Low-altitude networking method, device, equipment, storage medium and program product

Abstract

The invention relates to the technical field of low-altitude networking, and provides a low-altitude networking method and device, equipment, a storage medium and a program product, and the method comprises the steps: determining a candidate site set according to a multi-dimensional data set of a to-be-planned low-altitude region; the multi-dimensional data set comprises work parameter data of the base station and an air route of the low-altitude area; constructing an evaluation area of each candidate base station according to the candidate station set, and determining a preferred object set of the candidate base stations according to the evaluation areas; extracting spatial features of each preferred object in the preferred object set, and identifying a backbone site according to the spatial features; and carrying out networking on the low-altitude area based on the backbone station. Through quantization data and computational geometry of the low-altitude area and in combination with spatial features of the base station, the backbone station can be efficiently and accurately identified, networking of the low-altitude area is realized, and the networking efficiency of the low-altitude private network is improved.

Description

Technical Field

[0001] The present application relates to the technical field of low-altitude networking, and in particular, to a low-altitude networking method, device, equipment, storage medium and program product. Background Art

[0002] With the development of the low-altitude field, the requirements for low-altitude communication and sensing capabilities are gradually increasing. However, the low-altitude airspace is a brand-new signal coverage scenario, facing problems such as high networking costs, high low-altitude overlapping coverage, and strong interference. A backbone site refers to a site with a wide coverage range and good quality that can play a network support role. When networking a low-altitude private network, a site with a relatively high antenna height (a tower site or a site on a tall building) and less obstruction to the sky is used as the backbone site for low-altitude private network networking to provide a wider airspace coverage. To ensure ground coverage, after these sites perform low-altitude coverage, there are other sites within the coverage range that can supplement the coverage area of these sites, and the impact on ground coverage is controlled within the threshold.

[0003] On the one hand, in terms of airspace coverage, due to the high antenna height and less obstruction of the higher sites, a larger airspace can be covered, and the airspace overlapping and interference are reduced because the airspace coverage capabilities of other lower sites are weak; on the other hand, in terms of ground coverage, due to the dense ground base stations, while the higher sites provide ground coverage, they also bring higher overlapping coverage and co-channel interference. Using the higher sites for airspace coverage can also reduce ground interference.

[0004] Currently, the selection of backbone sites for low-altitude networking mainly relies on expert experience for manual judgment based on antenna height. Affected by individual cognitive differences, expert experience has subjective biases and lacks a unified quantification standard, resulting in inconsistent site selection; when the site scale is large, the time-consuming for manual screening in the face of a large-scale network grows exponentially. Therefore, there are problems of low efficiency and accuracy. Summary of the Invention

[0005] The present application provides a low-altitude networking method, device, equipment, storage medium and program product (hereinafter referred to as), to solve the defect that in the prior art, low-altitude networking relies on expert experience to manually select backbone sites, resulting in low efficiency and accuracy, and to achieve accurate and efficient identification of backbone sites.

[0006] The present application provides a low-altitude networking method, including: Determining a candidate site set according to a multi-dimensional data set of a to-be-planned low-altitude area; the multi-dimensional data set includes base station engineering parameter data and flight routes of the low-altitude area; Constructing an evaluation area for each candidate base station according to the candidate site set, and determining a set of preferred objects of the candidate base stations according to the evaluation area; Extracting spatial features of each preferred object in the set of preferred objects, and identifying backbone sites according to the spatial features; Network the low-altitude area based on the backbone sites.

[0007] In one embodiment, constructing the evaluation area of each candidate base station according to the candidate site set includes: Determine the coverage distance of the candidate base station according to the base station engineering parameter data of each candidate base station in the candidate site set; Taking the candidate base station as the center and the coverage distance as the radius, generate a circular coverage circle; Taking each candidate base station within the circular coverage circle as a node, use the triangulation algorithm to construct a triangular grid corresponding to the candidate base station; Based on the adjacency relationship of the triangular grid, determine the adjacent base stations of the candidate base station; Generate sectors according to the base station density of the adjacent base stations, screen out the key base stations in each sector as boundary points, and construct the evaluation area of the candidate base station.

[0008] In one embodiment, the base station engineering parameter data includes a coverage scenario; identifying the backbone sites according to the spatial features includes: Obtain the site recognition model corresponding to the coverage scenario of each candidate base station in the candidate site set; Input the spatial features into the site recognition model, classify each candidate base station, and obtain a candidate set of backbone sites; Iteratively optimize the candidate set to obtain the backbone sites.

[0009] In one embodiment, the multi-dimensional data further includes measurement report data; iteratively optimizing the candidate set to obtain the backbone sites includes: Based on each candidate backbone site in the candidate set, conduct a flight test on the low-altitude area according to the flight route to determine whether there are coverage holes in the low-altitude area; If there are coverage holes and the coverage holes are within the coverage range of the candidate backbone site and it, obtain the sub-optimal base station of the candidate backbone site as the backbone site to optimize the candidate set; After removing the sampling points of each backbone site in the candidate set from the measurement report data, evaluate the ground coverage rate of each base station; Mark the backbone base stations affecting ground coverage according to the ground coverage rate; Update the candidate site set according to the candidate set, return and execute the step of constructing the evaluation area of each candidate base station according to the candidate site set, until there are no coverage holes in the low-altitude area, and obtain each backbone site in the candidate set.

[0010] In one embodiment, before obtaining the site recognition models corresponding to the coverage scenarios of the candidate base stations in the candidate site set, the following steps are further included: Obtain an original data set; the original data set includes the antenna hanging height of sample base stations, the coverage scenario, and the height information of geographical objects; Based on the distance between the geographical object and the receiving point of the sample base station and the height information, construct a spatial influence factor; Compare the height information with the antenna hanging height, and generate a label for the spatial influence factor according to the comparison result; Based on the label and the spatial influence factor, construct a sample data set, and divide the sample data set according to the coverage scenario to obtain sample data subsets; Use each sample data subset to iteratively train a preset support vector machine model to obtain the site recognition models corresponding to the respective coverage scenarios.

[0011] In one embodiment, the base station engineering parameter data includes the base station location; the determining of the candidate site set according to the multi-dimensional data set of the low-altitude area to be planned includes: Establish a plurality of buffer zones on both sides of the flight path of the low-altitude area to be planned with a preset width; Deduplicate and merge the plurality of buffer zones to obtain a base station selection area; According to the base station locations of the base stations in the base station engineering parameter data, obtain each candidate base station within the base station selection area to obtain a candidate site set.

[0012] This application also provides a low-altitude networking device, including the following modules: A base station selection module, configured to determine a candidate site set according to the multi-dimensional data set of the low-altitude area to be planned; the multi-dimensional data set includes base station engineering parameter data and the flight path of the low-altitude area; A region construction module, configured to construct an evaluation region for each candidate base station according to the candidate site set, and determine a set of preferred objects for the candidate base stations according to the evaluation region; A site recognition module, configured to extract the spatial features of each preferred object in the set of preferred objects, and identify backbone sites according to the spatial features; A low-altitude networking module, configured to perform networking on the low-altitude area based on the backbone sites.

[0013] This application also provides an electronic device, including a memory, a processor, and a computer program stored on the memory and executable on the processor, where when the processor executes the computer program, it implements the low-altitude networking method as described in any one of the above.

[0014] The present application also provides a non-transitory computer-readable storage medium, on which a computer program is stored. When the computer program is executed by a processor, the low-altitude networking method described in any one of the above is implemented.

[0015] The present application also provides a computer program product, including a computer program. When the computer program is executed by a processor, the low-altitude networking method described in any one of the above is implemented.

[0016] The low-altitude networking method, device, equipment, storage medium and program product provided by the present application determine candidate base stations through a multi-dimensional data set in the low-altitude area, construct an evaluation area of the candidate base stations through computational geometry, extract the spatial features of the optimal objects in the evaluation area, identify the backbone sites according to the spatial features of the optimal objects, and network the low-altitude area. Through the quantitative data and computational geometry of the low-altitude area, combined with the spatial features of the base stations, the backbone sites can be efficiently and accurately identified, the networking of the low-altitude area can be realized, and the networking efficiency of the low-altitude private network is improved. BRIEF DESCRIPTION OF THE DRAWINGS

[0017] In order to more clearly illustrate the technical solutions in the present application or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are some embodiments of the present application. For those of ordinary skill in the art, without creative efforts, other drawings can also be obtained based on these drawings.

[0018] Figure 1 is a schematic flowchart of the low-altitude networking method provided by an embodiment of the present application.

[0019] Figure 2 is a schematic structural diagram of the low-altitude networking device provided by the present invention.

[0020] Figure 3 is a schematic structural diagram of the electronic device provided by an embodiment of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0021] To make the objectives, technical solutions and advantages of the present application clearer, the following will clearly and completely describe the technical solutions in the present application with reference to the drawings in the present application. Obviously, the described embodiments are some, but not all, of the embodiments of the present application. All other embodiments obtained by those of ordinary skill in the art without creative efforts based on the embodiments in the present application belong to the scope of protection of the present application.

[0022] An embodiment of the present application provides a low-altitude networking method, which can accurately and efficiently identify backbone sites in multiple scenarios through a dynamic decision-making method of space-propagation-topology, and realize networking of the low-altitude airspace. Specifically, Figure 1It is a schematic flowchart of the low-altitude networking method provided by an embodiment of the present application. As Figure 1 shown, the low-altitude networking method includes the following steps: Step 100: Determine a candidate site set according to the multi-dimensional data set of the to-be-planned low-altitude area; the multi-dimensional data set includes base station engineering parameter data and the flight routes of the low-altitude area; Step 200: Construct an evaluation area for each candidate base station according to the candidate site set, and determine a preferred object set of the candidate base stations according to the evaluation area; Step 300: Extract the spatial features of each preferred object in the preferred object set, and identify the backbone sites according to the spatial features; Step 400: Network the low-altitude area based on the backbone sites.

[0023] According to the multi-dimensional data set of the to-be-planned low-altitude area, determine a candidate site set. Among them, the multi-dimensional data set includes the engineering parameter data of each base station and the flight routes planned for the low-altitude area. Optionally, the engineering parameter data of the base station specifically includes the base station name, base station identifier (base station ID), longitude and latitude (base station location), coverage type, antenna hanging height, antenna azimuth angle, scene type of the coverage scenario, transmit power, and minimum access level, etc. Further, the coverage scenarios include multiple scenarios such as low-altitude + urban cruise, low-altitude + fire fighting, and low-altitude + agriculture.

[0024] Optionally, determine one or more candidate base stations through judgment based on the base station engineering parameter data to obtain a candidate site set. In one embodiment, according to the base station engineering parameters in the multi-dimensional data set, determine the positions of each base station, and determine one or more candidate base stations as candidate sites according to the base station positions to obtain a candidate site set.

[0025] Construct an evaluation area for each candidate base station according to the determined candidate site set. Optionally, construct the evaluation areas of each candidate base station in the candidate site set through methods such as computational geometry, and determine a preferred object set of the candidate base stations according to the evaluation area. Among them, for any candidate base station in the candidate site set, the purpose of constructing the evaluation area is to determine the maximum reasonable coverage area of each base station under the existing network base station layout as the preferred range for comparison with other objects, and it is also the range for evaluating the impact of the base station's ground coverage. The actual coverage of the base station is affected by factors such as parameter configuration and the surrounding environment, and it can be larger or smaller than this area.

[0026] Optionally, the multi-dimensional dataset further includes the feature information and distribution of geographical objects. The geographical objects include buildings such as buildings. The feature information of the geographical objects includes object location, object shape outline, object type, object name, object height, etc. Among them, the distribution of geographical objects is determined according to the object locations of each geographical object. The set of preferred objects for candidate base stations is composed of other base stations and geographical objects within the evaluation area.

[0027] Extract the spatial features of each preferred object in the set of preferred objects. The preferred objects include base stations and geographical objects. The spatial features of the preferred objects include height features. For the geographical objects in the preferred objects, their height features are determined according to the height of the geographical objects. For the base stations in the preferred objects, their height features are determined according to the antenna hanging height.

[0028] Optionally, the extracted spatial features of the preferred objects may further include the relative distance between different preferred objects. According to the extracted spatial features of each preferred object, the backbone sites are identified. In one embodiment, according to the antenna hanging height of each candidate base station in the preferred objects, combined with the height of the surrounding geographical objects, the backbone sites are determined. Exemplarily, the higher the antenna hanging height of the candidate base station, the lower the surrounding geographical objects or the farther the distance from the candidate base station, the greater the possibility of being a backbone site.

[0029] Furthermore, the identified backbone sites include one or more. Based on the identified backbone sites, the low-altitude area is networked to achieve airspace coverage of the low-altitude area.

[0030] In this embodiment, candidate base stations are determined through the multi-dimensional dataset of the low-altitude area, and the evaluation area of the candidate base stations is constructed through computational geometry. The spatial features of the preferred objects in the evaluation area are extracted, and the backbone sites are identified according to the spatial features of the preferred objects, and the low-altitude area is networked. Through the quantitative data of the low-altitude area and computational geometry, combined with the spatial features of the base stations, the backbone sites can be identified efficiently and accurately, the networking of the low-altitude area can be realized, and the networking efficiency of the low-altitude private network is improved.

[0031] In one embodiment, the base station engineering parameter data in the multi-dimensional dataset of the low-altitude area includes the base station location. In step 100, determining the candidate site set according to the multi-dimensional dataset of the to-be-planned low-altitude area includes: Step 101, establish a plurality of buffer zones on both sides of the flight path of the to-be-planned low-altitude area with a preset width; Step 102, de-duplicate and merge the plurality of buffer zones to obtain a base station selection area; Step 103, according to the base station locations of each base station in the base station engineering parameter data, obtain each candidate base station within the base station selection area to obtain a candidate site set.

[0032] Based on the flight route in the low-altitude area plan, multiple buffer zones are established on both sides of the flight route with a preset width D. There may be overlapping areas among the multiple buffer zones. The multiple buffer zones are de-duplicated and merged to obtain a base station selection area. According to the positions of the base stations in the base station engineering parameter data, the base stations within the base station selection area are obtained as candidate base stations, and a candidate site set is obtained.

[0033] Optionally, the base station selection area can be a polygon area. The buffer zones are merged to remove the overlapping parts and synthesized into a single polygon area, which is the base station selection area.

[0034] Optionally, in step 200, the evaluation areas of the candidate base stations in the candidate site set are constructed through computational geometry, including: Step 201, determine the coverage distance of the candidate base stations according to the base station engineering parameter data of the candidate base stations in the candidate site set; Step 202, with the candidate base station as the center and the coverage distance as the radius, generate a circular coverage circle; Step 203, with the candidate base stations within the circular coverage circle as nodes, use the triangulation algorithm to construct the triangular grid corresponding to the candidate base stations; Step 204, determine the adjacent base stations of the candidate base stations based on the adjacency relationship of the triangular grid; Step 205, generate sectors according to the base station density of the adjacent base stations, screen out the key base stations in each sector as boundary points, and construct the evaluation area of the candidate base stations.

[0035] According to the engineering parameter data of the candidate base stations in the candidate site set, determine the coverage distance of the candidate base stations. In one embodiment, according to the maximum access path loss of the candidate base station, the coverage distance of the propagation model of the candidate base station is calculated through a 3GPP channel model, etc. Among them, according to the coverage scenario in the engineering parameter data of the candidate base station, a suitable radio propagation model, such as the 3GPP channel model, etc., is selected to calculate the coverage distance of the propagation model of the candidate base station. In the 3GPP channel model, the path loss is a parameter of the distance from the terminal to the base station and the sector frequency. In this embodiment, the distance from the terminal corresponding to the maximum access path loss to the base station is the coverage distance of the candidate base station.

[0036] For any candidate base station B, with the base station B as the center and the coverage distance of the base station B as the radius, generate a circular coverage circle M. Further, with all the candidate base stations within the circular coverage circle M as nodes, use the triangulation algorithm to construct the triangular grid corresponding to the candidate base stations.

[0037] In one embodiment, all the candidate base stations within the preset distance range around the base station B are used as nodes, and all the nodes are set as the vertex set V = {v 0 , v 1,...,v n}, where v 0 is the base station B, n represents the number of nodes, and the preset distance range is, for example, less than or equal to 1.5 times the coverage distance Des of the base station B.

[0038] Furthermore, the triangulation algorithm is used to construct a triangular mesh. The triangulation algorithm is, for example, the Bowyer-Watson algorithm. The constructed triangular mesh is a Delaunay triangular mesh, which satisfies the empty circle property, that is, the circumcircle of any triangle does not contain other vertices. Add a constraint condition: delete the triangular edges whose side lengths exceed the coverage distance Des of the propagation model to ensure that the network topology conforms to the actual coverage ability.

[0039] Based on the adjacency relationship of the triangular mesh, determine the adjacent base stations of the candidate base station. According to the base station density of the adjacent base stations, dynamically generate sectors, and screen out the key base stations in the sectors as boundary points. Map the screened boundary points to construct an evaluation area of the candidate base station.

[0040] In one embodiment, an adjacency matrix is established through the adjacency relationship of the triangular mesh, adjacency_matrix[i][j]=1, v i and v j are connected by a Delaunay edge; extract the set N={n 1 ,n 2 ,...,n k} of the direct adjacent base stations of the base station B, where k≥3.

[0041] Based on the base station density, generate sectors by means of a dynamic sector generation algorithm, etc. Specifically, calculate the relative azimuth angles and relative distances of each adjacent base station, and sort the adjacent base stations according to the calculated relative azimuth angles: θ 1 ≤θ 2 ≤...≤θ k , calculate the difference and standard deviation of the relative azimuth angles of adjacent two adjacent base stations according to the sorting order, and dynamically divide the number of sectors in the manner shown in Formula 2 below: N=floor(360° / (median(θ i+1 -θ i )+3σ)); (1) where N is the number of sectors, σ is the standard deviation of the relative azimuth angle, and according to the dynamically divided number of sectors, it can be ensured that the sector division adapts to the distribution density of the adjacent base stations.

[0042] Furthermore, based on the dynamically divided number of sectors, generate sector boundary angles in the manner shown in Formula 2 below: Φ j =θ B+j×360° / N±Δθ,j = 0, 1, ..., N - 1; (2) where Δθ is the anti-overlap buffer angle. Optionally, Δθ = 0.1×min(θ i+1 - θ i ).

[0043] In each sector, select key base stations as boundary points. For the selection of key base stations, specifically, filter the adjacent base station n j satisfying the condition: Φ i ≤ θ j+1 < Φ i , and calculate the radial coverage weight W i : W i = α×r i + β×(h B / h i ) + γ×(PL max / PL i ), where α, β, and γ are configurable distance weight, height ratio weight, and path loss ratio weight. PL i is the path loss value from base station B to adjacent base station n i , PL max is the maximum path loss, r i is the relative distance from base station B to adjacent base station n i , h B is the antenna hanging height of base station B, and h i is the antenna hanging height of adjacent base station n i .

[0044] According to the calculated radial coverage weights of each adjacent base station of base station B, select the top m adjacent base stations with the largest weights as boundary points, project these boundary points, and construct the evaluation area of base station B. In one embodiment, use the α-shape algorithm, etc. to construct the optimal polygon area to obtain the evaluation area of base station B. Specifically, project all boundary points into the polar coordinate system centered on base station B, and calculate the radius α r of the polar coordinate system. Optionally, α r = 0.8×Des, where 0.8 is a configurable coefficient and Des is the maximum coverage distance of the propagation model of base station B.

[0045] Furthermore, generate the α-hull polygon, eliminate abnormal convex points, and then convert the α-hull to geographic polygon coordinates to obtain the maximum evaluation area of base station B.

[0046] Optionally, the base station engineering parameter data in the multi-dimensional dataset includes the coverage scenario of the base station. In step 300, identifying the backbone sites according to the spatial characteristics includes: Step 301: Obtain the site recognition models corresponding to the coverage scenarios of each candidate base station in the candidate site set; Step 302: Input the spatial features into the site recognition models to classify each of the candidate base stations, and obtain a candidate set of backbone sites; Step 303: Iteratively optimize the candidate set to obtain backbone sites.

[0047] Obtain the site recognition models corresponding to the coverage scenarios of each candidate base station in the candidate site set. Input the extracted spatial features into the site recognition models to classify the candidate site set and determine whether a candidate site is a backbone site. For the candidate base stations classified as backbone sites, obtain a candidate set of backbone sites. Further, optimize the candidate set to obtain the final backbone sites.

[0048] Optionally, if there are multiple backbone sites, obtain a backbone site set corresponding to the multiple backbone sites in a set form.

[0049] In one embodiment, use the site recognition models to identify whether the candidate site set is a backbone site, and include the base stations classified as backbone sites in a new set, which is the candidate set of backbone sites. In the candidate set of backbone sites, preferentially select the backbone sites for low-altitude area networking to remove ultra-close sites and ensure the optimization of site distribution. Calculate the distances between all adjacent base stations in the candidate set of backbone sites. For the case where the distance between two adjacent base stations is less than a preset distance threshold, such as 200 meters, select the base station with the highest height for retention and eliminate the base stations with lower heights, so as to optimize the candidate set, avoid over-dense sites, and reduce signal interference and resource waste.

[0050] Further, in step 303, iteratively optimizing the candidate set to obtain backbone sites may further include: Step 313: Based on each candidate backbone site in the candidate set, conduct flight tests on the low-altitude area according to the flight route to determine whether there are coverage holes in the low-altitude area; Step 323: If there are coverage holes and the coverage holes are within the coverage ranges of the candidate backbone sites and themselves, obtain the sub-optimal base stations of the candidate backbone sites as backbone sites to optimize the candidate set; Step 333: After removing the sampling points of each backbone site in the candidate set from the measurement report data, evaluate the ground coverage rate of each base station; Step 343: Mark the backbone base stations affecting ground coverage according to the ground coverage rate. Step 353: Update the candidate site set according to the candidate set, return and execute the step of constructing the evaluation area of each candidate base station according to the candidate site set, until there is no coverage hole in the low-altitude area, and obtain each backbone site in the candidate set.

[0051] The iterative optimization of the candidate set is to jointly evaluate the air and ground coverage of the candidate backbone sites in the candidate set to form a backbone site set for private network networking in the low-altitude area. Specifically, first, based on each candidate backbone site in the candidate set, conduct a flight test on the low-altitude area according to the flight routes planned in the low-altitude area to determine whether there is a coverage hole in the low-altitude area when networking based on the candidate backbone sites in the current candidate set. This coverage hole includes uncovered areas and weak coverage areas.

[0052] If there is a coverage hole and the coverage hole is within the coverage range of the candidate backbone site and itself, obtain the sub-optimal base station of the candidate backbone site as the backbone site, so as to realize the optimization of the candidate set. Among them, the sub-optimal base station of the candidate backbone site can be a candidate base station adjacent to the candidate backbone site with an antenna height lower than that of the candidate backbone site.

[0053] In one embodiment, for the candidate set of backbone sites, according to the distance between two adjacent candidate backbone sites, retain the higher site and eliminate the lower site. The eliminated candidate backbone sites are used as the backup site set. On this basis, conduct a joint air and ground coverage evaluation on the candidate set, check the air coverage loopholes of each candidate backbone site in the candidate set, and supplement the backbone site set for low-altitude private network networking through the sub-optimal base stations in the backup site set. Further, after the base station becomes a backbone site for low-altitude networking, it is necessary to adjust the antenna angle for air coverage, which will reduce the ground coverage ability of the base station. Therefore, it is necessary to conduct a ground coverage evaluation on the backbone site to check whether the backbone site affects the ground coverage ability, and mark the backbone sites that affect the ground coverage.

[0054] Specifically, when verifying the ground coverage of the backbone site from the measurement report data (i.e., MR data) of the multi-dimensional data set in the low-altitude area, eliminate the sampling points of the backbone site from the measurement report data in the base station evaluation area, and then re-evaluate the coverage rate in the evaluation area of the backbone site to determine whether there is an impact on the ground coverage in the evaluation area of the backbone site.

[0055] Optionally, if the coverage rate in the evaluation area is greater than or equal to the preset coverage rate threshold after eliminating the sampling points of the backbone site, the backbone site does not affect the ground coverage. Otherwise, if the coverage rate is lower than the preset coverage rate threshold, mark the backbone site.

[0056] Then, according to the optimized candidate set, update the candidate site set, remove each backbone site in the candidate set from the candidate site set to obtain a new candidate site set. For the updated candidate site set, reconstruct the evaluation area of each candidate site through computational geometry, screen out new backbone sites, form a supplementary site set for the backbone sites in the candidate set to address coverage holes, merge the backbone sites in the supplementary site set into the candidate set, and conduct joint air-ground coverage evaluation until there are no coverage holes in the low-altitude area. The final updated candidate set is obtained, and the base stations in this candidate set are the final backbone sites used for private network networking in the low-altitude area.

[0057] Optionally, the flight test of the low-altitude area can be performed by an unmanned aerial vehicle, and no specific limitation is imposed on this.

[0058] Optionally, the candidate backbone sites in the candidate set are obtained by classifying and identifying candidate base stations based on a site identification model. The site identification model is trained based on a sample data set under different coverage scenarios and can identify whether a candidate base station under different coverage scenarios is a backbone site. Or, different coverage scenarios correspond to different site identification models. The site identification model corresponding to any coverage scenario is trained based on the sample data set under this coverage scenario. Based on the coverage scenario corresponding to the candidate base station, the site identification model corresponding to this coverage scenario is used to classify and identify whether the candidate base station is a backbone site.

[0059] Before step 301, it may further include: Step 310, obtain an original data set; the original data set includes the antenna hanging height, coverage scenario, and height information of geographical objects of sample base stations; Step 320, construct a spatial influence factor based on the distance between the geographical object and the receiving point of the sample base station and the height information; Step 330, compare the height information with the antenna hanging height, and generate a label for the spatial influence factor according to the comparison result; Step 340, construct a sample data set based on the label and the spatial influence factor, and divide the sample data set according to the coverage scenario to obtain sample data subsets; Step 350, use each sample data subset to iteratively train a preset support vector machine model to obtain the site identification models corresponding to each coverage scenario.

[0060] In the model training stage, for the training of the site identification model, first obtain an original data set, which includes the engineering parameter data of sample base stations and the height information of geographical objects. The engineering parameter data of sample base stations further includes the antenna hanging height and the coverage scenario.

[0061] Among them, the geographical objects include buildings such as buildings. Different base stations in the sample base stations are distinguished by base station IDs. Each base station has a unique base station ID, which can be generated by a hash function.

[0062] According to the antenna hanging height of the sample base station and the height information of the geographical object, a spatial feature is constructed. The spatial feature includes a height feature and a spatial influence factor. The spatial influence factor is an index factor used to quantify the influence of the geographical object on signal propagation, and is calculated through the height of the geographical object and the distance between it and the base station receiving point. The spatial influence factor The specific calculation method is shown in the following formula 3: ; (3) represents the height of the th geographical object, represents the distance between the geographical object and the base station receiving point, while is the number of geographical objects around the base station. By calculating the spatial influence factor of the base station signal propagation in this way, the antenna hanging height of the base station and the height of the geographical object are comprehensively considered, ensuring that the geographical object closer and higher has a greater shielding effect on the base station signal. Moreover, it not only expands the data set for model training but also increases the geographical factors affecting the base station coverage area.

[0063] Compare the height information of the geographical object with the antenna hanging height of the sample base station, and construct the label of the spatial influence factor according to the comparison result. Comparing the height information of the geographical object with the antenna hanging height of the sample base station specifically means comparing the antenna hanging height of the sample base station with the height of the geographical objects around it. If the antenna hanging height of a certain sample base station is higher than the height of the geographical objects around it or the antenna hanging height of other base stations, its label is the first eigenvalue, otherwise it is the second label value. When a certain geographical object is higher than the geographical objects around it or the antenna hanging height of the sample base station, its label is the first label value, otherwise it is the second label value. Exemplarily, the first label value is 1 and the second label value is -1. That is, the label of the spatial influence factor is used to characterize Based on the label and the constructed spatial influence factor, construct a sample data set, and divide the constructed sample data set according to the coverage scenarios of each sample base station to obtain sample data subsets under each coverage scenario, and construct sample data for different scenarios.

[0064] Optionally, the site recognition model is constructed based on the support vector machine model. Using the sample data subsets under each coverage scenario, iteratively train the preset support vector machine model to obtain the site recognition model corresponding to this coverage scenario.

[0065] When training a support vector machine model using a subset of sample data, first split the sample data set by coverage scenarios. Each coverage scenario is split into a subset of sample data. Further, each subset of sample data is divided into a training set and a test set. The division of the training set and the test set is achieved according to a preset splitting ratio, which can be determined according to the requirements of model training and the characteristics of the data. Exemplarily, the splitting ratio is 8:2, with 80% of the data in the subset of sample data being divided into the training set and 20% being divided into the test set.

[0066] Further, adopt the method of stratified sampling to ensure the consistency of the distribution of data of each category in each subset of sample data with the original sample data set. Specifically, first classify and organize the sample data set, separating the sample data set by category. The categories include two types: geographical objects and base stations, so as to facilitate the separate processing of the data of each category. For each category separated, calculate the amount of data to be allocated to the training set and the test set, and adopt the method of random sampling to extract the corresponding proportion of samples from the data of each category and add them to the training set and the test set to ensure that the samples in each subset of sample data can represent their respective categories as much as possible. Finally, combine the samples extracted from the data of each category into the final training set, test set and validation set, and use the divided training set, test set and validation set to perform iterative training on the preset support vector machine model.

[0067] In the model training by scenarios, according to any one of the coverage scenarios for which model training is to be carried out, select the training set, test set and validation set under this coverage scenario for model training. Further, in the process of model training under this coverage scenario, adopt the method of grid search to optimize the parameters, and through five-fold cross-validation, evaluate the performance of the model for each parameter combination. Finally, select the parameter combination with the highest F1-score (balanced mean) in the cross-validation as the final model parameter to obtain the site recognition model under this coverage scenario.

[0068] The core of the support vector machine model (SVM) lies in finding a hyperplane to divide data points of different categories. Since in this embodiment, classification is carried out through height features, with fewer feature items but a large amount of data, therefore, the radial basis function RBF (Radial Basis Function), based on the Gaussian distribution, can effectively handle nonlinear problems and is particularly effective for the case of a small sample dimension but a large number of samples.

[0069] During the training process of the support vector machine model, the model parameters to be trained include the penalty parameter and the kernel function parameter. The penalty parameter C is a parameter that balances model complexity and training error. When C takes a large value, the model will try to minimize the training error, which may lead to overfitting; when C takes a small value, the model will be more inclined to have a larger margin, which may ignore the misclassification of some training data points and lead to underfitting. The kernel function parameter determines the shape of the Gaussian kernel function. The larger the value of the kernel function parameter, the smaller the range of action of the kernel function, the more complex the model, and it is easy to overfit; the smaller the value of the kernel function parameter, the larger the range of action of the kernel function, the simpler the model, and it may lead to underfitting.

[0070] During the iterative training process, after each round of iteration is completed, the trained model is verified through the validation set. Since base stations with relatively high antennas belong to the minority, the number of positive and negative samples in the sample data varies greatly. Therefore, the F1-score (balanced mean) is used as the performance evaluation index for the model parameter combination.

[0071] Select the parameter combination with the highest F1-score (balanced mean) as the optimal model parameter. Samples with relatively high antenna heights are called positive samples, and the rest are negative samples. An actually relatively high base station means that the label in the validation set is 1 and the prediction of being relatively high is that the label predicted by the model is 1. Actually not relatively high means that the label in the validation set is -1 and the prediction of not being relatively high is that the label predicted by the model is -1.

[0072] TP: The number of samples that predict positive samples as positive samples, that is, the number of samples that predict actually relatively high base stations as relatively high; TN: The number of samples that predict negative samples as negative samples, that is, the number of samples that predict actually not relatively high base stations as not relatively high; FP: The number of samples that predict negative samples as positive samples, that is, the number of samples that predict actually not relatively high base stations as relatively high; FN: The number of samples that predict positive samples as negative samples, that is, the number of samples that predict actually relatively high base stations as not relatively high; Calculate the F1-Score in the manner shown in Formula 4 below: ; (4) Evaluate the model on the test set. The parameter combination with an F1-score greater than the preset threshold is the model parameter with qualified performance indicators. Further select the parameter combination with the largest F1-score, and use the model configured with this parameter combination as the site recognition model for the antenna height in the corresponding coverage scenario.

[0073] According to the obtained multi-dimensional data set, the spatial impact factor of the candidate base station is constructed. Based on the spatial impact factor, the pre-trained candidate base station in the coverage scenario corresponding to the candidate base station is used to predict the spatial impact factor of the candidate set station. The spatial features constructed by the antenna height are used for prediction. The predicted label label is 1 or -1, so as to determine whether the antenna height of the candidate base station belongs to a high site or a low site. The candidate base station predicted as a high site is used as the backbone site to be selected, and the candidate set of backbone sites is obtained. Further iterative selection is performed on the selected set to obtain the final backbone site, and a dedicated network is established for the low-altitude area to achieve air coverage of the low-altitude area.

[0074] In this embodiment, the candidate base stations are determined through multi-dimensional data such as the engineering parameter data of the base station, the relevant information of the geographical objects and the planned routes. Based on the calculation set, the evaluation area of ​​the base station is dynamically constructed. Based on the spatial impact factor and the site recognition model obtained by the scene training, the backbone sites among the candidate base stations are identified and iteratively selected, which improves the scientificity and accuracy of the backbone site decision-making for low-altitude networking. In addition, in the face of large-scale networks, the decision-making method for backbone sites greatly improves the networking efficiency.

[0075] Furthermore, by calculating the coverage distance of the base station's propagation model, dynamically constructing the base station's triangular mesh based on the triangulation algorithm, and dynamically generating sectors based on the base station density of adjacent base stations, the key base stations are selected as boundary points for projection based on multi-dimensional weights to generate the base station evaluation area. The geographical objects and base stations in the evaluation area are selected as the optimal objects, and the backbone sites identified by the site recognition model obtained through scene-by-scene training are iteratively optimized, which is conducive to adapting to the networking needs of different scenarios, and can identify backbone sites in complex scenarios such as different building densities and heights, thereby improving the applicability of networking.

[0076] In the process of iterative optimization of backbone sites, through joint evaluation of air-ground coverage, the coverage holes of air coverage and the impact on ground coverage are checked, and the backbone sites of air coverage are supplemented to ensure that the final backbone sites can not only fill all the coverage holes of air coverage, but also ensure the quality and stability of air network coverage. While meeting the low-cost networking requirements of air-ground collaboration, the evaluation area can be accurately determined, and the backbone sites can be iteratively evaluated and optimized, which can maximize the reuse of existing network resources while ensuring the network quality of air coverage, effectively reducing the networking cost of low-altitude private networks.

[0077] The low-altitude networking device provided in an embodiment of the present application is described below. The low-altitude networking device described below and the low-altitude networking method described above can be referenced to each other.

[0078] Reference Figure 2, the low-altitude networking device provided by the embodiments of the present application includes: A base station selection module 10, configured to determine a set of candidate sites according to a multi-dimensional data set of a to-be-planned low-altitude area; the multi-dimensional data set includes base station engineering parameter data and flight routes of the low-altitude area; A region construction module 20, configured to construct an evaluation region for each candidate base station according to the set of candidate sites, and determine a set of preferred objects of the candidate base stations according to the evaluation region; A site identification module 30, configured to extract spatial features of each preferred object in the set of preferred objects, and identify backbone sites according to the spatial features; A low-altitude networking module 40, configured to perform networking on the low-altitude area based on the backbone sites.

[0079] In one embodiment, the region construction module 20 is further configured to: Determine the coverage distance of each candidate base station in the set of candidate sites according to the base station engineering parameter data of the candidate base stations; Taking the candidate base station as the center and the coverage distance as the radius, generate a circular coverage circle; Taking each candidate base station within the circular coverage circle as a node, use a triangulation algorithm to construct a triangular grid corresponding to the candidate base station; Based on the adjacency relationship of the triangular grid, determine the adjacent base stations of the candidate base station; Generate sectors according to the base station density of the adjacent base stations, screen out the key base stations in each sector as boundary points, and construct an evaluation region for the candidate base station.

[0080] In one embodiment, the base station engineering parameter data includes a coverage scenario; the site identification module 30 is further configured to: Obtain a site identification model corresponding to the coverage scenario of each candidate base station in the set of candidate sites; Input the spatial features into the site identification model, classify each candidate base station, and obtain a set of candidates for backbone sites; Perform iterative optimization on the set of candidates to obtain backbone sites.

[0081] In one embodiment, the multi-dimensional data further includes measurement report data; the site identification module 30 is further configured to: Based on each candidate backbone site in the set of candidates, perform a flight test on the low-altitude area according to the flight route to determine whether there are coverage holes in the low-altitude area; If there are coverage holes and the coverage holes are within the coverage range of the candidate backbone sites and themselves, obtain the sub-optimal base stations of the candidate backbone sites as backbone sites to optimize the set of candidates; After removing the sampling points of each backbone site in the candidate set from the measurement report data, evaluate the ground coverage rate of each base station; Mark the backbone base stations affecting ground coverage according to the ground coverage rate; Update the candidate site set according to the candidate set, return and execute the step of constructing the evaluation area of each candidate base station according to the candidate site set, until there is no coverage hole in the low-altitude area, and obtain each backbone site in the candidate set.

[0082] In one embodiment, the low-altitude networking device further includes a model training module, which is used for: Obtain an original data set; the original data set includes the antenna hanging height of the sample base station, the coverage scenario, and the height information of the geographical object; Based on the distance between the geographical object and the receiving point of the sample base station and the height information, construct a spatial influence factor; Compare the height information with the antenna hanging height, and generate a label for the spatial influence factor according to the comparison result; Based on the label and the spatial influence factor, construct a sample data set, and divide the sample data set according to the coverage scenario to obtain sample data subsets; Use each sample data subset to iteratively train a preset support vector machine model to obtain a site recognition model corresponding to each coverage scenario.

[0083] In one embodiment, the base station engineering parameter data includes the base station location; the base station selection module 10 is further used for: Establish a plurality of buffer zones on both sides of the flight path in the low-altitude area to be planned with a preset width; Deduplicate and merge the plurality of buffer zones to obtain a base station selection area; According to the base station location of each base station in the base station engineering parameter data, obtain each candidate base station in the base station selection area to obtain a candidate site set.

[0084] Figure 3 Illustrate a schematic diagram of the physical structure of an electronic device, as Figure 3 shown, the electronic device may include: a processor (processor) 310, a communication interface (Communications Interface) 320, a memory (memory) 330, and a communication bus 340. Among them, the processor 310, the communication interface 320, and the memory 330 communicate with each other through the communication bus 340. The processor 310 can call the logical instructions in the memory 330 to execute the low-altitude networking method, and the method includes: Determine a set of candidate sites based on the multi-dimensional dataset of the low-altitude area to be planned; the multi-dimensional dataset includes base station engineering parameter data and the flight routes of the low-altitude area; Construct an evaluation area for each candidate base station according to the set of candidate sites, and determine a set of preferred objects for the candidate base stations according to the evaluation area; Extract the spatial features of each preferred object in the set of preferred objects, and identify backbone sites according to the spatial features; Network the low-altitude area based on the backbone sites.

[0085] In addition, when the logical instructions in the above-mentioned memory 330 are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium and includes several instructions for causing a computer device (which may be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present application. The foregoing storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0086] On the other hand, the embodiments of the present application further provide a computer program product, the computer program product includes a computer program, the computer program can be stored on a non-transitory computer-readable storage medium, and when the computer program is executed by a processor, the computer can execute the low-altitude networking method provided by the above-mentioned various methods. The method includes: Determine a set of candidate sites based on the multi-dimensional dataset of the low-altitude area to be planned; the multi-dimensional dataset includes base station engineering parameter data and the flight routes of the low-altitude area; Construct an evaluation area for each candidate base station according to the set of candidate sites, and determine a set of preferred objects for the candidate base stations according to the evaluation area; Extract the spatial features of each preferred object in the set of preferred objects, and identify backbone sites according to the spatial features; Network the low-altitude area based on the backbone sites.

[0087] On yet another aspect, the embodiments of the present application further provide a non-transitory computer-readable storage medium, on which a computer program is stored, and when the computer program is executed by a processor, it is implemented to execute the low-altitude networking method provided by the above-mentioned various methods. The method includes: Determine a set of candidate sites according to the multi-dimensional dataset of the low-altitude area to be planned; the multi-dimensional dataset includes base station engineering parameter data and flight routes of the low-altitude area; Construct an evaluation area for each candidate base station according to the set of candidate sites, and determine a set of preferred objects for the candidate base stations according to the evaluation area; Extract the spatial features of each preferred object in the set of preferred objects, and identify backbone sites according to the spatial features; Network the low-altitude area based on the backbone sites.

[0088] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed to multiple network units. Some or all of the modules can be selected according to actual needs to achieve the purpose of the solution of this embodiment. Those of ordinary skill in the art can understand and implement it without creative work.

[0089] Through the description of the above embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus a necessary general hardware platform, and of course also by hardware. Based on this understanding, the above technical solutions, in essence, or the part that contributes to the prior art, can be embodied in the form of a software product, which can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., including several instructions to enable a computer device (which can be a personal computer, server, or network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0090] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present application, and are not intended to limit them; although the present application has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that they can still modify the technical solutions described in the foregoing embodiments, or perform equivalent replacements for some of the technical features; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of each embodiment of the present application.

Claims

1. A low-altitude networking method, characterized in that: include: Determine a candidate site set based on a multidimensional dataset of the low-altitude area to be planned; The multidimensional data set includes base station engineering parameter data and the flight path of the low-altitude area; Constructing an evaluation area for each candidate base station according to the candidate site set, and determining a preferential object set for the candidate base station according to the evaluation area; Extracting spatial features of each preferred object in the preferred object set, and identifying backbone sites according to the spatial features; The low-altitude area is networked based on the backbone site.

2. The low-altitude networking method according to claim 1, characterized in that: The constructing the evaluation area of ​​each candidate base station according to the candidate site set includes: Determining the coverage distance of the candidate base station according to the base station engineering parameter data of each of the candidate base stations in the candidate site set; Taking the candidate base station as the center and the coverage distance as the radius, a circular coverage circle is generated; Taking each candidate base station within the circular coverage circle as a node, a triangulation algorithm is used to construct a triangular mesh corresponding to the candidate base station; Determining adjacent base stations of the candidate base station based on the adjacency relationship of the triangular mesh; Sectors are generated according to the base station density of the adjacent base stations, key base stations in each sector are screened out as boundary points, and an evaluation area for the candidate base stations is constructed.

3. The low-altitude networking method according to claim 1, characterized in that: The base station engineering parameter data includes coverage scenarios; and the identifying backbone sites according to the spatial features includes: Obtaining a site identification model corresponding to a coverage scenario of each candidate base station in the candidate site set; Inputting the spatial features into the site identification model, classifying each of the candidate base stations, and obtaining a candidate set of backbone sites; The candidate set is iteratively optimized to obtain a backbone site.

4. The low-altitude networking method according to claim 3, characterized in that: The multi-dimensional data also includes measurement report data; the iterative optimization of the candidate set to obtain the backbone site includes: Based on each of the candidate backbone sites in the candidate set, a flight test is performed on the low-altitude area according to the route to determine whether there is a coverage hole in the low-altitude area; If there is a coverage hole and the coverage hole is within the coverage range of the candidate backbone site and the candidate backbone site, a suboptimal base station of the candidate backbone site is obtained as a backbone site to optimize the candidate set; After removing the sampling points of each backbone site in the candidate set from the measurement report data, evaluating the ground coverage of each base station; Marking the backbone base stations that affect ground coverage according to the ground coverage rate; The candidate site set is updated according to the candidate set, and the step of constructing the evaluation area of ​​each candidate base station according to the candidate site set is returned and executed until there is no coverage hole in the low-altitude area, and each backbone site in the candidate set is obtained.

5. The low-altitude networking method according to claim 3, characterized in that: Before acquiring the site identification model corresponding to the coverage scenario of each candidate base station in the candidate site set, the method further includes: Acquire an original data set; the original data set includes antenna heights of sample base stations, coverage scenes, and height information of geographic objects; constructing a spatial impact factor based on the distance between the geographical object and the receiving point of the sample base station and the height information; Comparing the height information with the antenna hanging height, and generating a label of the spatial impact factor according to the comparison result; Based on the label and the spatial impact factor, a sample data set is constructed, and the sample data set is divided according to the coverage scenario to obtain a sample data subset; The preset support vector machine model is iteratively trained using each of the sample data set subsets to obtain a site recognition model corresponding to each of the coverage scenarios.

6. The low-altitude networking method according to claim 1, characterized in that: The base station engineering parameter data includes the base station location; the candidate site set is determined according to the multidimensional data set of the low-altitude area to be planned, including: Establish multiple buffer zones with preset widths on both sides of the route in the low-altitude area to be planned; De-overlapping and merging the multiple buffers to obtain a base station selection area; According to the base station position of each base station in the base station engineering parameter data, each candidate base station in the base station selection area is obtained to obtain a candidate site set.

7. A low-altitude networking device, characterized in that: include: A base station selection module, used to determine a set of candidate sites based on a multi-dimensional data set of the low-altitude area to be planned; The multidimensional data set includes base station engineering parameter data and the flight path of the low-altitude area; A region construction module, used to construct an evaluation region for each candidate base station according to the candidate site set, and determine a preferential object set for the candidate base station according to the evaluation region; A site identification module, used for extracting the spatial features of each preferred object in the preferred object set, and identifying the backbone site according to the spatial features; The low-altitude networking module is used to network the low-altitude area based on the backbone site.

8. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein: When the processor executes the computer program, the low-altitude networking method as described in any one of claims 1 to 6 is implemented.

9. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by a processor, the low-altitude networking method as described in any one of claims 1 to 6 is implemented.

10. A computer program product, comprising a computer program, characterized in that When the computer program is executed by a processor, the low-altitude networking method as described in any one of claims 1 to 6 is implemented.

Images