Unmanned aerial vehicle base station deployment method, device, equipment, medium and product

Abstract

The invention provides an unmanned aerial vehicle base station deployment method and device, equipment, a medium and a product, and relates to the field of wireless communication. The method comprises the following steps: acquiring simulation model parameters of a to-be-deployed area, and generating a signal coverage simulation model based on ray tracing according to the simulation model parameters; obtaining physical environment data of the to-be-deployed area; configuring an initial neural network model according to the signal coverage simulation model based on ray tracing; according to the physical environment data, carrying out training processing and prediction processing on the initial neural network model to determine a signal intensity graph of the to-be-deployed area; and determining an unmanned aerial vehicle base station deployment scheme of the to-be-deployed area according to the signal intensity graph. The problem that in the prior art, a traditional signal coverage prediction method is influenced by various factors in an actual complex environment, so that the signal coverage condition of the unmanned aerial vehicle to the ground user cannot be accurately predicted, and the deployment efficiency of the unmanned aerial vehicle base station is low is solved.

Description

Technical Field

[0001] This application relates to the field of wireless communication, and in particular to a method, apparatus, equipment, medium and product for deploying a drone base station. Background Technology

[0002] With the rapid development of wireless communication technology, drones, as a flexible mobile platform, have been gradually introduced into the deployment and optimization of communication base stations to improve network coverage and flexibility. However, drone base stations face a number of complex challenges in practical applications, one of the most critical being how to accurately predict the signal coverage of drones to ground users.

[0003] Currently, traditional signal coverage prediction methods are mostly based on theoretical formulas and empirical models, such as path loss models and statistical fading models. These methods can only achieve relatively ideal prediction results under ideal or simplified environments.

[0004] However, traditional signal coverage prediction methods are affected by a variety of factors in the complex environment, making it impossible to accurately predict the signal coverage of drones to ground users, thus resulting in low efficiency in the deployment of drone base stations. Summary of the Invention

[0005] This application provides a method, apparatus, equipment, medium, and product for deploying unmanned aerial vehicle (UAV) base stations, which solves the problem that traditional signal coverage prediction methods are affected by various factors in the actual complex environment, resulting in the inability to accurately predict the signal coverage of UAVs to ground users, thus leading to low efficiency in the deployment of UAV base stations.

[0006] In a first aspect, this application provides a method for deploying a drone base station, comprising:

[0007] Obtain the simulation model parameters of the area to be deployed, and generate a ray-tracing-based signal coverage simulation model based on the simulation model parameters; wherein, the simulation model parameters include ray tracing parameters;

[0008] Obtain physical environment data for the area to be deployed;

[0009] Based on the ray tracing-based signal coverage simulation model, configure the initial neural network model;

[0010] Based on physical environment data, the initial neural network model is trained and predicted to determine the signal strength map of the area to be deployed.

[0011] Based on the signal strength map, determine the deployment plan for drone base stations in the area to be deployed.

[0012] In one possible design, the simulation model parameters of the area to be deployed are obtained, including:

[0013] Obtain the transmission model parameters used for ray tracing;

[0014] Obtain user location parameters;

[0015] Obtain building information data for the area to be deployed.

[0016] In one possible design, a ray-tracing-based signal coverage simulation model is generated based on simulation model parameters, including:

[0017] Based on user location parameters and building information data, the spatial relationship between buildings, user nodes, and candidate drone locations is displayed in the simulation environment to obtain a three-dimensional environmental model.

[0018] Initialize the ray tracing parameters based on the transmission model parameters; the ray tracing parameters include transmit power, path loss model, and fading factor.

[0019] Based on the ray tracing parameters, configure the simulation physics layer and import the 3D environment model to obtain a ray tracing-based signal coverage simulation model.

[0020] In one possible design, based on a ray-tracing-based signal coverage simulation model, an initial neural network model is configured, including:

[0021] Based on the physical environment data in the ray-tracing-based signal coverage simulation model, the input data for the initial neural network model is determined.

[0022] Based on the simulation results data in the ray tracing-based signal coverage simulation model, determine the output data of the initial neural network model;

[0023] Configure the initial neural network model based on the input and output data.

[0024] In one possible design, an initial neural network model is trained and predicted based on physical environment data to determine the signal strength map of the area to be deployed, including:

[0025] The physical environment data is normalized and converted to obtain standard physical environment data.

[0026] Based on standard physical environment data, the initial neural network model is trained and predicted to determine the signal strength map of the area to be deployed.

[0027] In one possible design, an initial neural network model is trained and predicted based on standard physical environment data to determine the signal strength map of the area to be deployed, including:

[0028] Obtain the scene source of the standard physical environment data, and add scene tags to the standard physical environment data according to the scene source;

[0029] Based on standard physical environment data with added scene labels, determine the prediction input dataset and the training input dataset;

[0030] Add signal intensity labels to the training input dataset and generate the training output dataset based on the signal intensity labels;

[0031] The initial neural network model is trained based on the training input dataset and the training output dataset to obtain a trained neural network model.

[0032] The trained neural network model is used to perform prediction processing on the input dataset to obtain the signal strength map of the area to be deployed based on the prediction results.

[0033] In one possible design, the deployment scheme for drone base stations in the area to be deployed is determined based on the signal strength map, including:

[0034] Obtain the preset signal strength threshold;

[0035] Based on a preset signal strength threshold, the signal strength map is filtered to obtain a set of feasible flight locations;

[0036] Based on the set of feasible flight locations, determine the deployment plan for drone base stations in the area to be deployed.

[0037] Secondly, this application provides a deployment device for a drone base station, comprising:

[0038] The first acquisition module is used to acquire the simulation model parameters of the area to be deployed, and generate a signal coverage simulation model based on ray tracing according to the simulation model parameters; wherein, the simulation model parameters include ray tracing parameters;

[0039] The second acquisition module is used to acquire physical environment data of the area to be deployed.

[0040] The configuration module is used to configure the initial neural network model based on the ray tracing-based signal coverage simulation model.

[0041] The first determining module is used to train and predict the initial neural network model based on physical environment data in order to determine the signal strength map of the area to be deployed.

[0042] The second determining module is used to determine the deployment plan for drone base stations in the area to be deployed based on the signal strength map.

[0043] Thirdly, this application provides a deployment device for a drone base station, including: a memory and a processor;

[0044] The memory stores the instructions that the computer executes;

[0045] The processor executes computer execution instructions stored in memory, causing the processor to perform a method for deploying a drone base station as described in the first aspect of the invention.

[0046] Fourthly, this application provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement a method for deploying a drone base station as described in the first aspect of the invention.

[0047] Fifthly, this application provides a computer program product, including a computer program that, when executed by a processor, implements a method for deploying a drone base station according to the first aspect of the invention.

[0048] This application provides a method, apparatus, equipment, medium, and product for deploying unmanned aerial vehicle (UAV) base stations, comprising: acquiring simulation model parameters of the area to be deployed, and generating a signal coverage simulation model based on ray tracing according to the simulation model parameters; acquiring physical environment data of the area to be deployed; configuring an initial neural network model based on the ray tracing-based signal coverage simulation model; training and predicting the initial neural network model based on the physical environment data to determine the signal strength map of the area to be deployed; and determining the UAV base station deployment scheme for the area to be deployed based on the signal strength map. Compared to existing technologies, traditional signal coverage prediction methods are affected by various factors in complex real-world environments, leading to inaccurate predictions of UAV signal coverage for ground users, resulting in low deployment efficiency of UAV base stations. This application solves the data scarcity problem by providing physically realistic signal propagation samples through ray tracing; it preserves spatial multi-scale features through the encoding-decoding structure of the neural network model, adapting to nonlinear propagation such as building obstruction and multipath effects; thereby accurately predicting signal coverage and achieving accurate deployment of UAV base stations, thus improving deployment efficiency. Attached Figure Description

[0049] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0050] Figure 1 A schematic diagram of the system architecture for a method of deploying a drone base station provided in an embodiment of this application;

[0051] Figure 2 A flowchart illustrating a method for deploying a drone base station provided in this application embodiment. Figure 1 ;

[0052] Figure 3 A flowchart illustrating a method for deploying a drone base station provided in this application embodiment. Figure 2 ;

[0053] Figure 4 A flowchart illustrating a method for deploying a drone base station provided in this application embodiment. Figure 3 ;

[0054] Figure 5 A flowchart illustrating a method for deploying a drone base station provided in this application embodiment. Figure 4 ;

[0055] Figure 6 This is a schematic diagram of the structure of a drone base station deployment device provided in an embodiment of this application;

[0056] Figure 7 This is a schematic diagram of the structure of a drone base station deployment device provided in an embodiment of this application;

[0057] Figure 8 This is a schematic diagram illustrating the simulation model construction and parameter configuration provided in the embodiments of this application;

[0058] Figure 9 This is a schematic diagram illustrating real-world data acquisition and preprocessing as provided in the embodiments of this application.

[0059] Figure 10 This is a schematic diagram illustrating the deployment, training, and prediction of the U-Net model provided in an embodiment of this application;

[0060] Figure 11 This is a schematic diagram illustrating the signal coverage prediction result processing and UAV location planning provided in an embodiment of this application.

[0061] Figure 12 This is a schematic diagram of the overall process provided for an embodiment of this application;

[0062] Figure 13 This is a schematic diagram illustrating the building environment input and signal prediction effects provided in an embodiment of this application.

[0063] Figure 14 This is a schematic diagram illustrating the multi-point of interest signal coverage prediction and deployable area selection provided in an embodiment of this application. Detailed Implementation

[0064] Exemplary embodiments will now be described in detail, examples of which are illustrated in the accompanying drawings. When the following description relates to the drawings, unless otherwise indicated, the same numbers in different drawings denote the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with this application. Rather, they are merely examples of apparatuses and methods consistent with some aspects of this application as detailed in the appended claims.

[0065] In the embodiments of this application, the terms "first" and "second" are used to distinguish identical or similar items with substantially the same function and effect. Those skilled in the art will understand that the terms "first" and "second" do not limit the quantity or execution order, nor do they necessarily imply difference. It should be noted that in the embodiments of this application, words such as "exemplary" or "for example" are used to indicate examples, illustrations, or explanations. Any embodiment or design scheme described as "exemplary" or "for example" in this application should not be construed as being more preferred or advantageous than other embodiments or design schemes. Specifically, the use of words such as "exemplary" or "for example" is intended to present the relevant concepts in a concrete manner. In the embodiments of this application, "at least one" refers to one or more, and "more than one" refers to two or more.

[0066] It should be noted that the phrase "at...time" in the embodiments of this application can refer to the instant at which a certain situation occurs, or to a period of time after the occurrence of a certain situation; the embodiments of this application do not specifically limit this. Furthermore, the method for deploying a drone base station provided in the embodiments of this application is merely an example, and the method for deploying a drone base station may include more or fewer elements.

[0067] It should also be noted that the user information (including but not limited to user device information, user personal information, etc.) and data (including but not limited to data used for analysis, stored data, displayed data, etc.) involved in this application are all information and data authorized by the user or fully authorized by all parties. Furthermore, the collection, storage, use, processing, transmission, provision, disclosure, and application of the relevant data all comply with the relevant laws, regulations, and standards of the relevant countries and regions, have taken necessary confidentiality measures, do not violate public order and good morals, and provide corresponding operation access points for users to choose to authorize or refuse.

[0068] With the rapid development of wireless communication technology, drones, as a flexible mobile platform, have been gradually introduced into the deployment and optimization of communication base stations to improve network coverage and flexibility.

[0069] However, drone base stations face a variety of complex challenges in practical applications, the most critical of which is how to accurately predict the signal coverage of drones to ground users. Traditional signal coverage prediction methods are mostly based on theoretical formulas and empirical models, such as path loss models and statistical fading models. These methods can achieve relatively ideal prediction results in ideal or simplified environments, but they often fail to achieve the expected accuracy in real-world complex environments.

[0070] In real-world applications, urban or campus environments often feature numerous buildings, undulating terrain, and complex user distribution. These factors can lead to multipath effects, shadowing fading, and rapid fading during signal propagation, posing significant challenges to signal coverage prediction. Traditional theoretical models struggle to adequately reflect the impact of building height, outline, materials, and distribution density on electromagnetic wave propagation, thus often exhibiting limitations in fine-grained prediction.

[0071] Currently, ray tracing technology, as a method to simulate the propagation path of electromagnetic waves, uses precise physical models to model in detail the reflection, refraction, and scattering of signals when they encounter obstacles, and can reproduce the propagation of electromagnetic waves in the real environment to a large extent.

[0072] However, ray tracing technology also suffers from problems in practical applications, such as high computational cost, insufficient real-time performance, and heavy reliance on environmental information. Especially in large-scale, complex environments, accurately acquiring 3D building models, material properties, and user location data, and efficiently applying this information to ray tracing simulations, remains a pressing challenge.

[0073] Based on this, embodiments of this application provide a method, apparatus, device, medium, and product for deploying unmanned aerial vehicle (UAV) base stations, which can be used in the field of wireless communication and are intended to solve the above-mentioned technical problems of the prior art.

[0074] The inventive concept of this application is as follows: Addressing the aforementioned problems, the inventors, during their research on the efficiency of UAV base station deployment, discovered that existing methods based on theoretical formulas and empirical models are susceptible to multipath effects, shadow fading, and rapid fading in complex environments. They also struggle to account for the influence of complex factors such as buildings on electromagnetic wave propagation, resulting in poor accuracy in fine-grained predictions. Ray tracing technology, on the other hand, involves high computational demands, insufficient real-time performance, and is highly dependent on environmental information, making data acquisition and application difficult in large-scale complex environments. Therefore, the inventors propose a method for deploying UAV base stations by fully utilizing real-world environmental data and employing neural networks to learn from limited ray tracing data. This improves the accuracy of predicting ground user signal coverage by UAV base stations in complex environments, overcoming the shortcomings of traditional methods and ray tracing technology. This provides a reliable basis for the rational planning and optimized deployment of UAV base stations, thereby improving the coverage quality and performance of communication networks. Based on this, this application proposes a deployment method for UAV base stations to further improve their deployment efficiency.

[0075] The technical solution of this application and how the technical solution of this application solves the above-mentioned technical problems are described in detail below with specific embodiments. These specific embodiments can be combined with each other, and the same or similar concepts or processes may not be described again in some embodiments. The embodiments of this application will now be described with reference to the accompanying drawings.

[0076] Figure 1 This is a schematic diagram of the system architecture for a method of deploying a drone base station according to an embodiment of this application. Figure 1 In the above architecture, at least one of data acquisition device 101, processing device 102 and display device 103 is included.

[0077] It is understood that the structures illustrated in the embodiments of this application do not constitute a specific limitation on the deployment system architecture of unmanned aerial vehicle (UAV) base stations. In other feasible embodiments of this application, the above architecture may include more or fewer components than illustrated, or combine some components, or divide some components, or arrange different components, which can be determined according to the actual application scenario and is not limited here. Figure 1 The components shown can be implemented in hardware, software, or a combination of both.

[0078] In the specific implementation process, the data acquisition device 101 may include an input / output interface or a communication interface, and the data acquisition device 101 can be connected to the processing device through the input / output interface or the communication interface.

[0079] Processing device 102 can determine the deployment plan for drone base stations in the area to be deployed.

[0080] The display device 103 can also be a touch screen or the screen of a terminal device, used to store data while displaying the above-mentioned content, so as to realize interaction with the user.

[0081] It should be understood that the aforementioned processing device can be implemented by a processor reading instructions from memory and executing those instructions, or it can be implemented by a chip circuit.

[0082] Furthermore, the network architecture and business scenarios described in the embodiments of this application are for the purpose of more clearly illustrating the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions provided in the embodiments of this application. As those skilled in the art will know, with the evolution of network architecture and the emergence of new business scenarios, the technical solutions provided in the embodiments of this application are also applicable to similar technical problems.

[0083] The technical solution of this application will be described in detail below with reference to specific embodiments:

[0084] Figure 2 A flowchart illustrating a method for deploying a drone base station provided in this application embodiment. Figure 1 ,like Figure 2 As shown, the method includes:

[0085] S201. Obtain the simulation model parameters of the area to be deployed, and generate a signal coverage simulation model based on ray tracing according to the simulation model parameters.

[0086] The simulation model parameters include ray tracing parameters.

[0087] Specifically, obtain the transmission model parameters used for ray tracing.

[0088] Specifically, obtain user location parameters.

[0089] Specifically, it acquires building information data for the area to be deployed.

[0090] For example, collecting the parameters needed to build a signal coverage prediction model includes the transmission model required for ray tracing, user and drone locations, and building information data.

[0091] S202. Obtain the physical environment data of the area to be deployed.

[0092] For example, physical environment data such as actual building height, outline, and user location are collected in different campus areas.

[0093] S203. Configure the initial neural network model based on the ray tracing-based signal coverage simulation model.

[0094] Specifically, the input data for the initial neural network model is determined based on the physical environment data in the ray-tracing-based signal coverage simulation model.

[0095] Specifically, the output data of the initial neural network model is determined based on the simulation results data in the ray tracing-based signal coverage simulation model.

[0096] Specifically, the initial neural network model is configured based on the input and output data.

[0097] S204. Based on the physical environment data, train and predict the initial neural network model to determine the signal strength map of the area to be deployed.

[0098] S205. Based on the signal strength map, determine the deployment plan for drone base stations in the area to be deployed.

[0099] This embodiment provides a method for deploying a drone base station, including: acquiring simulation model parameters of the area to be deployed, and generating a ray-tracing-based signal coverage simulation model based on the simulation model parameters; acquiring physical environment data of the area to be deployed; configuring an initial neural network model based on the ray-tracing-based signal coverage simulation model; training and predicting the initial neural network model based on the physical environment data to determine the signal strength map of the area to be deployed; and determining a drone base station deployment scheme for the area to be deployed based on the signal strength map. Compared to existing technologies, traditional signal coverage prediction methods are affected by various factors in complex real-world environments, leading to an inability to accurately predict the signal coverage of drones to ground users, resulting in low deployment efficiency for drone base stations. This application solves the data scarcity problem by providing physically realistic signal propagation samples through ray tracing; it preserves spatial multi-scale features through the encoding-decoding structure of the neural network model, adapting to nonlinear propagation such as building obstruction and multipath effects; thereby accurately predicting signal coverage and achieving accurate deployment of drone base stations, thus improving deployment efficiency.

[0100] Figure 3 A flowchart illustrating a method for deploying a drone base station provided in this application embodiment. Figure 2 ,like Figure 3 As shown, the specific implementation steps of S201 above include:

[0101] S301. Based on user location parameters and building information data, display the spatial relationship between buildings, user nodes and candidate UAV locations in the simulation environment to obtain a three-dimensional environmental model.

[0102] S302. Initialize the ray tracing parameters according to the transmission model parameters.

[0103] The ray tracing parameters include emission power, path loss model, and fading factor.

[0104] S303. Configure the simulation physical layer according to the ray tracing parameters and import the 3D environment model to obtain a signal coverage simulation model based on ray tracing.

[0105] For example, the spatial relationships between buildings, user nodes, and candidate drone locations are first demonstrated in a simulation environment.

[0106] Furthermore, set and initialize ray tracing parameters, including transmit power, path loss model, and fading factor, to ensure parameter consistency across nodes.

[0107] Furthermore, the simulation physical layer is configured, the 3D model of the building and other environmental characteristics are imported, and the initial simulation model is established.

[0108] Furthermore, an information display platform will be created to display the model construction status and key parameters in real time.

[0109] Furthermore, the correctness of the initial model configuration is verified, and the collected parameters are adjusted to ensure the accuracy of subsequent simulation data.

[0110] In one possible embodiment, Figure 8 This is a schematic diagram illustrating the simulation model construction and parameter configuration provided in the embodiments of this application, such as... Figure 8 As shown, it covers key aspects such as parameter acquisition, 3D display, platform configuration, and calibration verification.

[0111] Optionally, the basic data acquisition layer includes an initial parameter collection module that acquires data through dual channels: basic environmental parameters (such as terrain data and obstacle distribution) and wireless channel parameters (signal propagation characteristics obtained through ray tracing). The user and drone position acquisition module uses real-time positioning technology to obtain the three-dimensional coordinate data of the operator and the drone.

[0112] Optionally, the environment modeling layer includes a 3D environment model import module that integrates Geographic Information System (GIS) data and Building Information Modeling (BIM) data to construct a high-precision virtual scene. The initial simulation model creation module creates a dynamic simulation environment based on a physics engine (such as Unreal Engine or Gazebo), which includes the following sub-models: UAV dynamics model, wireless channel propagation model, and environmental interaction model.

[0113] Optionally, the visualization verification layer, a module for model display and parameter verification, enables both 3D visualization rendering and parameter validation.

[0114] Specifically, the visualization part uses LOD (Level of Detail) technology to adjust the model accuracy at different viewing distances.

[0115] Specifically, parameter verification includes three dimensions: physical authenticity verification (such as whether the gravitational acceleration is consistent), communication performance verification (such as signal attenuation model), and task logic verification (such as the rationality of path planning).

[0116] Optionally, a feedback optimization layer and parameter feedback and adjustment module establish a two-way correction mechanism:

[0117] Specifically, lateral correction involves making local adjustments to specific parameters (such as the electrical conductivity of the obstacle material).

[0118] Specifically, longitudinal correction: when a systematic deviation occurs, the model reconstruction process is triggered.

[0119] In addition, a PID control algorithm is used to achieve dynamic balance in parameter adjustment.

[0120] Optional, a cyclical verification mechanism, where the process forms a PDCA (Plan-Do-Check-Act) cycle:

[0121] First, the initial modeling (Plan).

[0122] Secondly, simulation run (Do).

[0123] Next, parameter validation (Check).

[0124] Finally, feedback adjustment (Act).

[0125] When the verification indicators (such as positioning error > 0.5m or communication latency > 100ms) exceed the limit, the system automatically backtracks to the model reconstruction stage, forming a continuous optimization closed loop.

[0126] It should be noted that this process achieves rapid construction and iterative optimization of the simulation environment through modular design, making it particularly suitable for applications requiring high-precision environmental simulation, such as UAV communication and path planning. Data flow between modules uses standardized interfaces (such as VRML format for 3D model transmission) to ensure system scalability and compatibility.

[0127] In this embodiment, to support the efficient operation of the overall process, a visual simulation and training platform was constructed. The platform covers ray tracing parameter configuration, 3D environment construction, and data standardization processing. It supports result visualization and real-time feedback of model effects, and has good system integration and scalability, thereby improving the deployment efficiency of UAV base stations.

[0128] Figure 4 A flowchart illustrating a method for deploying a drone base station provided in this application embodiment. Figure 3 ,like Figure 4 As shown, the specific implementation steps of S204 above include:

[0129] S401. Perform data normalization and format conversion on the physical environment data to obtain standard physical environment data.

[0130] For example, the collected data is preprocessed, including data cleaning, normalization, and format conversion, to ensure that the data meets the input requirements of deep learning networks.

[0131] Furthermore, the preprocessed data is abstracted into virtual data to form a standardized dataset for subsequent training and validation of neural networks.

[0132] Furthermore, the aforementioned virtualized data is stored in the data management module, and the data quality and format are monitored and feedback is provided through the information display platform.

[0133] In one possible embodiment, Figure 9 This is a schematic diagram of real-world data acquisition and preprocessing provided in the embodiments of this application, such as... Figure 9 The diagram illustrates the entire real-world data preprocessing workflow, including every step from raw data acquisition to standardized input.

[0134] Optional modular phase division, in which the process adopts a four-layer progressive architecture:

[0135] Specifically, the data acquisition layer involves dual-source parallel acquisition (static building attribute data and dynamic user behavior data).

[0136] Specifically, the conversion layer includes three operations: cleaning, normalization, and format conversion.

[0137] Specifically, the storage quality control layer integrates anomaly detection and persistent data storage.

[0138] Specifically, the output encapsulation layer generates structured datasets that conform to industry standards.

[0139] Optional, refined processing steps:

[0140] First, data cleaning: missing value handling (median filling / record deletion), outlier detection (based on the 3σ principle of Z-score), and deduplication of duplicate data (hash fingerprint algorithm).

[0141] Secondly, normalization transformation: continuous value standardization (Min-Max scaling), categorical value encoding (one-hot encoding / label encoding), and time series alignment (interpolation to complete timestamps).

[0142] Finally, format conversion: structured data conversion (CSV → Parquet) and spatial data conversion (Shapefile → GeoJSON).

[0143] Optional, dynamic data flow, where the main process is: collection → cleaning → normalization → storage → output.

[0144] Specifically, the abnormal branch is triggered when data anomalies are detected. In the case of minor anomalies, automatic correction (such as time format standardization) is performed; in the case of severe anomalies, the process is backtracked to the data acquisition stage to reacquire the data.

[0145] Specifically, an adaptive processing mechanism with feedback loops is formed to ensure controllable data quality.

[0146] Optional, key quality control points:

[0147] First, the data cleaning station sets up three verification rules: business rule verification (e.g., building area > 0), statistical distribution verification (e.g., user dwell time conforms to a normal distribution), and spatial topology verification (building boundaries do not self-intersect).

[0148] Secondly, normalization monitoring: box plots are used to visualize feature distribution to ensure that the transformed data meets the expected range.

[0149] Finally, format conversion verification: ensure that the field type and length conform to the metadata definition through schema validation.

[0150] Optional exception handling mechanism, through the construction of a three-level response system:

[0151] Specifically, Level 1 (system level): Automatically correct formatting errors (such as coordinate system transformations).

[0152] Specifically, Level 2 (business level): Manual intervention to handle logical errors (such as incorrect building function classification).

[0153] Specifically, at level three (data source level): trigger the data re-collection mechanism and record the data source reliability score.

[0154] In addition, abnormal data is placed in an isolation zone, and issues are traced through version control.

[0155] Optional quality assessment system:

[0156] First, implement the DQA (Data Quality Assessment) three-dimensional model:

[0157] Specifically, the completeness metric is: required field fill rate > 99.5%.

[0158] Specifically, the accuracy metric is: spatial data topology error rate <0.1%.

[0159] Specifically, the consistency metric is: 100% matching rate of cross-table join fields.

[0160] Secondly, an automated quality inspection toolkit is adopted:

[0161] Specifically, Great Expectations performs expected data validation.

[0162] Specifically, Deepu constructs data quality constraint rules.

[0163] Specifically, OpenRefine enables interactive data cleaning.

[0164] It should be noted that this process, through modular design and closed-loop quality control, ensures that the output data conforms to the GB / T36344-2018 "Information Technology Data Quality Evaluation Index" standard, providing a reliable data foundation for subsequent data analysis and machine learning modeling. The anomaly backtracking mechanism and quality gates set up in the process effectively guarantee the robustness of the data processing chain.

[0165] S402. Obtain the scene source of the standard physical environment data, and add scene tags to the standard physical environment data according to the scene source.

[0166] S403. Based on the standard physical environment data with added scene labels, determine the prediction input dataset and the training input dataset.

[0167] S404. Add signal intensity labels to the training input dataset and generate the training output dataset based on the signal intensity labels.

[0168] S405 trains the initial neural network model based on the training input dataset and the training output dataset to obtain a trained neural network model.

[0169] S406. Perform prediction processing on the prediction input dataset based on the trained neural network model to obtain the signal strength map of the area to be deployed based on the prediction results.

[0170] For example, deploy and configure the U-Net neural network model, and instantiate the model in a simulation environment.

[0171] Furthermore, the preprocessed building information and user location are used as inputs into the neural network model.

[0172] Specifically, data processing for the same scenario.

[0173] Specifically, spatial coordinate calibration is performed on building information collected within the same campus area to ensure that the data format is consistent with the expected format.

[0174] Specifically, user location data in the same area is labeled, binarized, and aligned with building information in the same coordinate system.

[0175] Specifically, the calibrated and tagged building information is integrated with user location data at a fixed ratio to generate standardized input data.

[0176] Specifically, the generated standardized data is input into a neural network model to train and predict signal coverage within the same scene.

[0177] Specifically, data processing for different scenarios

[0178] Specifically, building information from different campuses or regions is normalized to eliminate scale and distribution differences and ensure data consistency across numerical ranges.

[0179] Specifically, user location data for different scenarios are binarized, and a unique scenario identifier is assigned to each scenario for subsequent differentiation.

[0180] Specifically, based on the characteristics of each scenario, the normalized building information and the identified user location data are converted into standard input formats respectively.

[0181] Specifically, standardized data from multiple scenarios are input into a neural network model in the form of a joint dataset to train the model and enhance its generalization ability to predict signal coverage in different scenarios.

[0182] Specifically, the model parameters are adjusted through network training until the model can accurately output the received signal strength data corresponding to the location of each drone base station.

[0183] Specifically, the model is validated using a test set, and the prediction results are stored in the data management module. At the same time, the training and prediction status is displayed in real time through an information display platform.

[0184] Figure 10 This application provides a schematic diagram of U-Net model deployment, training, and prediction in its embodiments. Figure 4 ,like Figure 10 The diagram shows the deployment flowchart of U-Net, including input processing, model training, performance evaluation, and prediction output.

[0185] Optional, modular phase division, the process adopts a five-layer progressive architecture:

[0186] First, the data preparation layer includes two modules: input preprocessing and data calibration.

[0187] Secondly, the model building layer: feature extraction and reconstruction are implemented based on the U-Net architecture.

[0188] Secondly, the training optimization layer integrates parameter iteration and loss monitoring mechanisms.

[0189] Then, the effect verification layer: quantifies the model performance through multi-dimensional evaluation indicators.

[0190] Finally, the encapsulation layer is output, which generates standardized prediction results and model weight files.

[0191] Optional, refined processing steps include:

[0192] First, the input preprocessing includes: image normalization (scaling pixel values ​​to the [0,1] range), channel splitting (for multimodal MRI data), and random cropping (generating 512×512 pixel training patches).

[0193] Secondly, data calibration includes: spatial alignment (multimodal registration using affine transformation), intensity normalization (Z-score normalization of brain tumor regions), and label mapping (converting expert annotations into one-hot encoding).

[0194] Secondly, the model architecture:

[0195] Specifically, the encoder path consists of 4 convolutional blocks (each containing two 3×3 convolutions + BN + ReLU).

[0196] Specifically, the decoder path corresponds to the 4-layer upsampling block (combined with skip connections).

[0197] Specifically, feature fusion: the third-layer skip connection adopts a channel splicing method.

[0198] Finally, training strategy:

[0199] Specifically, the loss function is a weighted combination of Dice loss and cross-entropy loss (weight ratio 3:1).

[0200] Specifically, the optimizer is AdamW (learning rate = 1e-4, weight decay = 1e-5).

[0201] Specifically, regularization: L2 regularization + Dropout (rate = 0.5).

[0202] Optional, dynamic data flow, where the main process is: preprocessing → calibration → training → validation → output.

[0203] Specifically, the abnormal branch is triggered when the validation metric (such as Dice coefficient < 0.7) fails to meet the standard: if there is slight underfitting, the number of training epochs is increased (epoch + 20); if there is severe overfitting, an early stopping mechanism is introduced (patience = 5).

[0204] In addition, an adaptive training mechanism with feedback loops is formed to ensure the model's convergence stability.

[0205] Optional key technology nodes include:

[0206] Specifically, skip connections enable cross-layer connections between encoder feature maps and decoders, preserving high-resolution spatial information and improving segmentation boundary accuracy.

[0207] Specifically, in terms of loss function design, Dice loss alleviates the class imbalance problem, while cross-entropy loss ensures pixel-level classification accuracy.

[0208] Specifically, the training process involves monitoring and plotting the loss curve and Dice coefficient changes in real time, and saving the optimal model weights (based on validation set performance) for each round.

[0209] Optional exception handling mechanism, constructing a three-level tuning strategy:

[0210] Specifically, Level 1 (parameter level): dynamically adjust the learning rate (using cosine annealing).

[0211] Specifically, at the second level (architectural level): increase or decrease the number of hop connections (3 → 4 or 2).

[0212] Specifically, at the third level (data level): expand the training set (introduce data augmentation strategies).

[0213] In addition, abnormal data enters the isolation zone, and the training process can be traced back through version control.

[0214] Optional, a quality assessment system, implementing a multi-dimensional assessment approach:

[0215] Specifically, pixel-level metrics include: accuracy and sensitivity.

[0216] Specifically, regional indicators include: Dice similarity coefficient (DSC) and Hausdorff distance (HD).

[0217] Specifically, clinical acceptability: expert visual assessment (Turing test).

[0218] Additionally, the output criteria are: when DSC > 0.8 and HD < 5mm, the model passes validation.

[0219] It should be noted that this process, through modular design and closed-loop quality control, ensures that the U-Net model meets clinical application standards in medical image segmentation tasks. The skip connection optimization mechanism and multi-scale loss function set in the process effectively solve the balance problem between detail preservation and semantic understanding in traditional convolutional networks.

[0220] In this embodiment, U-Net was selected as the basic architecture for the neural network structure design. Its ability to model local and global features in parallel enables pixel-level signal intensity map prediction. This network structure is not only suitable for environments with complex building distributions, but also preserves multi-scale spatial information, improving the model's prediction stability under nonlinear propagation factors such as occlusion and fading, thereby increasing the deployment efficiency of UAV base stations.

[0221] Figure 5 A flowchart illustrating a method for deploying a drone base station provided in this application embodiment. Figure 4 ,like Figure 5 As shown, the specific implementation steps of S205 above include:

[0222] S501. Obtain the preset signal strength threshold.

[0223] S502. Based on the preset signal strength threshold, the signal strength map is filtered to obtain a set of feasible flight locations.

[0224] S503. Based on the set of feasible flight locations, determine the deployment plan for drone base stations in the area to be deployed.

[0225] For example, the predicted received signal strength data is imported into the subsequent processing module.

[0226] Furthermore, a predetermined signal threshold is set, and the predicted signal strength of each UAV's flight location is used to filter and initially form a set of feasible flight locations.

[0227] Furthermore, the screening results are optimized and analyzed, taking into account factors such as coverage overlap between drones, airspace utilization, and network connectivity, to further determine the final drone base station locations.

[0228] Furthermore, the final planning results will be presented through an information display platform, and the drone location planning data will be output to the automatic deployment module to realize the automatic planning and optimized deployment of drone base stations.

[0229] Figure 11 This is a schematic diagram of signal coverage prediction result processing and UAV location planning provided in the embodiments of this application, such as... Figure 11 As shown, a flowchart of path optimization and drone deployment is presented, highlighting the closed-loop control logic of prediction-analysis-deployment.

[0230] Optional, modular phase division, with the process adopting a three-level progressive architecture:

[0231] Specifically, the environment perception layer: spatial feature extraction is achieved based on an improved U-Net.

[0232] Specifically, the decision optimization layer integrates multi-objective optimization algorithms with dynamic deployment strategies.

[0233] Specifically, the execution feedback layer enables the automatic deployment and status monitoring of drone swarms.

[0234] Optional, fine-grained processing module:

[0235] First, the environment prediction module takes as input multi-source sensor data (RGB-D imagery, LiDAR point cloud), processes it using an improved U-Net (with added attention mechanism) to generate a semantic segmentation map, and outputs obstacle distribution and signal attenuation heatmap.

[0236] Secondly, the decision-making module is optimized, with threshold settings including dynamic signal strength thresholds (automatically adjusted according to task type), candidate selection based on the NSGA-II algorithm for multi-objective optimization (coverage / energy consumption / latency), and path planning using a hybrid A*-Dijkstra algorithm to generate the optimal route.

[0237] Finally, the deployment execution module enables cluster collaboration: it implements drone formation and dynamic adjustment based on the Pigeon Optimization (PIO) algorithm, and monitors signal strength (RSSI) in real time to trigger path replanning.

[0238] Optional closed-loop control logic, where the main loop is: environmental perception → decision optimization → deployment execution → effect evaluation.

[0239] Specifically, the feedback branch: when the signal strength is detected to be below the threshold or a new obstacle appears: if it is a minor anomaly, the local path is fine-tuned (using the RRT* algorithm); if it is a severe anomaly, the global path is replanned (triggering U-Net to re-predict).

[0240] In addition, a continuous optimization loop encompassing prediction, deployment, and feedback is formed to ensure the system adapts to dynamic environments.

[0241] Optional, key algorithm nodes:

[0242] First, the improved U-Net adds a spatial attention module (CBAM) to improve obstacle detection accuracy and adopts a deep supervision mechanism to accelerate network convergence.

[0243] Secondly, hybrid path planning is used: global planning: A* algorithm generates the skeleton path; local optimization: Dijkstra's algorithm refines waypoints.

[0244] Finally, a dynamic deployment strategy is adopted, which uses a market mechanism (auction algorithm) to allocate tasks based on the coverage area division of the Voronoi diagram.

[0245] Optional exception handling mechanism, constructing a three-level response system:

[0246] Specifically, Level 1 (signal level): when RSSI < -75dBm, the backup link is activated.

[0247] Specifically, at level two (path level): path replanning is triggered when the proportion of obstacles is greater than 30%.

[0248] Specifically, Level 3 (system level): If the drone loses contact for more than 30 seconds, the backup drone will be activated.

[0249] In addition, abnormal data enters the isolation zone and is used for fault diagnosis through a Bayesian network.

[0250] Optional, a quality assessment system, implementing a multi-dimensional assessment approach:

[0251] Specifically, the coverage metric is: effective coverage area / total area.

[0252] Specifically, the efficiency metric is: deployment time / path length.

[0253] Specifically, robustness metrics include: anomaly recovery time / task completion rate.

[0254] Specifically, the closed-loop control effect is achieved by continuously optimizing the decision-making strategy through reinforcement learning (PPO algorithm).

[0255] It should be noted that this process, through the fusion of deep learning and classical optimization algorithms, achieves autonomous decision-making and dynamic adaptation for UAV deployment. A specially designed two-layer feedback mechanism (local fine-tuning and global replanning) ensures high system reliability in complex environments and meets real-time requirements.

[0256] In this embodiment, during the prediction result application stage, the present invention uses a method of setting a received signal strength threshold to filter location regions that meet communication requirements from the signal image output by U-Net. This provides a region determination method based on image prediction results, which is simple, fast, and adjustable, and can provide direct parameters for subsequent deployment schemes, thereby improving the deployment efficiency of UAV base stations.

[0257] This application also provides a possible embodiment. Figure 12 The overall process diagram provided for the embodiments of this application is as follows: Figure 12 As shown, it includes four core modules: signal simulation modeling, data acquisition and processing, neural network modeling and prediction, and base station deployment optimization, and the following implementation steps are designed:

[0258] Step 1: Create a signal coverage prediction model in the simulation environment, set the ray tracing parameters, user position, UAV position resolution, and building information input, and establish the initial simulation model configuration.

[0259] Step 2: Collect physical environment data such as actual building height, outline, and user distribution in different areas, and preprocess the data to abstract it into virtual data and convert it into a standard format for use by deep neural networks.

[0260] Step 3: Deploy and configure the U-Net neural network model, taking the preprocessed building information and user points of interest as input, and output the received signal strength data corresponding to the location of each UAV base station through network training and prediction.

[0261] Step 4: Based on the received signal strength predicted in Step 3, and combined with the set signal threshold, the flight positions of each UAV are screened to form a set of feasible flight positions that are higher than the threshold, and the automatic planning and optimized deployment of UAV base station positions are completed.

[0262] It should be noted that this flowchart constructs a complete closed-loop system from data acquisition to decision optimization. The process begins with multi-source data acquisition and parameter setting. After data acquisition and processing, it enters the U-Net model training, prediction, and optimization stage, and the generated control commands drive the deployment of the UAV. The data flow is divided into two parallel paths: one is input into the decision module after data preprocessing, and the other directly generates decision results through model prediction. Finally, continuous iteration is achieved through result display and parameter adjustment. Key nodes include multi-source data fusion, U-Net model optimization, and closed-loop feedback control to ensure that the system can achieve adaptive deployment in dynamic environments.

[0263] It should be noted that the modules are logically linked through a three-level control mechanism: 1) The data layer uses multi-source heterogeneous data fusion to provide high-quality input for model training; 2) The algorithm layer integrates U-Net prediction and hybrid path planning, triggering dynamic replanning through confidence thresholds; 3) The decision layer optimizes deployment strategies based on reinforcement learning, feeding back the actual coverage effect to the data preprocessing and model training stages. In abnormal situations, a dual-mode switch (local fine-tuning / global replanning) is initiated to ensure that the system maintains service continuity when the signal weakens or obstacles change abruptly.

[0264] It should also be noted that this technical solution encompasses the entire process, from the collection of real-world geographic data and the modeling and ray-tracing simulation of building spatial information to the training and prediction output of the U-Net neural network model. During the experimental phase, several typical campus areas were selected as test scenarios. A 3D simulation environment was constructed to generate datasets for training and testing, and model training and performance verification were completed. Experimental results show that the signal strength prediction map output by the system performs well in terms of spatial structure consistency and physical trend fitting, demonstrating high practical application value.

[0265] In one possible embodiment, Figure 13This is a schematic diagram illustrating the building environment input and signal prediction effect provided in an embodiment of this application, such as... Figure 13 As shown, in the experiment, the normalized building height map and the point of interest location map were used as inputs to the neural network, and the signal intensity map output by the neural network was... Figure 13 The bottom left image shows a spatial distribution that corresponds to the actual results generated by ray tracing simulation. Figure 13 The high consistency between the right and left sides indicates that the method has good prediction accuracy and model convergence ability.

[0266] It should be noted that, Figure 13 This is only an example of effectiveness and does not affect the scope of protection.

[0267] Furthermore, in more complex communication scenarios, when there are multiple points of interest (i.e., multiple user receiving locations) and each location has specific threshold requirements for signal quality, this application can also achieve prediction and feasible solution extraction for multi-target deployment areas.

[0268] In one possible embodiment, Figure 14 This is a schematic diagram illustrating the multi-point of interest signal coverage prediction and deployable area filtering provided in an embodiment of this application, as shown below. Figure 14 As shown, the system targets three points of interest (respectively...). Figure 14 (The three points marked with dark red squares on the left) are assigned different signal strength thresholds. Their respective signal maps are predicted by the model and then fused and judged to select the set of UAV deployment locations that simultaneously meet the communication requirements of all points of interest, i.e., the blue area marked in the figure.

[0269] It should be noted that, Figure 14 This is only an example of effectiveness and does not affect the scope of protection.

[0270] In this embodiment, the system achieves three core advantages through the integration of deep learning and control theory: First, improved efficiency: automated data processing and model optimization shorten deployment time and improve coverage accuracy. Second, enhanced robustness: a closed-loop feedback mechanism reduces anomaly recovery time and improves task completion rate. Finally, optimized scalability: the modular design supports multi-UAV collaboration, significantly reducing manual intervention costs in practical applications, making it suitable for scenarios requiring dynamic deployment, such as emergency communication and precision agriculture.

[0271] Figure 6 This is a schematic diagram of the structure of a drone base station deployment device provided in an embodiment of this application, as shown below. Figure 6 As shown, the device includes: a first acquisition module 61, a second acquisition module 62, a configuration module 63, a first determination module 64, and a second determination module 65.

[0272] The first acquisition module 61 is used to acquire the simulation model parameters of the area to be deployed, and generate a signal coverage simulation model based on ray tracing according to the simulation model parameters; wherein, the simulation model parameters include ray tracing parameters;

[0273] The second acquisition module 62 is used to acquire physical environment data of the area to be deployed.

[0274] Configuration module 63 is used to configure the initial neural network model based on the ray tracing-based signal coverage simulation model;

[0275] The first determining module 64 is used to train and predict the initial neural network model based on physical environment data in order to determine the signal strength map of the area to be deployed.

[0276] The second determining module 65 is used to determine the deployment scheme of the drone base station in the area to be deployed based on the signal strength map.

[0277] In one possible design, the simulation model parameters of the area to be deployed are obtained, including:

[0278] The first acquisition module 61 is also used to acquire transmission model parameters for ray tracing;

[0279] Obtain user location parameters;

[0280] Obtain building information data for the area to be deployed.

[0281] In one possible design, a ray-tracing-based signal coverage simulation model is generated based on simulation model parameters, including:

[0282] The first acquisition module 61 is also used to display the spatial relationship between buildings, user nodes and candidate drone positions in the simulation environment based on user location parameters and building information data, so as to obtain a three-dimensional environmental model.

[0283] Initialize the ray tracing parameters based on the transmission model parameters; the ray tracing parameters include transmit power, path loss model, and fading factor.

[0284] Based on the ray tracing parameters, configure the simulation physics layer and import the 3D environment model to obtain a ray tracing-based signal coverage simulation model.

[0285] In one possible design, based on a ray-tracing-based signal coverage simulation model, an initial neural network model is configured, including:

[0286] Configuration module 63 is also used to determine the input data of the initial neural network model based on the physical environment data in the ray-tracing-based signal coverage simulation model.

[0287] Based on the simulation results data in the ray tracing-based signal coverage simulation model, determine the output data of the initial neural network model;

[0288] Configure the initial neural network model based on the input and output data.

[0289] In one possible design, an initial neural network model is trained and predicted based on physical environment data to determine the signal strength map of the area to be deployed, including:

[0290] The first determining module 64 is also used to perform data normalization and format conversion processing on the physical environment data to obtain standard physical environment data;

[0291] Based on standard physical environment data, the initial neural network model is trained and predicted to determine the signal strength map of the area to be deployed.

[0292] In one possible design, an initial neural network model is trained and predicted based on standard physical environment data to determine the signal strength map of the area to be deployed, including:

[0293] The first determining module 64 is also used to obtain the scene source of the standard physical environment data and add scene tags to the standard physical environment data according to the scene source;

[0294] Based on standard physical environment data with added scene labels, determine the prediction input dataset and the training input dataset;

[0295] Add signal intensity labels to the training input dataset and generate the training output dataset based on the signal intensity labels;

[0296] The initial neural network model is trained based on the training input dataset and the training output dataset to obtain a trained neural network model.

[0297] The trained neural network model is used to perform prediction processing on the input dataset to obtain the signal strength map of the area to be deployed based on the prediction results.

[0298] In one possible design, the deployment scheme for drone base stations in the area to be deployed is determined based on the signal strength map, including:

[0299] The second determining module 65 is also used to obtain a preset signal strength threshold;

[0300] Based on a preset signal strength threshold, the signal strength map is filtered to obtain a set of feasible flight locations;

[0301] Based on the set of feasible flight locations, determine the deployment plan for drone base stations in the area to be deployed.

[0302] This embodiment provides a drone base station deployment device that can execute a drone base station deployment method described in the above embodiment. Its implementation principle and technical effects are similar, and will not be repeated here.

[0303] In a specific implementation of the aforementioned method for deploying a drone base station, each module can be implemented as a processor. The processor can execute computer execution instructions stored in the memory, thereby enabling the processor to execute the aforementioned method for deploying a drone base station.

[0304] Figure 7 This is a schematic diagram of the structure of a drone base station deployment device provided in an embodiment of this application. Figure 7 As shown, the deployment device 70 for the drone base station includes at least one processor 71 and a memory 72. The deployment device 70 also includes a communication component 73. The processor 71, memory 72, and communication component 73 are connected via a second bus 74.

[0305] In the specific implementation process, at least one processor 71 executes computer execution instructions stored in memory 72, causing at least one processor 71 to execute a drone base station deployment method executed by the deployment device side of the above-mentioned drone base station.

[0306] The specific implementation process of processor 71 can be found in the above method embodiments, and its implementation principle and technical effect are similar. It will not be repeated here.

[0307] In the above embodiments, it should be understood that the processor can be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), etc. The general-purpose processor can be a microprocessor or any conventional processor. The steps of the method disclosed in this invention can be directly implemented by a hardware processor, or implemented by a combination of hardware and software modules within the processor.

[0308] The memory may include high-speed RAM, and may also include non-volatile storage (NVM), such as at least one disk storage.

[0309] The second bus can be an Industry Standard Architecture (ISA) bus, a Peripheral Component Interconnect (PCI) bus, or an Extended Industry Standard Architecture (EISA) bus, etc. Buses can be categorized as address buses, data buses, control buses, etc. For ease of illustration, the buses shown in the accompanying drawings are not limited to a single bus or a single type of bus.

[0310] The above description of the functions implemented by the deployment equipment and main control equipment for drone base stations illustrates the solution provided by the embodiments of the present invention. It is understood that, in order to achieve the above functions, the deployment equipment or main control equipment for drone base stations includes hardware structures and / or software modules corresponding to the execution of each function. By combining the units and algorithm steps of the various examples described in the embodiments of the present invention, the embodiments of the present invention can be implemented in hardware or a combination of hardware and computer software. Whether a function is executed by hardware or by computer software driving hardware depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of the technical solution of the embodiments of the present invention.

[0311] This application also provides a computer-readable storage medium storing computer-executable instructions, which, when executed by a processor, are used to implement the above-described method for deploying a drone base station.

[0312] The aforementioned readable storage medium can be implemented by any type of volatile or non-volatile storage device or a combination thereof, such as static random access memory (SRAM), electrically erasable programmable read-only memory (EEPROM), erasable programmable read-only memory (EPROM), programmable read-only memory (PROM), read-only memory (ROM), magnetic storage, flash memory, magnetic disk, or optical disk. The readable storage medium can be any available medium accessible to a general-purpose or special-purpose computer.

[0313] An exemplary readable storage medium is coupled to a processor, enabling the processor to read information from and write information to the readable storage medium. Alternatively, the readable storage medium can be an integral part of the processor. The processor and the readable storage medium can reside in application-specific integrated circuits (ASICs). Alternatively, the processor and the readable storage medium can exist as discrete components in the deployment equipment or main control equipment of a drone base station.

[0314] This application also provides a computer program product, comprising: a computer program stored in a readable storage medium, wherein at least one processor of the deployment device of the drone base station can read the computer program from the readable storage medium, and the at least one processor executes the computer program to cause the deployment device of the drone base station to perform the scheme provided in any of the above embodiments.

[0315] Those skilled in the art will understand that all or part of the steps of the above method embodiments can be implemented by hardware related to program instructions. The aforementioned program can be stored in a computer-readable storage medium. When the program is executed, it performs the steps of the above method embodiments; and the aforementioned storage medium includes various media capable of storing program code, such as ROM, RAM, magnetic disk, or optical disk.

[0316] The technical solutions of this application have been described above with reference to the preferred embodiments shown in the accompanying drawings. However, it is readily understood by those skilled in the art that the scope of protection of this application is obviously not limited to these specific embodiments. The above embodiments are only used to illustrate the technical solutions of this application and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features therein. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application.

Claims

1. A method for deploying a drone base station, characterized in that, include: Obtain the simulation model parameters of the area to be deployed, and generate a signal coverage simulation model based on ray tracing according to the simulation model parameters; wherein, the simulation model parameters include ray tracing parameters; Obtain the physical environment data of the area to be deployed; Based on the ray tracing-based signal coverage simulation model, configure the initial neural network model; Based on the physical environment data, the initial neural network model is trained and predicted to determine the signal strength map of the area to be deployed. Based on the signal strength map, determine the deployment plan for drone base stations in the area to be deployed.

2. The method according to claim 1, characterized in that, The process of obtaining the simulation model parameters for the region to be deployed includes: Obtain the transmission model parameters used for ray tracing; Obtain user location parameters; Obtain building information data for the area to be deployed.

3. The method according to claim 2, characterized in that, The step of generating a ray-tracing-based signal coverage simulation model based on the simulation model parameters includes: Based on the user location parameters and the building information data, the spatial relationship between buildings, user nodes and candidate drone locations is displayed in the simulation environment to obtain a three-dimensional environmental model. Based on the transmission model parameters, initialize the ray tracing parameters; wherein, the ray tracing parameters include transmit power, path loss model, and fading factor; Based on the ray tracing parameters, configure the simulation physics layer and import the 3D environment model to obtain a ray tracing-based signal coverage simulation model.

4. The method according to any one of claims 1 to 3, characterized in that, The step of configuring the initial neural network model based on the ray-tracing-based signal coverage simulation model includes: Based on the physical environment data in the ray-tracing-based signal coverage simulation model, the input data of the initial neural network model is determined; Based on the simulation results data in the ray tracing-based signal coverage simulation model, determine the output data of the initial neural network model; Configure the initial neural network model based on the input data and the output data.

5. The method according to claim 4, characterized in that, The step of training and predicting the initial neural network model based on the physical environment data to determine the signal strength map of the area to be deployed includes: The physical environment data is subjected to data normalization and format conversion to obtain standard physical environment data; Based on the standard physical environment data, the initial neural network model is trained and predicted to determine the signal strength map of the area to be deployed.

6. The method according to claim 5, characterized in that, The step of training and predicting the initial neural network model based on the standard physical environment data to determine the signal strength map of the area to be deployed includes: Obtain the scene source of the standard physical environment data, and add scene tags to the standard physical environment data according to the scene source; Based on the standard physical environment data with the scene labels added, determine the prediction input dataset and the training input dataset; Add signal intensity labels to the training input dataset, and generate a training output dataset based on the signal intensity labels; The initial neural network model is trained based on the training input dataset and the training output dataset to obtain a trained neural network model. The trained neural network model is used to perform prediction processing on the prediction input dataset to obtain the signal strength map of the area to be deployed based on the prediction results.

7. The method according to claim 6, characterized in that, The step of determining the drone base station deployment plan for the area to be deployed based on the signal strength map includes: Obtain the preset signal strength threshold; Based on the preset signal strength threshold, the signal strength map is filtered to obtain a set of feasible flight locations; Based on the set of feasible flight locations, a deployment plan for drone base stations in the area to be deployed is determined.

8. A deployment device for a drone base station, characterized in that, include: The first acquisition module is used to acquire simulation model parameters of the area to be deployed, and generate a signal coverage simulation model based on ray tracing according to the simulation model parameters; wherein, the simulation model parameters include ray tracing parameters; The second acquisition module is used to acquire physical environment data of the area to be deployed. A configuration module is used to configure an initial neural network model based on the ray tracing-based signal coverage simulation model. The first determining module is used to train and predict the initial neural network model based on the physical environment data to determine the signal strength map of the area to be deployed. The second determining module is used to determine the deployment scheme of the drone base station in the area to be deployed based on the signal strength map.

9. A deployment device for a drone base station, characterized in that, include: Memory, processor; The memory stores computer-executed instructions; The processor executes computer execution instructions stored in the memory, causing the processor to perform the method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions, which, when executed by a processor, are used to implement the method as described in any one of claims 1-7.

11. A computer program product, characterized in that, Includes a computer program that, when executed by a processor, implements the method described in any one of claims 1-7.

Images