Channel angle map construction method based on path cluster

Abstract

The invention relates to a channel angle map construction method and device based on a path cluster, computer equipment and a storage medium. The method comprises the following steps: mapping a communication coverage area in a target scene into a planar map; selecting a preset number of sampling points from the grid point set of the plane map, and measuring multi-path angle data of each sampling point; performing clustering analysis on the multi-path angle data in each sampling point to obtain a path cluster belonging to each sampling point; performing angle feature extraction based on the path cluster of each sampling point to obtain a multi-dimensional angle feature quantity of each sampling point; for the multi-dimensional angle characteristic quantity of each sampling point, adopting a spatial interpolation algorithm to predict the multi-dimensional angle characteristic quantity of non-sampling points in the grid point set; and combining the multi-dimensional angle characteristic quantities of the sampling points and the non-sampling points to construct a channel angle map. By adopting the method, the calculation process can be simplified to obtain the channel angle map, and beam scanning, beam scheduling and beam codebook design of the receiving and transmitting end can be better realized.

Description

Technical Field

[0001] The present application relates to the technical field of wireless communication channel estimation, and in particular to a method, device, computer equipment and storage medium for constructing a channel angle map based on a path cluster. Background Art

[0002] The future-oriented sixth generation mobile communication (6G) system will further enhance the exploitation of spatial resources. Massive multiple-input multiple-output (MIMO) is a key physical layer technology to enhance the spatial multiplexing capability of 6G. However, the application of massive MIMO communication systems requires a large amount of accurate angular channel state information (CSI) as support. Usually, angular CSI is obtained by channel estimation, but as the antenna scale continues to increase, the traditional channel estimation method based on pilot measurement will introduce greater pilot overhead, making the system unbearable.

[0003] The channel angle map (CAM) directly reflects the localized channel angle characteristics of a specific location, and can realize the effective reasoning and prediction of the channel angle based on the location information, thus helping to alleviate the pilot overhead problem caused by the acquisition of angle CSI. The traditional method of constructing a channel angle map requires the spatial continuity of the channel angle characteristics at different locations as a prerequisite, but actual measurements have found that there is not necessarily a strict one-to-one correspondence between the angle characteristics of multiple propagation paths measured at different locations, and the angle characteristics of the path cluster show a better correspondence. Therefore, it is necessary to design a channel angle map construction method based on the spatial continuity of the angle characteristics of the path cluster. Summary of the invention

[0004] Based on this, it is necessary to provide a channel angle map construction method, device, computer equipment and storage medium based on path clusters, which can simplify the calculation process to obtain the channel angle map and better realize the beam scanning, beam scheduling and beam codebook design of the transceiver.

[0005] In a first aspect, the present application provides a method for constructing a channel angle map based on a path cluster, comprising:

[0006] Step S1, mapping the communication coverage area in the target scene into a plane map, where the plane map includes a target number of grid point sets;

[0007] Step S2, selecting a preset number of sampling points from the grid point set of the plane map, and measuring multipath angle data of each sampling point, wherein the multipath angle data includes angle data corresponding to each path;

[0008] Step S3, performing cluster analysis on the multipath angle data at each sampling point and obtaining a path cluster belonging to each sampling point;

[0009] Step S4, extracting angle features based on the path cluster of each sampling point to obtain a multi-dimensional angle feature value of each sampling point;

[0010] Step S5, using a spatial interpolation algorithm to predict the multi-dimensional angular feature values ​​of non-sampling points in the grid point set based on the multi-dimensional angular feature values ​​of each sampling point;

[0011] Step S6: construct a channel angle map by combining the multi-dimensional angle feature quantities of the sampling points and the non-sampling points.

[0012] In one embodiment, the process before step S3 includes:

[0013] The angle data of each path in the multipath angle data is weighted and normalized.

[0014] In one embodiment, step S4 comprises:

[0015] Statistical analysis of path clusters is used to extract angle features of the path clusters of each sampling point, and multi-dimensional angle features of each sampling point are obtained, where the multi-dimensional angle features include the number of path clusters, the central angle of each path cluster, the angle extension, and the number of paths within the cluster.

[0016] In one embodiment, step S4 includes:

[0017] In each sampling point, each path cluster is arranged in ascending order according to the size of the center angle of the path cluster, and the corresponding relationship between the angle features of each sampling point is established according to the sorting result.

[0018] In one embodiment, step S5 comprises:

[0019] The inverse distance weighted interpolation algorithm based on sliding window is used to predict the multi-dimensional angular feature values ​​of non-sampling points in the grid point set.

[0020] In one embodiment, using an inverse distance weighted interpolation algorithm based on a sliding window to predict and obtain multi-dimensional angular feature quantities of non-sampling points in a grid point set includes:

[0021] Preset the size of the sliding window;

[0022] Taking the designated non-sampling point as the center, a sliding window frame is used to select sampling points that meet the preset distance from the designated non-sampling point;

[0023] Based on the selected sampling points, the inverse distance weighted interpolation algorithm is used to predict the angle information of the specified non-sampling points.

[0024] In one embodiment, step S6 further includes:

[0025] The resolution of the channel angle map is positively correlated with the density of sampling points. The higher the density of sampling points, the higher the spatial angle resolution of the constructed channel map.

[0026] In a second aspect, the present application also provides a channel angle map construction device based on a path cluster, comprising:

[0027] A map mapping module, used to map the communication coverage area in the target scene into a plane map, where the plane map includes a target number of grid point sets;

[0028] A multipath angle data measurement module is used to select a preset number of sampling points from a grid point set of a plane map, and measure multipath angle data of each sampling point, wherein the multipath angle data includes angle data corresponding to each path;

[0029] A cluster analysis module is used to perform cluster analysis on multipath angle data at each sampling point and obtain path clusters belonging to each sampling point;

[0030] An angle feature extraction module, used to extract angle features based on the path cluster of each sampling point to obtain a multi-dimensional angle feature value of each sampling point;

[0031] A spatial interpolation prediction module is used to predict the multi-dimensional angular feature quantity of non-sampling points in the grid point set using a spatial interpolation algorithm for the multi-dimensional angular feature quantity of each sampling point;

[0032] The channel angle map construction module is used to construct a channel angle map by combining the multi-dimensional angle feature quantities of sampling points and non-sampling points.

[0033] In a third aspect, the present application also provides a computer device, including a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of a method for constructing a channel angle map based on a path cluster when executing the computer program.

[0034] In a fourth aspect, the present application also provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of a method for constructing a channel angle map based on a path cluster.

[0035] The above-mentioned method, device, computer equipment and storage medium for constructing a channel angle map based on path clusters, in order to construct a channel angle map, firstly excavates the correspondence between the channel angle characteristics and the spatial position, and through in-depth analysis of the angle data obtained by sampling measurement, finds that the angle statistical characteristics of the path clusters obtained after multi-path clustering show the characteristic of spatially stable changes as the position moves. Based on the path cluster construction, a one-to-one correspondence between the angle characteristics and the displacement changes can be generated, thereby generating an angle channel map and simplifying the calculation process. The present application conducts cluster analysis on the different path angle characteristics contained in the same sampling point, obtains multi-dimensional angle feature quantities, including determination features and statistical features, and lays the foundation for angle space interpolation. At the same time, in order to construct a complete channel angle map using the angle information of a small number of known sampling points, the present application performs spatial interpolation on the basis of the sampling data, and predicts the angle data of the interpolation point, wherein the prediction point is the non-sampling point in the grid point set. Combined with the characteristics of the channel environment and utilizing the flexible focusing capability of the sliding window, an inverse distance weighted interpolation algorithm based on the sliding window is designed for spatial interpolation to achieve angle prediction. Based on the multi-dimensional angular features of known sampling points, a spatial interpolation algorithm is used to predict the remaining points. Then, the multi-dimensional angular features of the sampling points and the predicted points are combined to construct a complete channel angle map, where the predicted points are non-sampling points in the grid point set. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] Figure 1 A schematic diagram of a flow chart of a method for constructing a channel angle map based on a path cluster in one embodiment;

[0037] Figure 2 is a schematic diagram of a specific processing process of spatial interpolation in an embodiment;

[0038] Figure 3 A block diagram of the overall idea of ​​a method for constructing a channel angle map based on a path cluster in one embodiment;

[0039] Figure 4 It is a flowchart of a specific technical implementation method of a channel angle map construction method based on a path cluster in one embodiment;

[0040] Figure 5 A schematic diagram of a street coverage scenario in a city area according to an embodiment;

[0041] Figure 6 A schematic diagram of a distribution map of effective measurement points in a coverage area in one embodiment;

[0042] Figure 7 is a graph showing the variation of NMSE with sampling interval of two spatial interpolation algorithms in one embodiment;

[0043] Figure 8 A heat map of channel angle maps constructed in one embodiment. DETAILED DESCRIPTION

[0044] The future-oriented sixth generation mobile communication (6G) system will further enhance the exploitation of spatial resources. Massive multiple input multiple output (MIMO) is a key physical layer technology to enhance the spatial multiplexing capability of 6G. However, the application of massive MIMO communication systems requires a large amount of accurate angular channel state information (CSI) as support. Usually, angular CSI is obtained by channel estimation, but as the antenna scale continues to increase, the traditional channel estimation method based on pilot measurement will introduce greater pilot overhead, making the system unbearable.

[0045] The channel angle map (CAM) directly reflects the localized channel angle characteristics of a specific location, and can achieve effective reasoning and prediction of the channel angle based on location information, thus helping to alleviate the pilot overhead problem caused by angle CSI acquisition. The use of CAM helps to assist the antenna beam design of the transceiver in large-scale MIMO communication systems and reduce the channel estimation transmission overhead. To this end, it is necessary to study the construction method of the channel angle map.

[0046] Since the channel angle characteristics have spatial continuity in the local area, we can first select a small number of sample points based on the spatial position to measure the angle data, then use the spatial interpolation method to predict the angle data of the remaining points, and finally combine the angle data of the sampling points and the predicted points to jointly construct a complete channel angle map. This method requires the spatial continuity of the channel angle characteristics at different locations as a premise, but actual measurements have found that this spatial continuity in multipath propagation scenarios is not based on the path, but on the path cluster. In other words, there is not necessarily a strict one-to-one correspondence between the angle characteristics of multiple propagation paths measured at different locations, so it is necessary to design a channel angle map construction method based on the spatial continuity of the angle characteristics of the path cluster.

[0047] The purpose of this application is to extract and use the channel angle characteristics of known sampling points to predict the angle characteristics of the remaining points, so as to jointly construct a channel angle map based on path clusters. The channel angle map can be used to quickly query and reason about local channel angle characteristics, guide beam codebook design, optimize beam scheduling, and reduce the acquisition overhead of angle CSI, thereby improving system transmission efficiency.

[0048] In the communication coverage area, the present application predicts the angle information of non-sampling points (points to be interpolated) based on the location information of the divided grid points and the multipath angle information of the known sampling points, thereby constructing a complete channel angle map.

[0049] 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.

[0050] In an exemplary embodiment, Figure 1 As shown, a method for constructing a channel angle map based on a path cluster is provided, comprising:

[0051] Step S1, mapping the communication coverage area in the target scene into a plane map, where the plane map includes a target number of grid point sets;

[0052] Step S2, selecting a preset number of sampling points from the grid point set of the plane map, and measuring multipath angle data of each sampling point, wherein the multipath angle data includes angle data corresponding to each path;

[0053] Step S3, performing cluster analysis on the multipath angle data at each sampling point and obtaining a path cluster belonging to each sampling point;

[0054] Step S4, extracting angle features based on the path cluster of each sampling point to obtain a multi-dimensional angle feature value of each sampling point;

[0055] Step S5, using a spatial interpolation algorithm to predict the multi-dimensional angular feature values ​​of non-sampling points in the grid point set based on the multi-dimensional angular feature values ​​of each sampling point;

[0056] Exemplarily, the sampling methods include random sampling and uniform spatial sampling.

[0057] Step S6: construct a channel angle map by combining the multi-dimensional angle feature quantities of the sampling points and the non-sampling points.

[0058] In this embodiment, a small number of sampling points are selected within the communication coverage area to measure angle data and obtain multi-path angle data. Through the analysis of the measured angle data, it can be seen that the angle characteristics of multiple propagation paths within the sampling points do not change significantly with the displacement, and there is not necessarily a strict one-to-one correspondence between the angle characteristics of multiple propagation paths measured by sampling points at different positions. However, the path clusters obtained by path clustering have the characteristic of continuous change in angle characteristics. Therefore, the statistical characteristics of the path clusters can be used to generate multi-dimensional angle feature quantities, and spatial interpolation prediction can be performed based on this. The multi-dimensional angle feature quantities here can include the number of path clusters, the angle center of each cluster, the angle extension, the number of paths within the cluster, etc.

[0059] In an exemplary embodiment, step S3 includes:

[0060] The angle data of each path in the multipath angle data is weighted and normalized.

[0061] Specifically, before multi-path clustering is performed to obtain a path cluster, the various angle data of each path need to be weighted and normalized to avoid weakening the influence of other angle data due to a certain angle value being too large.

[0062] For example, the azimuth angle θ and the elevation angle φ are first normalized to obtain the equations (1) and (2):

[0063] θ norm =(θ-θ min ) / (θ max -θ min ) (1)

[0064] φ norm =(φ-φ min ) / (φ max -φ min ) (2)

[0065] Among them, θ max and θ min represent the maximum and minimum azimuth angles in multiple paths, respectively, and φ max and φ min Represent the maximum and minimum elevation angles respectively; then the normalized azimuth angle θ norm Multiplying by the weighting factor α, the normalized elevation angle φ norm Multiply by the weighting factor β and use the weighted value αθ norm and βφ norm As two independent dimensions, they participate in clustering. Based on the results of weighted normalization, the joint angle feature of each path can be expressed as shown in formula (3):

[0066] γ=(αθ norm ,βφ norm ) (3)

[0067] In an exemplary embodiment, step S4 includes:

[0068] Statistical analysis of path clusters is used to extract angle features of the path clusters of each sampling point, and multi-dimensional angle features of each sampling point are obtained, where the multi-dimensional angle features include the number of path clusters, the central angle of each path cluster, the angle extension, and the number of paths within the cluster.

[0069] Specifically, spatial clustering is performed on the joint angle feature γ of each path within the sampling point to obtain multiple path clusters belonging to the sampling point. Then, statistical analysis of the path clusters can be used to extract angle features including the number of path clusters, the central angle of each path cluster (average), angle spread (standard deviation), the number of paths within the cluster, and other information to obtain multi-dimensional angle features of each sampling point.

[0070] In an exemplary embodiment, step S4 includes:

[0071] In each sampling point, each path cluster is arranged in ascending order according to the size of the center angle of the path cluster, and the corresponding relationship between the angle features of each sampling point is established according to the sorting result.

[0072] Specifically, in order to further explore the corresponding relationship of angle information between sampling points at different locations, it is necessary to realize the alignment of path clusters between sampling points. Specifically, firstly, each path cluster is sorted in ascending order according to the size of its central angle in each spatial sampling point, and then a one-to-one correspondence between the angle features of each point is established based on the sorting results.

[0073] Furthermore, the number of path clusters, cluster center angle, angle extension, number of paths within the cluster and other statistical features after the path clusters are aligned are used as the angle feature ρ of the path cluster based on the sampling point, and combined with the corresponding sampling point position information, the angle feature ρ belonging to any point p can be constructed. x,y Multidimensional angular feature Where L x,y Represents point p x,y The number of path clusters included. Multi-dimensional angular feature of sampling points It lays the data foundation for subsequent spatial interpolation prediction.

[0074] In an exemplary embodiment, step S5 includes:

[0075] The inverse distance weighted interpolation algorithm based on sliding window is used to predict the multi-dimensional angular feature values ​​of non-sampling points in the grid point set.

[0076] In an exemplary embodiment, using an inverse distance weighted interpolation algorithm based on a sliding window to predict and obtain multi-dimensional angular feature quantities of non-sampling points in a grid point set includes:

[0077] Preset the size of the sliding window;

[0078] Taking the designated non-sampling point as the center, a sliding window frame is used to select sampling points that meet the preset distance from the designated non-sampling point;

[0079] Based on the selected sampling points, the inverse distance weighted interpolation algorithm is used to predict the angle information of the specified non-sampling points.

[0080] Specifically, in obtaining the multi-dimensional angle feature quantity based on the path cluster After that, the angle information of the known sampling points is used for spatial interpolation to predict the angle information of the remaining points to be measured, that is, the multi-dimensional angle feature of each sampling point is predicted by using the spatial interpolation algorithm for the multi-dimensional angle feature of the non-sampling points in the grid point set. By integrating the angle information of all the grid points in the coverage area, a complete channel angle map based on the path cluster can be constructed, that is, the channel angle map is constructed by combining the multi-dimensional angle feature of the sampling points and the non-sampling points, as shown in formula (4):

[0081]

[0082] Where M and N represent the number of rows and columns for constructing the channel angle map, respectively. The angle channel map can reflect the statistical information of all channel angle characteristics based on path clusters contained in each grid point, including the number of path clusters, cluster center angle, angle extension, number of paths within the cluster, etc.

[0083] Due to the slow change of the spatial propagation characteristics of the channel, the angle information contained in the grid points with similar positions on the map also shows a steady change trend. By using the positional relationship between the known sampling points and the unknown interpolation points, combined with the known angle information of the sampling points, spatial interpolation can be performed to obtain the angle information of the interpolation points. Spatial interpolation prediction can be mainly divided into two categories, one is interpolation prediction based on the overall space, and the other is interpolation prediction based on the local area. Since the grid points are discretely distributed and the influence of the sampling points on the interpolation points is inversely proportional to the distance between the two points, the inverse distance weighting (IDW) in the classic algorithm is a more suitable spatial interpolation method under the scene model constructed by the channel map. This method determines the value of the interpolation point by calculating the inverse of the distance between the interpolation point and the sampling point. The influence of the sampling point on the interpolation result decreases as the distance increases. The inverse distance weighted interpolation method is based on the overall spatial sampling points. Although the algorithm is simple, it requires a large amount of calculation because it needs to calculate the distance between the interpolation point and all known sampling points, especially when there are a large number of spatial sampling points. In addition, although the angle feature has the characteristic of slow change in a small range, it also has the characteristic of jumping in a larger range. Therefore, sampling points beyond a certain distance range not only cannot provide information assistance, but may cause negative effects. Therefore, selecting too many sampling points not only increases the computational complexity, but also makes it difficult to ensure the accuracy of the prediction results.

[0084] In order to improve the accuracy of spatial interpolation, the inverse distance weighted interpolation method is improved, and a sliding window-based inverse distance weighted (SW-IDW) interpolation algorithm is proposed. First, the size of the rectangular sliding window is set; second, with the specified interpolation point as the center, a small number of sampling points close to the point are selected using the sliding window frame; then, based on the selected sampling points, the inverse distance weighted interpolation algorithm is used to predict the angle information of the target interpolation point.

[0085] Among them, the size of the sliding window is set according to the number of sampling points that need to be selected. Taking the uniform sampling of grid points on the map as an example, the more sampling points are required to be selected, the larger the sliding window size is, but a sliding window that is too large is prone to introduce sampling points with angle feature jumps; conversely, the fewer sampling points are required to be selected, the smaller the sliding window size is, but a sliding window that is too small will lead to insufficient prediction ability of the sampling points, which will seriously affect the angle prediction accuracy of the interpolation points. Usually, setting the number of sampling points required to be selected to 9 can achieve a better spatial interpolation effect. At this time, the side length a of the sliding window can be correspondingly set to 3 times the sampling interval Δs, that is, a=3Δs.

[0086] Specifically, the sliding window frame is used to select a small number of sampling points that are close to the point. Taking the uniform sampling of grid points on the map as an example, for any point p to be interpolated (that is, the prediction point, non-sampling point in the grid point set), first find the sampling point q0 with the closest Euclidean distance to it, and then find the 8-neighborhood sampling point set {q1, q2, ..., q8} of q0. Then, the above 9 sampling points are taken as sampling points that are close to p, and based on this, the angle prediction of the point to be interpolated is realized, that is, the angle prediction of the non-sampling point in the grid point set.

[0087] Furthermore, the specific processing process of spatial interpolation is as follows Figure 2 As shown, first input the sampling point angle data set Φ s , the number of sampling points N to be selected; set the sliding window side length a according to N; calculate the sampling point q0 that is closest to the point to be interpolated p; find the 8-neighborhood sampling point set {q1, q2, ..., q8} of the sampling point q0; from Φ s Find the multi-dimensional angular feature quantity corresponding to {q1, q2, ..., q8} based on Use the IDW algorithm to find the multi-dimensional angle feature value corresponding to point p Final output

[0088] The formula involved is as follows:

[0089] When calculating the side length a of the sliding window, assuming that the uniform sampling interval on the map is Δs and the number of sampling points to be selected is N, the side length a of the sliding window should satisfy the following formula (5):

[0090]

[0091] When the central sampling point q0 is found, the position coordinates of any point on the channel map R can be expressed as c; first calculate the distance d from the interpolation point p to all sampling points q q , as shown in formula (6):

[0092] d q =||c p -c q ‖,q∈R (6)

[0093] Then find the sampling point q0 corresponding to the minimum distance, as shown in formula (7):

[0094] q0=argmin q (d q ) (7)

[0095] Next, when finding the 8-neighborhood sampling points, the 8-neighborhood definition of pixel points in the field of digital image processing is used to determine the positions of the 8-neighborhood sampling points. Assuming that the coordinates of q0 are (x, y) and the sampling interval is Δs, the coordinates of the 8-neighborhood sampling points are (x-Δs, y-Δs), (x, y-Δs), (x+Δs, y-Δs), (x-Δs, y), (x+Δs, y), (x-Δs, y+Δs), (x, y+Δs), and (x+Δs, y+Δs).

[0096] When implementing IDW spatial interpolation, it is assumed that the angle data corresponding to the above 9 sampling points are The distances from the interpolation point p are Then the angle data corresponding to point p can be calculated according to the IDW algorithm, as shown in formula (8):

[0097]

[0098] In an exemplary embodiment, step S6 further includes:

[0099] The resolution of the channel angle map is positively correlated with the density of sampling points. The higher the density of sampling points, the higher the spatial angle resolution of the constructed channel map.

[0100] Specifically, the resolution of the channel angle map is related to the density of the sampling points. The higher the spatial sampling density, the higher the spatial angle resolution of the constructed channel map. In the process of constructing the channel angle map, low-resolution sampling can be achieved through sparse uniform spatial sampling, that is, a sampling point is selected every fixed number of grid points, and a relatively coarse channel angle map is formed by the selected sampling points. Then, spatial interpolation is performed on the basis of the coarse angle channel map to obtain a detailed and complete channel angle map.

[0101] In one embodiment, the present application proposes a general idea and implementation method for constructing a channel angle map, which specifically includes channel angle feature extraction and spatial interpolation prediction. The channel feature extraction includes deterministic angle feature extraction in the original measurement data and statistical angle feature extraction after multi-path clustering; the spatial interpolation prediction includes two types of methods based on overall spatial interpolation and sliding window interpolation. The overall idea of ​​the channel angle map construction method based on path cluster is shown in the following figure. Figure 3 As shown, it includes sampling point measurement, angle feature extraction, interpolation point prediction, and target CAM construction, where the interpolation point is the non-sampling point in the grid point set.

[0102] Furthermore, the specific technical implementation method of the channel angle map construction method based on the path cluster is as follows: Figure 4 As shown in the figure. First, the communication coverage area in the real scene of the target is mapped into a plane map; second, a small number of sampling points are selected from the map to measure the multipath angle data in the channel; third, the measured multipath angle data is clustered and analyzed in each sampling point to obtain the path cluster; then, the multidimensional angle feature is calculated based on the path cluster; then, the spatial interpolation algorithm is used to predict the remaining points based on the multidimensional angle feature; finally, the multidimensional angle feature of the sampling point and the predicted point are combined to jointly construct a complete channel angle map. The remaining points and the predicted points are non-sampling points in the grid point set.

[0103] In a specific embodiment, firstly, the urban street communication coverage scenario is established by Ray-tracing, and the corresponding channel characteristics are collected to construct the angle data set. The database is created by using the WallMan module of WinProp software, and basic parameters such as wall height and building materials are set. Then, the 3D urban street propagation model under the real target scenario is drawn in proportion using the obstacle type components given in the software. The urban street coverage scenario is as follows: Figure 5 shown.

[0104] Figure 5 The effective coverage of the urban street scene shown is 632×470m 2, the measurement height of the sampling point is fixed at 1.5m. In this scenario, the Ray-tracing technology is used to obtain the angle-related information of each grid point in advance to simulate the true value of the angle data actually measured. According to the angle data measured at the known sampling points, the angle information of the remaining points is predicted by spatial interpolation to construct the channel angle map. In addition, the normalized mean square error (NMSE) is used as the performance evaluation indicator to simulate and verify the accuracy of the constructed map.

[0105] Due to the influence of building shielding, the effective receiving signal cannot be directly measured in the area where the building is located. Therefore, when constructing the channel angle map, it is necessary to Figure 5 Based on the street scene shown in FIG. 1 , the area occupied by the buildings is removed to obtain a distribution map of effective measurement points. The distribution map of effective measurement points in the coverage area is shown in FIG. Figure 6 The figure shows a flat map. The blank area in the figure indicates the area where the measurable grid points are located, and the gray area indicates the area occupied by the building that needs to be removed. Spatial sampling interpolation and channel angle map construction can be completed based on the blank area.

[0106] During the simulation, the row and column sampling intervals of the channel map were set to 3, 5, 7, 9, and 11 m, respectively, and the corresponding sliding window sizes were set to 9 × 9, 15 × 15, 21 × 21, 27 × 27, and 33 × 33 m. 2 Based on uniform spatial sampling, the NMSE performance curves of the channel angle maps obtained by the two spatial interpolation methods are shown in Figure 2. Figure 7 As shown in the figure, under different sampling intervals, the channel angle maps obtained by sampling the two spatial interpolation methods both meet the lower NMSE index, and the performance of the sliding window-based inverse distance weighted interpolation method is better than that of the inverse distance weighted interpolation method. And the angle prediction error will decrease as the sampling interval continues to decrease.

[0107] The angle data predicted by the sliding window inverse distance weighted interpolation method is combined with the angle data measured at the sampling point to construct a heat map of the channel angle map. Figure 8 The channel angle map can be used to quickly query and infer local channel angle characteristics, guide beam codebook design, optimize beam scheduling, and reduce the acquisition overhead of angle CSI, thereby improving system transmission efficiency.

[0108] This application proposes an overall idea and implementation method for constructing an angle channel map, and designs key algorithms for channel angle feature extraction and angle space interpolation prediction. First, the channel angle features are extracted from a small amount of sampled measurement data; secondly, the angle features are clustered to obtain path clusters; then, multi-dimensional angle feature quantities of sampling points are constructed based on the path clusters; then, the multi-dimensional angle feature quantities of known sampling points are used to spatially interpolate the angle information of the estimated points, thereby obtaining a complete channel angle map. Utilizing the angle channel map, beam scanning, beam scheduling, and beam codebook design at the transceiver can be better implemented. While achieving rapid acquisition of channel angle information, the channel estimation overhead is greatly reduced, and the overall transmission efficiency is improved, thereby alleviating the bandwidth pressure introduced by large-dimensional pilot transmission in large-scale MIMO technology application scenarios, and strongly supporting low-complexity transmission applications for multi-antenna systems for future 6G.

[0109] At least a portion of the steps in the flowcharts involved in the embodiments described above may include multiple steps or multiple stages. These steps or stages may be executed at different times, or may be executed in turn or alternation with other steps or at least a portion of the steps or stages in other steps.

[0110] Based on the same inventive concept, the embodiment of the present application also provides a channel angle map construction device based on a path cluster for implementing the channel angle map construction method based on a path cluster involved above. The implementation scheme for solving the problem provided by the device is similar to the implementation scheme recorded in the above method, so the specific limitations in one or more embodiments of the channel angle map construction device based on a path cluster provided below can refer to the limitations of the channel angle map construction method based on a path cluster above, and will not be repeated here.

[0111] In an exemplary embodiment, a channel angle map construction device based on a path cluster is provided, comprising:

[0112] A map mapping module, used to map the communication coverage area in the target scene into a plane map, where the plane map includes a target number of grid point sets;

[0113] A multipath angle data measurement module is used to select a preset number of sampling points from a grid point set of a plane map, and measure multipath angle data of each sampling point, wherein the multipath angle data includes angle data corresponding to each path;

[0114] A cluster analysis module is used to perform cluster analysis on multipath angle data at each sampling point and obtain the path clusters belonging to each sampling point;

[0115] An angle feature extraction module, used to extract angle features based on the path cluster of each sampling point to obtain a multi-dimensional angle feature value of each sampling point;

[0116] A spatial interpolation prediction module is used to predict the multi-dimensional angular feature quantity of non-sampling points in the grid point set by using a spatial interpolation algorithm for the multi-dimensional angular feature quantity of each sampling point;

[0117] The channel angle map construction module is used to construct a channel angle map by combining the multi-dimensional angle feature quantities of sampling points and non-sampling points.

[0118] Each module in the above-mentioned channel angle map construction device based on path cluster can be implemented in whole or in part by software, hardware and their combination. Each of the above-mentioned modules can be embedded in or independent of the processor in the computer device in the form of hardware, or can be stored in the memory in the computer device in the form of software, so that the processor can call and execute the operations corresponding to each of the above modules.

[0119] In one embodiment, a computer device is further provided, including a memory and a processor, wherein a computer program is stored in the memory, and the processor implements the steps in the above method embodiments when executing the computer program.

[0120] In one embodiment, a computer-readable storage medium is provided, on which a computer program is stored. When the computer program is executed by a processor, the steps in the above-mentioned method embodiments are implemented.

[0121] For those skilled in the art, several modifications and improvements can be made without departing from the concept of the present application, and these modifications and improvements are all within the scope of protection of the present application. Therefore, the scope of protection of the present application shall be subject to the attached claims.

Claims

1. A method for constructing a channel angle map based on a path cluster, characterized in that: The method comprises: Step S1, mapping the communication coverage area in the target scene into a plane map, where the plane map includes a target number of grid point sets; Step S2, selecting a preset number of sampling points from the grid point set of the plane map, and measuring multipath angle data of each sampling point, wherein the multipath angle data includes angle data corresponding to each path; Step S3, performing cluster analysis on the multipath angle data at each sampling point and obtaining a path cluster belonging to each sampling point; Step S4, extracting angle features based on the path cluster of each sampling point to obtain a multi-dimensional angle feature value of each sampling point; Step S5, using a spatial interpolation algorithm to predict the multi-dimensional angular feature values ​​of non-sampling points in the grid point set based on the multi-dimensional angular feature values ​​of each sampling point; Step S6: construct a channel angle map by combining the multi-dimensional angle feature quantities of the sampling points and the non-sampling points.

2. The method according to claim 1, characterized in that The step S3 includes: The angle data of each path in the multipath angle data is weighted and normalized.

3. The method according to claim 2, characterized in that The step S4 comprises: Statistical analysis of path clusters is used to extract angle features of the path clusters of each sampling point, and multi-dimensional angle features of each sampling point are obtained, where the multi-dimensional angle features include the number of path clusters, the central angle of each path cluster, the angle extension, and the number of paths within the cluster.

4. The method according to claim 3, characterized in that The step S4 then includes: In each sampling point, each path cluster is arranged in ascending order according to the size of the center angle of the path cluster, and the corresponding relationship between the angle features of each sampling point is established according to the sorting result.

5. The method according to claim 4, characterized in that The step S5 comprises: The inverse distance weighted interpolation algorithm based on sliding window is used to predict the multi-dimensional angular feature values ​​of non-sampling points in the grid point set.

6. The method according to claim 5, characterized in that The method of using the sliding window-based inverse distance weighted interpolation algorithm to predict and obtain the multi-dimensional angular feature quantity of the non-sampling points in the grid point set includes: Preset the size of the sliding window; Taking the designated non-sampling point as the center, a sliding window frame is used to select sampling points that meet the preset distance from the designated non-sampling point; Based on the selected sampling points, the inverse distance weighted interpolation algorithm is used to predict the angle information of the specified non-sampling points.

7. The method according to claim 6, characterized in that The step S6 further comprises: The resolution of the channel angle map is positively correlated with the density of sampling points. The higher the density of sampling points, the higher the spatial angle resolution of the constructed channel map.

8. A channel angle map construction device based on path cluster, characterized in that: The device comprises: A map mapping module, used to map the communication coverage area in the target scene into a plane map, where the plane map includes a target number of grid point sets; A multipath angle data measurement module is used to select a preset number of sampling points from a grid point set of a plane map, and measure multipath angle data of each sampling point, wherein the multipath angle data includes angle data corresponding to each path; A cluster analysis module is used to perform cluster analysis on multipath angle data at each sampling point and obtain the path clusters belonging to each sampling point; An angle feature extraction module, used to extract angle features based on the path cluster of each sampling point to obtain a multi-dimensional angle feature value of each sampling point; A spatial interpolation prediction module is used to predict the multi-dimensional angular feature quantity of non-sampling points in the grid point set by using a spatial interpolation algorithm for the multi-dimensional angular feature quantity of each sampling point; The channel angle map construction module is used to construct a channel angle map by combining the multi-dimensional angle feature quantities of sampling points and non-sampling points.

9. A computer device comprising a memory and a processor, wherein the memory stores a computer program, wherein: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 7 are implemented.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the steps of the method according to any one of claims 1 to 7 are implemented.

Images