and determine spatial position parameters of a plurality of sampling points on the signal channel; Inputting the spatial position parameters into a pre-trained wireless channel model, mapping the spatial position parameters corresponding to each sampling point from the same signal channel to obtain spatial signal features, and rendering the spatial signal features to obtain discrete angle power spectra corresponding to each sampling point; Determine the predicted reference signal power of the signal channel according to the array coefficient and the discrete angle power spectrum corresponding to each sampling point; Signal modeling is performed on the target area based on the predicted reference signal powers corresponding to all the signal channels to obtain a wireless channel modeling result of the target area.

2. The method according to claim 1, characterized in that The target area includes a plurality of spatial grids; The determining of at least one signal channel formed between the receiving position and the target area, and determining spatial position parameters of a plurality of sampling points on the signal channel, comprises: Acquire a two-dimensional image corresponding to the target area, and determine the camera position and direction when taking the two-dimensional image; Selecting one of the plurality of spatial grids as a target spatial grid, determining at least one signal channel formed between the target spatial grid and the receiving position, wherein the signal channel includes a plurality of sampling points; According to the camera posture and the camera direction, the three-dimensional coordinate information and the observation direction information of each sampling point on each signal channel are determined to obtain the spatial position parameters.

3. The method according to claim 2, characterized in that The spatial position parameters corresponding to the sampling points from the same signal channel are mapped to obtain spatial signal features, including: Acquire point cloud data of the target area; The point cloud data is used as a signal constraint condition, and the three-dimensional coordinate information and the observation direction information are mapped based on a preset encoding function to obtain a spatial signal feature, wherein the spatial signal feature includes a discrete radiation signal value and a volume density value.

4. The method according to claim 3, characterized in that The rendering process of the spatial signal feature to obtain a discrete angle power spectrum corresponding to each sampling point includes: For each of the sampling points, obtaining the interval distance value between the current sampling point and a neighboring sampling point; Calculating the transmittance value and the opacity value of the sampling point according to the interval distance value and the volume density value; Determining a signal weight converted from the volume density value based on a product of the transmittance value and the opacity value; The discrete angle power spectrum of the sampling point is obtained based on the product of the discrete radiation signal value and the signal weight.

5. The method according to claim 1, characterized in that The wireless channel model is obtained by training through the following steps, which include: Acquire sample array coefficients of a sample base station located in a sample area, and a sample receiving position of at least one sample receiver in the sample area; Determine at least one sample signal channel formed between the sample receiving position and the sample area, and determine sample space position parameters of a plurality of sample sampling points on the sample signal channel; Inputting the sample space position parameters into the initial wireless channel model, mapping the sample space position parameters corresponding to each sample sampling point from the same sample signal channel to obtain sample space signal features, and rendering the sample space signal features to obtain sample discrete angle power spectra corresponding to each sample sampling point; Determine the sample reference signal power of the sample signal channel according to the sample array coefficient and the sample discrete angle power spectrum corresponding to each sample sampling point; A model loss value is calculated according to the sample reference signal power, and parameters of the initial wireless channel model are adjusted according to the model loss value to obtain the trained wireless channel model.

6. The method according to claim 5, characterized in that The model loss value includes a wireless signal loss value and a point cloud depth loss value; The calculating the model loss value according to the sample reference signal power includes: Acquire the verification reference signal power and the verification point cloud depth value of the sample area; Calculating the wireless signal loss value according to the difference between the sample reference signal power and the verification reference signal power; Determine a sample point cloud depth value based on a plurality of sample sampling points from the same sample signal channel; The point cloud depth loss value is calculated based on the difference between the sample point cloud depth value and the verification point cloud depth value.

7. The method according to claim 6, characterized in that The step of calculating a model loss value according to the sample reference signal power, and adjusting parameters of the initial wireless channel model according to the model loss value to obtain the trained wireless channel model includes: Superimposing the wireless signal loss value and the point cloud depth loss value to obtain a total loss value; The parameters of the initial wireless channel model are jointly adjusted according to the total loss value to obtain the trained wireless channel model.

8. A multi-modal wireless environment prediction device, characterized in that: The device comprises: An acquisition module, used to acquire array coefficients of a target base station located in a target area and a receiving position of at least one receiver in the target area; A sampling point determination module, used to determine at least one signal channel formed between the receiving position and the target area, and to determine spatial position parameters of a plurality of sampling points on the signal channel; A rendering module, used to input the spatial position parameters into a pre-trained wireless channel model, map the spatial position parameters corresponding to each sampling point from the same signal channel to obtain spatial signal features, and render the spatial signal features to obtain a discrete angle power spectrum corresponding to each sampling point; A prediction module, used to determine the predicted reference signal power of the signal channel according to the array coefficient and the discrete angle power spectrum corresponding to each sampling point; The target result is used to perform signal modeling on the target area based on the predicted reference signal powers corresponding to all the signal channels to obtain a wireless channel modeling result of the target area.

9. An electronic device, characterized in that: The electronic device includes a memory and a processor, the memory stores a computer program, and the processor implements the multimodal wireless environment prediction method according to any one of claims 1 to 7 when executing the computer program.

10. A computer-readable storage medium storing a computer program, characterized in that: When the computer program is executed by a processor, the multimodal wireless environment prediction method according to any one of claims 1 to 7 is implemented.

Images