An intelligent beam prediction method, device and equipment

Abstract

The embodiment of the application discloses a kind of intelligent beam prediction method, device and equipment, wherein the method comprises: obtaining environment image, and the environment image includes the environment position information of base station and terminal;Determine the obstacle information on the direct path from the base station to the terminal based on the environment position information in the environment image;In response to the obstacle information representing that there is an obstacle on the direct path, determine the target boundary point based on the obstacle;Determine the emission angle, the incidence angle and the propagation distance of the target beam between the base station and the terminal based on the target boundary point, and determine the target beam direction based on the emission angle, the incidence angle and the propagation distance of the target beam.

Description

Technical Field

[0001] This application relates to, but is not limited to, the field of intelligent beam prediction, and particularly to an intelligent beam prediction method, apparatus, and device. Background Technology

[0002] Massive MIMO (Massively Multi-Tone) technology has significantly improved the capacity of communication systems. In high-frequency scenarios, more precise and lower-overhead intelligent beamforming is key to achieving even greater capacity increases in future 6G communication systems. However, with the increase in the number of beams, the overhead of globally scanning and measuring all beams is enormous and unacceptable in practical systems. Summary of the Invention

[0003] In view of the above, this application provides an intelligent beam prediction method, apparatus, and device to address the problems existing in the prior art.

[0004] The technical solution of this application embodiment is implemented as follows:

[0005] In a first aspect, embodiments of this application provide a smart beam prediction method, including:

[0006] An environmental image is acquired, including environmental location information of a base station and a terminal. Based on the environmental location information in the environmental image, obstacle information on the direct path from the base station to the terminal is determined. In response to the obstacle information indicating the presence of obstacles on the direct path, a target boundary point is determined based on the obstacles. Based on the target boundary point, the emission angle, incident angle, and propagation distance of the target beam between the base station and the terminal are determined. Based on the emission angle, incident angle, and propagation distance of the target beam, the direction of the target beam is determined.

[0007] Secondly, embodiments of this application provide an intelligent beam prediction system, the system comprising:

[0008] A receiver, a transmitter, and a processor, wherein the processor, when executing the program, implements the steps in the method described in the first aspect.

[0009] In this application embodiment, different intelligent beam prediction strategies can be used depending on the scenario. Data preprocessing can transform the transmission environment of a non-line-of-sight (NLOS) channel into a line-of-sight (LOS) channel. Therefore, when determining the target beam direction using only environmental images, significant resources can be saved while achieving good results. For determining the target beam direction for a specific terminal, the most likely beam directions can be obtained first using the location information of the environmental image. This eliminates the sampling process and saves resources. Furthermore, this method requires only a small amount of data to train a model. By decomposing the model into a horizontal direction model of the transmitted beam, a vertical direction model of the transmitted beam, and a received beam model, much data can be reused, and the trained model is easily transferable, including to different environments and carrier frequencies. This method can be used to handle situations where beam scanning is not feasible due to system overhead, or where environmental images, base station, and terminal location information are unavailable, making it widely applicable. In practical use, the trained machine learning model runs quickly, is small, and has low resource consumption and storage overhead, facilitating practical deployment. Attached Figure Description

[0010] In the accompanying drawings (which are not necessarily drawn to scale), similar reference numerals may describe similar parts in different views. Similar reference numerals with different letter suffixes may indicate different examples of similar parts. The drawings illustrate, by way of example and not limitation, the various embodiments discussed herein.

[0011] Figure 1 A schematic diagram illustrating the implementation process of an intelligent beam prediction method provided in this application embodiment;

[0012] Figure 2 This is a schematic diagram illustrating the implementation process of the massive MIMO technology provided in the embodiments of this application;

[0013] Figure 3 A schematic diagram illustrating the analysis process of environmental images provided in this application embodiment;

[0014] Figure 4 This is a schematic diagram of the beam strength information analysis process provided in the embodiments of this application;

[0015] Figure 5 This is another schematic diagram of the analysis process for beam intensity information provided in the embodiments of this application;

[0016] Figure 6 A schematic diagram of another analysis process for beam intensity information provided in an embodiment of this application;

[0017] Figure 7 A schematic diagram illustrating the data preprocessing implementation flow of an intelligent beam prediction method provided in this application embodiment;

[0018] Figure 8A This is a schematic diagram illustrating the implementation process of environmental image processing provided in the embodiments of this application;

[0019] Figure 8B This is a schematic diagram illustrating another implementation process of environmental image processing provided in an embodiment of this application;

[0020] Figures 9A to 9E This is a schematic diagram illustrating the implementation process of boundary information processing provided in an embodiment of this application;

[0021] Figure 10A A schematic diagram illustrating the implementation process of model training for an intelligent beam prediction method provided in this application embodiment;

[0022] Figure 10B A schematic diagram illustrating another implementation process of model training for an intelligent beam prediction method provided in this application embodiment;

[0023] Figure 10C This is a schematic diagram illustrating another implementation process of model training for an intelligent beam prediction method provided in this application embodiment;

[0024] Figure 11 This is a schematic diagram illustrating the actual deployment process of an intelligent beam prediction method provided in this application embodiment. Detailed Implementation

[0025] It should be understood that the specific embodiments described herein are for illustrative purposes only and are not intended to limit the scope of this application.

[0026] In the following description, the use of suffixes such as "module," "part," or "unit" to denote elements is solely for the purpose of illustration and has no specific meaning in itself. Therefore, "module," "part," or "unit" may be used interchangeably.

[0027] Electronic devices can be implemented in various forms. For example, the electronic devices described in this application may include mobile electronic devices such as personal digital assistants (PDAs), navigation devices, and wearable devices, as well as fixed electronic devices such as digital TVs and desktop computers that can collect fingerprints.

[0028] The following description will use mobile devices or base station devices as examples. Those skilled in the art will understand that, in addition to components specifically designed for mobile purposes, the construction according to the embodiments of this application can also be applied to fixed-type electronic devices.

[0029] Based on this, this application provides an intelligent beam prediction method that can select different beam prediction strategies according to different scenarios, convert NLOS to LOS, eliminate the sampling process, and save resource consumption. Model decomposition allows for the reuse of much data, resulting in a well-transferable trained model with fast execution speed and low resource consumption. In this application embodiment, the intelligent beam prediction method can be executed by the processor of an intelligent beam prediction system. Figure 1 This is a schematic diagram illustrating the implementation process of an intelligent beam prediction method provided in an embodiment of this application, as shown below. Figure 1 As shown, the method includes the following steps S101 to S103:

[0030] Step S101: Obtain an environmental image, which includes the environmental location information of the base station and the terminal.

[0031] Here, the environmental image refers to the environment corresponding to the current base station and device. The environmental image includes the environmental location information of the base station and terminal; the environmental location information includes at least the location coordinates of the base station, the location coordinates of the terminal, and the locations of surrounding buildings. By acquiring the environmental image, we can understand the location coordinates of the base station, the location coordinates of the terminal, and the locations of surrounding buildings.

[0032] In some possible implementations, environmental images can be obtained from input data or from an image acquisition device.

[0033] Step S102: Based on the environmental location information in the environmental image, determine the obstacle information on the direct path from the base station to the terminal.

[0034] Here, the direct path is the straight path between the base station and the terminal. Obstacle information is used to characterize whether there are obstacles on the direct path, including cases where obstacles are present and cases where obstacles are not present. Based on the analysis of the location coordinates of the base station, the location coordinates of the terminal, and the locations of surrounding buildings, it is determined whether there are obstacles such as buildings on the direct path from the base station to the terminal.

[0035] In some possible implementations, obstacle information can be determined by analyzing the location information of buildings in environmental images.

[0036] Step S103: In response to the obstacle information indicating the existence of obstacles on the direct path, determine the target boundary point based on the obstacles.

[0037] Here, the presence of obstacles on the direct path indicates that the transmission environment between the base station and the terminal is NLOS (Normally Inoperable). By using the target boundary point, the obstacle can be avoided to determine the LOS (Normally Inoperable) path from the base station to the terminal. When obstacles exist on the direct path between the base station and the terminal, the target boundary point is determined by calculating and analyzing the location information of the obstacles, the base station, and the terminal.

[0038] In some possible implementations, the equation of the line segment is obtained by using the location coordinates of the base station and the terminal, thereby obtaining all the points that the line segment passes through. It can be determined whether the line segment is a LOS path by judging whether all the points it passes through fall on obstacles.

[0039] Step S104: Based on the target boundary point, determine the transmission angle, incident angle, and propagation distance of the target beam between the base station and the terminal.

[0040] Here, the transmission angle, incident angle, and propagation distance of the target beam between the base station and the terminal are the corresponding transmission angle, incident angle, and propagation distance of the target beam. By combining the location information of the target boundary point with the location information of the base station and the terminal, beam information such as the transmission angle, incident angle, and propagation distance of the target beam between the base station and the terminal can be determined.

[0041] Step S105: Determine the direction of the target beam based on the emission angle, incident angle and propagation distance of the target beam.

[0042] Here, the target beam is the strongest beam between the base station and the terminal, and the direction of the target beam is the beam direction corresponding to the strongest beam. By combining beam information such as the transmission angle, incident angle, and propagation distance of the target beam between the base station and the terminal, the specific direction of the target beam between the base station and the terminal can be determined.

[0043] In this embodiment, after acquiring an environmental image between the base station and the terminal, the presence of obstacles on the direct path between them is determined using the image. If obstacles exist, the transmission angle, incident angle, and propagation distance of the target beam between the base station and the terminal are determined by identifying target boundary points, thereby determining the target beam direction. Thus, the target beam direction between the base station and the terminal can be determined using the environmental image. Furthermore, when obstacles exist between the base station and the terminal, and the transmission environment between them is NLOS (Normally Ordinary System), determining the target boundary points through positional analysis of the base station, terminal, and obstacles allows the transmission environment between the base station and the terminal to switch to LOS (Lessly System), effectively converting NLOS to LOS and thus more accurately determining the transmission angle, incident angle, and propagation distance of the target beam between them.

[0044] In some embodiments, after determining the transmission angle, incident angle, and propagation distance of the target beam between the base station and the terminal based on the target boundary point, the target beam direction is accurately determined by analyzing whether beam strength information between the base station and the terminal is obtained. That is, step S102 above can be implemented through the following steps:

[0045] Step S121: Determine whether beam strength information between the base station and the terminal has been obtained.

[0046] Here, beam strength information is obtained by sampling the beam between the base station and the terminal. After determining the transmission angle, incident angle, and propagation distance of the target beam between the base station and the terminal using information acquired from the environmental image, it can be determined through analysis whether beam strength information between the base station and the terminal is simultaneously obtained.

[0047] Step S122: In response to obtaining the beam strength information, determine the intensity characteristics corresponding to the beam strength information.

[0048] Here, the corresponding intensity features can be extracted from the beam strength information between the base station and the terminal. Therefore, when obtaining the beam strength information between the base station and the terminal, the corresponding intensity features can be determined by extracting features from the beam strength information.

[0049] In some possible implementations, the strength features include the index of the strongest sampled beam between the base station and the terminal, the value of the strongest sampled beam, the index of the second strongest sampled beam, and the value of the second strongest sampled beam, etc.

[0050] Step S123: Determine the direction of the target beam based on the emission angle, incident angle, propagation distance, and intensity characteristics of the target beam.

[0051] Here, the emission angle, incident angle, and propagation distance of the target beam determined from the environmental image are combined with the intensity characteristics determined from the beam intensity information, and used to determine the direction of the target beam.

[0052] In this embodiment, when both environmental imagery and beam strength information between the base station and the terminal are provided, the transmission angle, incident angle, and propagation distance of the target beam determined from the environmental image can be combined with the intensity characteristics determined from the beam strength information to determine the target beam direction. Thus, by enriching the analysis data, the target beam direction can be determined more accurately, improving the accuracy of target beam direction prediction.

[0053] In some embodiments, in step S105 above, the target beam direction can be determined in the following manner:

[0054] The first step is to determine the horizontal direction of the transmitting beam, the vertical direction of the transmitting beam, and the receiving beam in the target beam based on the transmission angle, incident angle, and propagation distance of the target beam.

[0055] Here, the horizontal direction of the transmitted beam is the same as the horizontal transmission direction of the transmitted beam in the target beam, the vertical direction of the transmitted beam is the same as the vertical transmission direction of the transmitted beam in the target beam, and the receiving beam is the same as the receiving direction of the receiving beam in the target beam. The horizontal direction, vertical direction, and receiving beam of the transmitted beam in the target beam are further determined by the transmission angle, incident angle, and propagation distance of the target beam.

[0056] The second step is to determine the direction of the target beam based on the horizontal direction of the transmitted beam, the vertical direction of the transmitted beam, and the received beam.

[0057] Here, the horizontal direction of the transmitted beam, the vertical direction of the transmitted beam, and the direction of the received beam are combined to determine the direction of the target beam.

[0058] In this embodiment, the horizontal direction, vertical direction, and receiving beam of the transmitting beam are first determined by the transmission angle, incident angle, and propagation distance of the target beam. Then, the direction of the target beam is further determined based on the horizontal direction, vertical direction, and receiving beam of the transmitting beam. This decomposes the problem of finding the target beam from hundreds of beams into identifying the strongest beam in three sets of problems: horizontal transmission direction, vertical transmission direction, and receiving direction. This decomposition improves the model's versatility and facilitates transfer; the intensity features in the decomposed beam intensity information are easier to extract; and the predicted data for the receiving direction can be used in both horizontal and vertical transmission directions, improving data reusability and saving resource consumption.

[0059] In some embodiments, when no environmental image is acquired, the smart beam prediction process includes:

[0060] The first step is to obtain beam strength information between the base station and the terminal in response to the failure to acquire the environmental image.

[0061] Here, when no environmental image between the base station and the terminal is obtained, the beam strength information between the base station and the terminal is obtained by scanning the beam between the base station and the terminal.

[0062] The second step is to determine the horizontal direction of the transmitted beam, the vertical direction of the transmitted beam, and the received beam based on the intensity characteristics corresponding to the beam intensity information.

[0063] Here, the corresponding intensity characteristics are determined by analyzing the beam strength information. Based on the beam strength characteristics of all beams between the base station and the terminal contained in the intensity characteristics, the beam with the largest beam strength in the horizontal direction of the transmitting beam, the vertical direction of the transmitting beam, and the receiving beam is determined as the horizontal direction of the transmitting beam, the vertical direction of the transmitting beam, and the receiving beam in the target beam.

[0064] The third step is to determine the direction of the target beam based on the horizontal direction of the transmitted beam, the vertical direction of the transmitted beam, and the received beam.

[0065] Here, the horizontal direction of the transmitted beam, the vertical direction of the transmitted beam, and the direction of the received beam are combined to determine the target beam direction.

[0066] In this embodiment, when environmental images of the base station and terminal are not available, and only beam strength information is provided, the horizontal direction, vertical direction, and receiving beam of the target beam are determined by extracting corresponding intensity features from the beam strength information. This method enables the determination of the target beam by extracting intensity features even with only beam strength information. Furthermore, decomposing the problem improves the model's versatility and facilitates transfer; the intensity features in the decomposed beam strength information are easier to extract; and the predicted data for the receiving direction is applicable to both horizontal and vertical transmission directions, improving data reusability and saving resource consumption.

[0067] In some embodiments, where the obstacle information indicates that there are no obstacles on the direct path, determining the emission angle, incident angle, and propagation distance of the target beam can be achieved through the following process:

[0068] In response to the obstacle information indicating that there are no obstacles on the direct path, the emission angle, incident angle and propagation distance of the target beam are determined based on the direct path.

[0069] Here, if there are no obstacles on the direct path between the base station and the terminal, the transmission environment between the base station and the terminal is considered to be LOS. Therefore, the transmission angle, incident angle, and propagation distance of the target beam can be directly determined based on the direct path.

[0070] In this embodiment, when there are no obstacles on the direct path between the base station and the terminal, the transmission angle, incident angle and propagation distance of the target beam can be directly determined based on the direct path, thereby determining the direction of the target beam. This eliminates the need to sample the beam between the base station and the device to obtain beam strength information, thus saving resource consumption.

[0071] In some embodiments, in step S103 above, the target boundary point can be determined in the following manner:

[0072] The first step is to determine the boundary points that are directly visible to the base station and the terminal based on the boundary information of the obstacle.

[0073] Here, the boundary information of the obstacle is obtained from the obstacle boundary in the environmental image through the Laplacian operator; by traversing all the boundary information, the boundary points that are LOS with both the base station and the terminal are found; thus, the boundary points that can be seen in a straight line from the base station and the terminal are determined by the boundary information of the obstacle.

[0074] The second step is to determine the target boundary point based on the path information from each boundary point to the base station and the terminal.

[0075] Here, the path information includes the path from each boundary point to the base station and the path from each boundary point to the terminal. The optimal boundary point is determined as the target boundary point through these paths.

[0076] In this embodiment, boundary points that are both LOS (Low-Incidence) and LOS are determined by analyzing the boundary information of obstacles between the base station and the terminal. The optimal boundary point is then selected as the target boundary point based on the path from each boundary point to the base station and the terminal. This method enables the determination of target boundary points that can convert NLOS to LOS in an NLOS transmission environment, thereby determining the transmission angle, incident angle, and propagation distance of the target beam between the base station and the terminal.

[0077] In some embodiments, in step S104 above, the emission angle, incident angle, and propagation distance of the target beam can be determined in the following ways:

[0078] The first step is to determine the angle from the base station to the target boundary point as the transmission angle.

[0079] Here, the angle corresponding to the line connecting the base station to the target boundary point is determined as the beam transmission angle, i.e., the transmission angle.

[0080] The second step is to determine the angle from the target boundary point to the terminal as the incident angle.

[0081] Here, the angle corresponding to the line connecting the target boundary point to the terminal is determined as the incident angle of the beam, i.e., the incident angle.

[0082] The third step is to determine the propagation distance as the sum of the distances from the base station to the target boundary point and from the target boundary point to the terminal.

[0083] Here, the path from the base station to the target boundary point and the path from the target boundary point to the terminal are determined, and the sum of the distances of the two paths is determined as the beam propagation distance.

[0084] In this embodiment, the angle corresponding to the line connecting the base station to the target boundary point is determined as the beam's transmission angle, the angle corresponding to the line connecting the target boundary point to the terminal is determined as the beam's incident angle, and the sum of the distances of the two paths from the base station and the terminal to the target boundary point is determined as the beam's propagation distance. This method determines the beam information representing the target, thereby accurately determining the target beam direction.

[0085] In some embodiments, after acquiring the environmental image, the environmental image can also be updated in the following ways:

[0086] The first step is to detect the environmental location information of the base station and the terminal.

[0087] Here, after acquiring the environmental image, the current environmental location information of the base station and terminal corresponding to the current environmental image is detected in real time.

[0088] The second step is to update the environmental image based on the environmental location information if the environmental location information changes.

[0089] Here, when changes in the environmental location information corresponding to the base station and the terminal are detected, such as changes in the number and location information of user terminals or changes in the location information of buildings around the base station and the terminal, the environmental image is updated according to the current environmental location information, so that the environmental image can be updated in real time and change with the changes in the environmental location information corresponding to the current base station and the terminal.

[0090] In some possible implementations, upon receiving user requests for environmental images, the current environmental location information is detected according to the user's specified requests to update the environmental image. The user's requests may include personalized requirements regarding the elements included in the environmental image and specific time intervals for detecting environmental location information. For example, if the user specifies that the environmental image should be updated every hour, then if one hour has passed since the last update, the environmental location information of the base station and the terminal is detected, and the environmental image is updated accordingly.

[0091] In this embodiment, the environmental location information corresponding to the environmental image can be updated in real time according to changes in the environmental location information such as the current base station, terminal, and surrounding obstacles. This method improves the timeliness of the environmental image and reduces the inaccuracy of the target beam direction determined from the environmental image when environmental location information changes.

[0092] In some embodiments, the target beam direction can be determined through the following steps in step S105 above:

[0093] The first step is to receive the horizontal direction model, the vertical direction model, and the receiving beam model of the transmitted beam; wherein, the horizontal direction model, the vertical direction model, and the receiving beam model of the transmitted beam can be updated based on the emission angle, incident angle, propagation distance, and intensity characteristics corresponding to the beam intensity information of the sample beam.

[0094] Here, the horizontal direction model, vertical direction model, and receiving beam model of the transmitted beam can be updated based on the intensity features corresponding to the emission angle, incident angle, propagation distance, and beam intensity information of the sample beams. The horizontal direction model of the transmitted beam can predict the horizontal direction corresponding to the strongest horizontal beam among all transmitted beams, the vertical direction model of the transmitted beam can predict the vertical direction corresponding to the strongest vertical beam among all transmitted beams, and the receiving beam model can predict the receiving direction corresponding to the strongest horizontal beam among all received beams. The horizontal direction model, vertical direction model, and receiving beam model of the transmitted beam are obtained by using the intensity features corresponding to the extracted emission angle, incident angle, propagation distance, and beam intensity information of the sample beams as input data, and using the horizontal direction, vertical direction, and receiving beam as labels.

[0095] The second step involves inputting the emission angle, incident angle, and propagation distance of the target beam and / or the intensity characteristics corresponding to the beam intensity information into the horizontal direction model, vertical direction model, and receiving beam model of the transmitting beam, respectively, to obtain the horizontal direction of the transmitting beam, the vertical direction of the transmitting beam, and the receiving beam.

[0096] Here, the horizontal direction of the transmitted beam is the horizontal direction corresponding to the strongest horizontal beam among all transmitted beams, the vertical direction of the transmitted beam is the vertical direction corresponding to the strongest vertical beam among all transmitted beams, and the receiving beam is the receiving direction corresponding to the strongest receiving beam among all received beams. The horizontal direction, vertical direction, and receiving beam are obtained by inputting the intensity characteristics corresponding to the transmission angle, incident angle, propagation distance, and / or beam strength information of the target beam into the horizontal direction model, vertical direction model, and receiving beam model of the transmitted beam, respectively.

[0097] The third step is to determine the direction of the target beam based on the horizontal direction of the transmitted beam, the vertical direction of the transmitted beam, and the received beam.

[0098] Here, the direction of the target beam is determined by combining the determined horizontal direction of the transmitted beam, the vertical direction of the transmitted beam, and the received beam.

[0099] In some possible implementations, the horizontal direction corresponding to the strongest horizontal beam among all transmitted beams is determined by the horizontal direction model of the transmitted beam, the vertical direction corresponding to the strongest vertical beam among all transmitted beams is determined by the vertical direction model of the transmitted beam, and the receiving direction corresponding to the strongest horizontal beam among all received beams is determined by the receiving beam model.

[0100] In this embodiment, the direction of the target beam is determined by inputting the emission angle, incident angle, propagation distance, and / or beam intensity information of the target beam into the horizontal direction model, vertical direction model, and receiving beam model of the transmitted beam, respectively. This process yields the horizontal direction, vertical direction, and receiving beam of the transmitted beam. By using these three models—the horizontal direction model, the vertical direction model, and the receiving beam model—the model for finding the target beam from hundreds of beams is decomposed into three separate models, improving model versatility and facilitating transfer. Furthermore, the data in the receiving beam model can be used in both the horizontal and vertical direction models of the transmitted beam, enhancing data reusability and saving resource consumption.

[0101] The following describes the application of the intelligent beam prediction method provided in this application in a real-world scenario, taking the beam between a base station device and a terminal device as an example.

[0102] The deep integration of communication and artificial intelligence technologies has become one of the most important directions in the development of wireless communication systems. Looking towards 6G, the angle and depth of this integration will be further expanded. For example... Figure 2 As shown, Massive Multiple Input Multiple Output (MIMO) technology significantly improves the capacity of communication systems. In high-frequency scenarios, more accurate and lower-overhead intelligent beamforming is key to achieving even greater capacity improvements in future 6G communication systems. Correct MIMO beam selection is based on accurate beam measurement. However, as the number of beams increases, the overhead of globally scanning and measuring all beams becomes enormous, making it unacceptable in practical systems. A more practical approach is to first perform sparse beam scanning measurements, then predict other unmeasured beams based on the measurement results, and finally select the strongest beam. How to use Artificial Intelligence (AI) technology to obtain the most accurate beam prediction results with a given measurement overhead is a very important research topic for future 6G communications.

[0103] In practical applications, beam prediction and model transfer issues for MIMO systems need to be considered. Model transfer includes transfer between different carrier frequency systems, as well as transfer from a general transmission environment dataset to a specific transmission environment. Environment transfer refers to the model's ability to predict beams for various transmission environment scenarios. The transmission environment image provides the layout of buildings in the generated data scenario. Carrier frequency system transfer refers to transfer between different carrier frequencies; we need the model to be equally applicable to systems with different carrier frequencies.

[0104] The first method involves a communication system with a carrier frequency of f1, consisting of a 64x4 beam pair set comprising 64 transmit beams and 4 receive beams. For each receive beam, 8 transmit beams are scanned to obtain the measurement results for 8x4 beam pairs. The data structure is <intensity of 64x4 beam pairs, transmission environment image, base station (BS) location, and terminal (UE) location>.

[0105] The second approach, for a communication system with a carrier frequency of f2, considers a set of 128x4 beam pairs consisting of 128 transmit beams and 4 receive beams. For each receive beam, scan the 8 transmit beams to obtain the measurement results for 8x4 beam pairs.

[0106] Because there may or may not be building obstructions between the base station and the terminal, these transmission environments may be NLOS or LOS environments. For example... Figure 3 As shown, in the environmental image, the shaded areas represent buildings, the squares represent base stations, and the triangles represent user terminals.

[0107] like Figure 4 The diagram illustrates beam strength information. The center image shows environmental data, including building locations, base station and terminal coordinates. The surrounding heatmaps display beam strength for each location, with two graphs for each position: the right side shows the original decibel (dB) value, and the left side shows the value linearized to 10**(dB / 10). It's clear that a full 64x4 beam is too much; therefore, in practice, only these beams are sampled. This dataset is centrally sampled as 8x4. For example, with 64 transmitting directions × 4 receiving directions, the base station samples 8 beams in 8 directions and transmits them with equal strength. The terminal receives these 8 beams in the four receiving directions and predicts which beam was strongest during transmission. This prediction is then fed back to the base station, which subsequently uses this strongest beam to send a signal to the terminal, which receives the signal from the predicted strongest receiving direction.

[0108] like Figure 5As shown, by analyzing the beam information data under three scenarios (tasks), it can be found that task1 and task3 have 16 beams in the horizontal direction and 4 beams in the vertical direction, while task2 has 16 beams in the horizontal direction and 8 beams in the vertical direction. Therefore, the three tasks are consistent in the horizontal direction of the beams.

[0109] In some possible implementations, trigonometric functions can be used to convert the spatial coordinates of the base station and the terminal into the angle and distance between the base station and the terminal.

[0110] Regarding the case of LOS, the following conclusions can be drawn from the analysis of LOS:

[0111] 1. The determination of the horizontal beam of the transmission beam is mainly related to the transmission angle from the base station to the terminal. This is the strongest feature and is not affected by the carrier frequency. In other words, the horizontal beam of the transmission beam can be determined by the same transmission angle under different carrier frequencies.

[0112] 2. The determination of the vertical beam of the transmitted beam is mainly related to the propagation distance from the base station to the terminal, and this is the strongest feature.

[0113] 3. The determination of the receiving beam is mainly related to the angle of incidence from the base station to the terminal, and is not affected by the carrier frequency. In other words, the horizontal beam of the transmitting beam can be determined by the same angle of incidence under different carrier frequencies.

[0114] 4. The receiving beam decomposed from the direction of the strongest beam is usually equal to the receiving antenna where the maximum value of the sampled beam is located. In other words, the receiving beam can be accurately determined in most cases by indexing the maximum value of the sampled beam.

[0115] In the case of NLOS, it can be converted to LOS based on the environmental image and location information for unified processing.

[0116] Table 1

[0117] 0 z0 z8 z16 z24 1 2 3 1 4 5 6 7 2 8 9 z1 z9 z17 z25 11 3 12 13 14 15 4 16 z2 z10 z18 z26 18 19 5 20 21 22 23 6 24 25 26 z3 z11 z19 z27 7 28 29 30 31 8 z4 z12 z20 z28 33 34 35 9 36 37 38 39 10 40 41 z5 z13 z21 z29 43 11 44 45 46 47 12 48 z6 z14 z22 z30 50 51 13 52 53 54 55 14 56 57 58 z7 z15 z23 z31 15 60 61 62 63

[0118] Table 2

[0119] 0 z0 z8 z16 z24 1 2 3 4 5 6 7 1 8 9 10 11 12 13 14 15 2 16 17 z1 z9 z17 z25 19 20 21 22 23 3 24 25 26 27 28 29 30 31 4 32 33 34 35 z2 z10 z18 z26 37 38 39 5 40 41 42 43 44 45 46 47 6 48 49 50 51 52 53 Z3 z11 z19 z27 55 7 56 57 58 59 60 61 62 63 8 64 Z4 z12 z20 z28 66 67 68 69 70 71 9 72 73 74 75 76 77 78 79 10 80 81 82 z5 z13 z21 z29 84 85 86 87 11 88 89 90 91 92 93 94 95 12 96 97 98 99 100 z6 z14 z22 z30 102 103 13 104 105 106 107 108 109 110 111 14 112 113 114 115 116 117 118 z7 z15 z23 z31 15 120 121 122 123 124 125 126 127

[0120] As shown in Tables 1 and 2, the beam direction information data reveals that each beam covers approximately 7.5 degrees horizontally, with 16 beams covering 120 degrees, and the beams repeat three times to cover the entire plane; the received beams are symmetrical about the horizontal axis. Based on the analyzed beam direction information, the sampling strategy can be deduced as follows: sampling occurs every other horizontal direction and every other vertical direction. The difference lies in that scenarios 1 and 3 have four vertical directions, while scenario 2 has eight.

[0121] The above analysis reveals that, for example Figure 6 As shown, the beam prediction problem can be decomposed into three parts: the horizontal beam of the transmitted beam, the vertical beam of the transmitted beam, and the direction of the received beam. The horizontal beam model, the vertical beam model, and the received beam model are used to predict the horizontal beam, vertical beam, and received beam of the transmitted beam, respectively. Then, these three beams are combined to obtain the final target beam. The prediction models for tasks 1 and 3 are general; if only the transmission angle and propagation distance are considered, task 2 can also use a single prediction model. For the prediction of the received beam direction, if the maximum value of the eight transmitted beams in each receiving direction is taken as the feature, a single prediction model can be used.

[0122] The vertical direction of the transmit beam uses the following features: the maximum value of the four beams in the receiving direction at the same position of the eight sampled transmit beams, the index of the strongest sampled beam, the value of the strongest sampled beam, the index of the second strongest sampled beam, the value of the second strongest sampled beam, the transmit angle, and the propagation distance. When using the transmit angle and propagation distance features, the prediction models for task 1 and task 3 can be shared.

[0123] The horizontal features used for the transmit beam include: the maximum value of the four beams in the receiving direction at the same position of the eight sampled transmit beams; the index of the strongest sampled beam; the value of the strongest sampled beam; the index of the second strongest sampled beam; the value of the second strongest sampled beam; the transmit angle; and the propagation distance. When using the transmit angle and propagation distance features, the prediction models for task 1 and task 3 can be shared.

[0124] The receiving beam direction uses the following features: incident angle, propagation distance, and the maximum value of the eight beams in the receiving direction.

[0125] Figure 7 This application provides a schematic diagram of the data preprocessing implementation flow for an intelligent beam prediction method, as illustrated in the embodiments of this application. Figure 7 As shown, the method includes the following steps:

[0126] The first step is to acquire environmental images containing location information of base stations and terminals.

[0127] The second step is to obtain the diameter distance and angle from the base station to the terminal using trigonometric functions, and to determine whether the direct path passes through buildings or not.

[0128] Here, when the direct path does not pass through a building, the transmission environment is determined to be LOS; otherwise, the transmission environment is determined to be NLOS.

[0129] The third step, in the case of LOS, is to use the emission angle, incident angle, and propagation distance of the direct path.

[0130] The fourth step, in the case of NLOS, is to process the environmental image to obtain building boundary information.

[0131] The fifth step is to obtain building boundary information. Based on the building boundary information and the location information of the base station and the terminal, the shortest path is calculated to determine whether it belongs to transmission or diffraction, thereby obtaining a more accurate transmission angle, incident angle and propagation distance.

[0132] The sixth step is to obtain information such as the emission angle, incident angle, and propagation distance.

[0133] Here, there are many methods to process environmental images and obtain obstacle boundary information. A simple method is to use the Laplacian operator for direct processing, which is fast and consumes few resources. For example... Figure 8A The environmental image shown is obtained by processing it with the Laplacian operator. Figure 8B The boundary information of each building is shown. For example... Figures 9A to 9E As shown, under different tasks, different environments, and different terminals (UEs), the reflection or diffraction path can be calculated based on the boundary information of the building, thereby more accurately determining the transmission angle and incident angle of the beam from the base station to the terminal in the case of NLOS.

[0134] The method for converting NLOS to LOS is as follows:

[0135] The first step is to use the Laplace operator to obtain the boundary points of the building.

[0136] The second step is to traverse each boundary point, find the boundary points that are both LOS (Local Out of Memory) with the base station and the terminal, and calculate the sum of the path lengths from these points to the base station and the terminal.

[0137] Here, the method to determine whether it is a LOS (Loss in Situation) is to obtain the equation of the line segment based on the coordinates of the starting and ending points, thereby obtaining all the points that the line segment passes through, and then determining whether all the points that the line segment passes through fall on the building.

[0138] The third step is to find the optimal boundary point as the target boundary point, and then obtain the optimal path based on the target boundary point.

[0139] The fourth step is to use the angle from the base station to the target boundary point as the transmission angle.

[0140] Fifth, use the angle from the target boundary point to the terminal as the angle of incidence.

[0141] Step 6: The propagation distance is the sum of the distances of the two LOS segments.

[0142] Step S806: Obtain information such as emission angle, incident angle, and propagation distance.

[0143] The model training part includes three cases: the first is to provide only the environmental image containing the location information of the base station and the terminal; the second is to provide only the beam strength information between the base station and the terminal; and the third is to provide both the environmental image and the beam strength information. The model is trained for each of these three cases.

[0144] The first type, such as Figure 10A As shown, considering the reduction of sampling overhead, for the case where only environmental images, base station, and terminal location information are provided, the training process includes the following:

[0145] The first step is to obtain data such as emission angle, incident angle, and propagation distance as features by preprocessing the environmental image.

[0146] The second step is to train the model using the transverse beam of the transmitted beam as the label, the longitudinal beam of the transmitted beam as the label, and the received beam as the label.

[0147] The third step is to obtain the horizontal direction prediction model of the transmitted beam: sendbeamHorizontal4onlyenv, the vertical direction prediction model of the transmitted beam: sendbeamvertical4onlyenv, and the prediction model of the received beam: recvbeamvertical4onlyenv.

[0148] The second type, such as Figure 10B As shown, considering the inability to obtain environmental information or base station and terminal location information, only the scanned beam intensity information is available, including the following training process:

[0149] The first step is to extract intensity features from the beam intensity information.

[0150] The second step is to train the model using the horizontal direction of the transmitted beam as the label, the vertical direction of the transmitted beam as the label, and the received beam as the label.

[0151] The third step is to obtain the horizontal direction prediction model of the transmitted beam: sendbeamHorizontal4onlyscan, the vertical direction prediction model of the transmitted beam: sendbeamvertical4onlyscan, and the prediction model of the received beam: vertical4onlyscanrecvbeam.

[0152] Here, the intensity features extracted from the beam strength information include the maximum beam strength, the second largest beam strength, the index of the maximum beam strength, and the index of the second largest beam strength.

[0153] The third type, such as Figure 10C As shown, the simultaneous provision of environmental images and beam intensity information includes the following training process:

[0154] The first step involves obtaining information such as emission angle, incident angle, and propagation distance as features through image preprocessing, which are then combined with intensity features extracted from the intensity information of the scanning beam.

[0155] The second step is to train the model using the horizontal direction of the transmitted beam as the label, the vertical direction of the transmitted beam as the label, and the received beam as the label.

[0156] The third step is to obtain the horizontal prediction model of the transmitted beam: sendbeam Horizontal, the vertical prediction model of the transmitted beam: sendbeam vertical, and the prediction model of the received beam: recvbeam vertical.

[0157] The deployment and usage component includes: depending on whether environmental location information and scanned beam strength information can be obtained in the actual deployment environment, corresponding features can be extracted and different trained models can be selected to predict the horizontal direction of the strongest transmitted beam, the vertical direction of the strongest transmitted beam, and the strongest received beam, and then combined into the final target beam direction. Figure 11 This is a schematic diagram illustrating the actual deployment implementation process of an intelligent beam prediction method provided in this application embodiment, as shown below. Figure 11 As shown, the method specifically includes the following steps:

[0158] The first step is to determine whether an environmental image has been provided.

[0159] Here, if an environment image is provided, proceed to step two; otherwise, proceed to step three.

[0160] The second step is to extract feature information such as emission angle, incident angle, and propagation distance, and then proceed to the fourth step.

[0161] The third step is to extract features, input them into sendbeamHorizontal4onlyscan to predict the horizontal direction of the transmitted beam, input them into sendbeamvertical4onlyscan to predict the vertical direction of the transmitted beam, input them into recvbeamvertical4onlyscan to predict the received beam, and then execute the seventh step.

[0162] The fourth step is to determine whether the scanning beam intensity information is provided at the same time.

[0163] Here, if beam strength information is provided, proceed to step five; otherwise, proceed to step six.

[0164] Step 5: Extract features, input them into sendbeamHorizontal to predict the horizontal direction of the transmitted beam, input them into sendbeamvertical to predict the vertical direction of the transmitted beam, input them into recvbeamvertical to predict the received beam, and then execute Step 7.

[0165] Step 6: Extract features, input them into sendbeamHorizontal4onlyenv to predict the horizontal direction of the transmitted beam, input them into sendbeamvertical4onlyenv to predict the vertical direction of the transmitted beam, input them into recvbeamvertical4onlyenv to predict the received beam, and then execute Step 7.

[0166] The seventh step is to combine the predicted horizontal direction of the transmitted beam, the vertical direction of the transmitted beam, and the received beam to obtain the target beam direction.

[0167] In this embodiment, different beam prediction strategies can be used depending on the scenario, and NLOS can be converted to LOS through data preprocessing. Therefore, when determining the target beam direction using only environmental images, significant resources can be saved while achieving good results. For determining the target beam direction for a specific terminal, the most likely beam directions can be obtained first through the location information of the environmental images. This eliminates the sampling process and saves resources. Furthermore, this method can build a model using only a small amount of data. By decomposing the model into a horizontal direction model of the transmitted beam, a vertical direction model of the transmitted beam, and a received beam model, much data can be reused, and the trained model can be easily transferred, including to different environments and carrier frequencies. This method can be used to handle situations where beam scanning is not possible due to system overhead, or where environmental images, base station, and terminal location information are unavailable, making it widely applicable. In practical use, the trained machine learning model runs quickly, is small, and has low resource consumption and storage overhead, facilitating practical deployment.

[0168] This application provides an intelligent beam prediction device, which includes various modules and units included in each module. It can be implemented by a processor in a terminal; of course, it can also be implemented by specific logic circuits. In the implementation process, the processor can be a central processing unit, a microprocessor, a digital signal processor, or a field-programmable gate array, etc.

[0169] This application provides an intelligent beam prediction system, which includes a receiver, a transmitter, and a processor; wherein the processor is used to implement the method described above; the transmitter is used to transmit a target beam based on the target beam direction; and the receiver is used to receive the target beam.

[0170] It should be noted that, in the embodiments of this application, if the above-mentioned problem discovery method is implemented as a software functional module and sold or used as an independent product, it can also be stored in a terminal-readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to the prior art, can be embodied in the form of a software product. This terminal software product is stored in a storage medium and includes several instructions to cause a terminal (which may be a personal computer or a server, etc.) to execute all or part of the methods described in the various embodiments of this application.

[0171] Correspondingly, embodiments of this application provide a storage medium storing executable instructions for inducing the processor to execute the aforementioned problem discovery method.

[0172] The descriptions of the storage medium and device embodiments above are similar to those of the method embodiments above, and have similar beneficial effects. For technical details not disclosed in the storage medium and device embodiments of this application, please refer to the descriptions of the method embodiments of this application for understanding.

[0173] In the several embodiments provided in this application, it should be understood that the disclosed devices and methods can be implemented in other ways. The device embodiments described above are merely illustrative. For example, the division of units is only a logical functional division, and in actual implementation, there may be other division methods, such as: multiple units or components can be combined, or integrated into another system, or some features can be ignored or not executed. In addition, the coupling, direct coupling, or communication connection between the various components shown or discussed can be through some interfaces, and the indirect coupling or communication connection between devices or units can be electrical, mechanical, or other forms. The above-mentioned separated components may or may not be physically separated, and the components shown may or may not be physical units; they can be located in one place or distributed across multiple network units; some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0174] Furthermore, the functional units in the various embodiments of this application can all be integrated into one processing unit, or each unit can be a separate unit, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or in a combination of hardware and software functional units. Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a readable storage medium, and when executed, it performs the steps of the above method embodiments. The aforementioned storage medium includes various media capable of storing program code, such as mobile storage devices, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks. Alternatively, if the integrated unit of this application is implemented as a software functional module and sold or used as an independent product, it can also be stored in a readable storage medium. Based on this understanding, the technical solution of the embodiments of this application, or the part that contributes to the prior art, can be embodied in the form of a software product. This software product is stored in a storage medium and includes several instructions to cause the terminal to execute all or part of the methods described in the various embodiments of this application. The aforementioned storage media include various media capable of storing program code, such as mobile storage devices, ROM, RAM, magnetic disks, or optical disks. The above description is merely a specific embodiment of this application, but the scope of protection of this application is not limited thereto. Any variations or substitutions that can be easily conceived by those skilled in the art within the scope of the technology disclosed in this application should be included within the scope of protection of this application. Therefore, the scope of protection of this application should be determined by the scope of the claims.

Claims

1. A smart beam prediction method, characterized in that, The method includes: Acquire an environmental image, which includes environmental location information of the base station and the terminal; Based on the environmental location information in the environmental image, the obstacle information on the direct path from the base station to the terminal is determined; In response to the obstacle information indicating the presence of obstacles on the direct path, the boundary points that are directly visible to the base station and the terminal are determined based on the boundary information of the obstacles; and the target boundary points are determined based on the path information from each boundary point to the base station and the terminal. Based on the target boundary points, determine the transmission angle, incident angle, and propagation distance of the target beam between the base station and the terminal; The direction of the target beam is determined based on the emission angle, incident angle, and propagation distance of the target beam.

2. The method according to claim 1, characterized in that, After determining the transmission angle, incident angle, and propagation distance of the target beam between the base station and the terminal based on the target boundary point, the method further includes: Determine whether beam strength information between the base station and the terminal has been obtained; In response to acquiring the beam strength information, the intensity characteristics corresponding to the beam strength information are determined; The direction of the target beam is determined based on the emission angle, incident angle, propagation distance, and intensity characteristics of the target beam.

3. The method according to claim 1, characterized in that, Determining the target beam direction based on the target beam's emission angle, incident angle, and propagation distance includes: Based on the transmission angle, incident angle, and propagation distance of the target beam, the horizontal direction of the transmitting beam, the vertical direction of the transmitting beam, and the receiving beam are determined. The direction of the target beam is determined based on the horizontal direction of the transmitted beam, the vertical direction of the transmitted beam, and the received beam.

4. The method according to claim 1, characterized in that, The method further includes: In response to the failure to acquire the environmental image, beam strength information between the base station and the terminal is acquired; Based on the intensity characteristics corresponding to the beam intensity information, the horizontal direction of the transmitted beam, the vertical direction of the transmitted beam, and the received beam are determined. The direction of the target beam is determined based on the horizontal direction of the transmitted beam, the vertical direction of the transmitted beam, and the received beam.

5. The method according to claim 1, characterized in that, After determining the obstacle information on the direct path from the base station to the terminal based on the environmental location information, the method further includes: In response to the obstacle information indicating that there are no obstacles on the direct path, the emission angle, incident angle and propagation distance of the target beam are determined based on the direct path.

6. The method according to claim 1, characterized in that, The step of determining the transmission angle, incident angle, and propagation distance of the target beam between the base station and the terminal based on the target boundary point includes: The angle from the base station to the target boundary point is defined as the transmission angle; The angle from the target boundary point to the terminal is defined as the incident angle; The propagation distance is the sum of the distances from the base station to the target boundary point and from the target boundary point to the terminal.

7. The method according to claim 1, characterized in that, After acquiring the environmental image, the method further includes: Detect the environmental location information of the base station and the terminal; If the environmental location information changes, the environmental image is updated based on the environmental location information.

8. The method according to claim 1, characterized in that, Determining the target beam direction based on the target beam's emission angle, incident angle, and propagation distance includes: The system includes a horizontal model of the transmitted beam, a vertical model of the transmitted beam, and a received beam model; wherein the horizontal model of the transmitted beam, the vertical model of the transmitted beam, and the received beam model can be updated based on the emission angle, incident angle, and propagation distance of the sample beam, as well as the intensity characteristics corresponding to the beam intensity information. The emission angle, incident angle, and propagation distance of the target beam and / or the intensity characteristics corresponding to the beam intensity information are respectively input into the horizontal direction model, vertical direction model, and receiving beam model of the transmitting beam to obtain the horizontal direction of the transmitting beam, the vertical direction of the transmitting beam, and the receiving beam. The direction of the target beam is determined based on the horizontal direction of the transmitted beam, the vertical direction of the transmitted beam, and the received beam.

9. A smart beam prediction system, characterized in that, The intelligent beam prediction system includes: Receiver, transmitter, and processor; among which, The processor is configured to implement the method according to any one of claims 1 to 8; The transmitter is used to transmit a target beam based on the target beam direction; The receiver is used to receive the target beam.

Images