A multi-base station networking method, positioning method and device of a UAV

Abstract

This application discloses a multi-base station networking method, positioning method, and apparatus for unmanned aerial vehicles (UAVs), relating to the field of UAV networking and positioning technology. The multi-base station networking method includes: dividing each base station into multiple cooperative sensing groups based on the location of each base station in the area to be networked, and determining multiple sectors of each base station in each cooperative sensing group according to the requirements of airspace layering; acquiring echo signal data of each UAV in the area to be networked based on each cooperative sensing group, determining the first horizontal distance of each UAV, and determining the first horizontal distance and target altitude of the corresponding UAV by combining a fusion network model; calculating the theoretical beam tilt angle of each sector of each base station, and constructing a tilt angle optimization function with the minimum overlap range and blind zone range of the sector coverage as the optimization objective to optimize the theoretical beam tilt angle and obtain the actual execution tilt angle; and constructing a network for each base station and each UAV based on the actual execution tilt angle of each sector of each base station.

Description

Technical Field

[0001] This application relates to the field of unmanned aerial vehicle (UAV) networking technology, and in particular to a multi-base station networking method, positioning method and device for UAVs. Background Technology

[0002] In the field of integrated monitoring of unmanned aerial vehicles (UAVs), with the expansion of low-altitude UAV application scenarios, the demand for precise UAV positioning, full airspace coverage, and compatibility with existing communication networks is becoming increasingly urgent. Therefore, it is necessary to further study UAV monitoring networking technologies.

[0003] Existing drone surveillance network solutions mainly revolve around two directions: single-base station sensing and fixed-beam coverage. Single-base station sensing technology transmits detection signals to the target airspace through a single base station, and calculates the target's position and altitude after receiving the reflected echo from the drone. It does not require the coordination of multiple base stations, thus having a lower deployment cost and is widely used in simple scenarios such as open suburban areas. Fixed-beam coverage technology configures fixed beam tilt angles for base station sectors through preset airspace layering, ensuring that each sector covers the corresponding airspace. The beam parameters are set once during deployment and do not change with the real-time position of the drone. Its operation is simple and does not require complex beam optimization algorithms, making it suitable for scenarios where drone flight paths are fixed and airspace is limited. However, both approaches have their drawbacks: Single-base station sensing relies solely on the echo signal from a single base station, leading to significant calculation errors when the drone is in interference. Furthermore, single-base station coverage is limited, and effective sensing is impossible in base station signal blind spots, resulting in low positioning accuracy and weak anti-interference capabilities, failing to meet actual regulatory needs. On the other hand, fixed-beam coverage cannot dynamically adjust its beam according to the drone's real-time position. When the drone flies at the boundary of airspace layers, multiple base station beam overlap can cause signal interference, or beam coverage gaps can lead to lost sensing. Additionally, because fixed beams require configuration parameters based on maximum coverage requirements, a large amount of idle beam resources are generated in low-density drone airspace, while insufficient beam coverage in high-density drone airspace can lead to regulatory oversights, resulting in poor airspace coverage flexibility and low airspace resource utilization, again failing to meet actual regulatory needs. Therefore, improving positioning accuracy and optimizing airspace resource utilization during drone monitoring to meet actual drone monitoring needs remains a critical technical problem that needs to be solved by existing technologies. Summary of the Invention

[0004] This application provides a multi-base station networking method, positioning method, and apparatus for unmanned aerial vehicles (UAVs) to solve the technical problem that existing UAV networking and positioning methods cannot meet practical needs.

[0005] According to a first aspect of the embodiments of this application, a multi-base station networking method for unmanned aerial vehicles (UAVs) is provided, comprising: Based on the location of each base station in the area to be networked, each base station is divided into multiple cooperative sensing groups, and based on the preset spatial layering requirements, multiple sectors of each base station in each cooperative sensing group are determined. Based on each cooperative sensing group, the echo signal data of each UAV in the area to be networked is acquired. Based on the echo signal data of each UAV, the first horizontal distance of each UAV is determined. Based on a preset fusion network model and combined with the first horizontal distance of each UAV, the target altitude of each UAV is determined. The fusion network model is constructed based on a convolutional neural network and a long short-term memory neural network. Based on the first horizontal distance and target altitude of each UAV, the theoretical beam tilt angle of each sector of each base station is calculated. With the minimum of the overlap range and blind zone range of the sector coverage as the optimization objective, a tilt angle optimization function is constructed to optimize the theoretical beam tilt angle of each sector of each base station in each cooperative sensing group, so as to obtain the actual execution tilt angle. Based on the actual execution tilt angle of each sector of each base station, network construction is carried out for each base station and each UAV in the area to be networked.

[0006] This application first divides each base station in the area to be networked into multiple cooperative sensing groups. Based on the spatial layering requirements, it determines multiple sectors of each base station in each cooperative sensing group. Then, based on the echo signal data of each UAV acquired by each cooperative sensing group, it determines the first horizontal distance of each UAV and the target altitude determined by the fusion network model. This enables cooperative positioning of UAVs through multiple base stations in the cooperative sensing groups, avoiding the problems of small sensing range and weak anti-interference capability caused by existing single-base station positioning, thereby improving the positioning accuracy of UAVs. Simultaneously, through the fusion network model constructed based on convolutional neural networks and long short-term memory neural networks, it can fully analyze the echo signal data of UAVs. Improving the accuracy of the obtained target altitude, i.e., improving the positioning accuracy of drones, is necessary to meet the actual needs of drone surveillance networking. Then, based on the first horizontal distance and target altitude of each drone, the theoretical beam tilt angle of each sector of each base station is calculated. With the goal of minimizing the overlap and blind zone ranges of the sector coverage, a tilt angle optimization function is constructed to optimize the theoretical beam tilt angle of each sector of each base station in each collaborative sensing group, obtaining the actual execution tilt angle. This optimizes the overlapping idle parts and blind zone omissions in the sector coverage determined by the theoretical beam tilt angle, thereby improving the airspace coverage and resource utilization of the base stations to meet the actual needs of drone surveillance networking.

[0007] In some embodiments of this application, the step of dividing each base station into multiple cooperative sensing groups based on the location of each base station in the area to be networked, and determining multiple sectors of each base station in each cooperative sensing group according to preset spatial layering requirements, specifically includes: Based on the location coordinates of each base station in the area to be networked, the actual distance between any two base stations is calculated using the great circle distance formula. Each base station is divided into multiple collaborative sensing groups in a triangular distribution; within each collaborative sensing group, the base stations are adjacent and the actual distance between them meets the preset spacing requirements. Based on the preset spatial layering requirements and the base station antenna height, the sector coverage range and initial tilt angle of each sector of each base station in each cooperative sensing group are obtained, so as to determine multiple sectors of each base station in each cooperative sensing group.

[0008] This application first calculates the actual distance between any two base stations based on the location coordinates of each base station in the area to be networked, using the great circle distance formula. Based on the location and actual distance, the base stations are divided into multiple cooperative sensing groups in a triangular distribution manner. This triangular distribution method can ensure the sensing coverage as much as possible. Then, based on the spatial layering requirements and the base station antenna height, the sector coverage range and initial tilt angle of each sector of each base station in each cooperative sensing group are obtained. This provides an implementation basis for subsequent UAV cooperative positioning through multiple base stations in each cooperative sensing group, thereby meeting the actual needs of UAV monitoring network construction.

[0009] In some embodiments of this application, the step of acquiring echo signal data of each UAV within the area to be networked based on each cooperative sensing group, determining the first horizontal distance of each UAV based on the echo signal data of each UAV, and determining the target altitude of each UAV based on a preset fusion network model and the first horizontal distance of each UAV, specifically includes: Based on the flight trajectories of each UAV in the area to be networked, the corresponding cooperative sensing group is determined, and the echo signal data of the corresponding UAV is obtained based on the cooperative sensing group. Based on the echo signal data of each UAV, and based on the corresponding cooperative sensing group, the initial target altitude and first horizontal distance of each UAV are determined. Based on the fusion network model, the initial target altitude of each UAV is corrected by using the echo signal data of each UAV and the first horizontal distance, thereby determining the target altitude of each UAV.

[0010] This application first determines the corresponding cooperative sensing group based on the flight trajectory of each UAV to obtain the UAV echo signal data. It can perform cooperative positioning of UAVs through multiple base stations in the cooperative sensing group, avoiding the problems of small sensing range and weak anti-interference capability caused by existing single base station positioning, thereby improving the positioning accuracy of UAVs. Then, it determines the initial target altitude and first horizontal distance of each UAV, and corrects the initial target altitude based on the fusion network model to obtain the target altitude of each UAV. It can fully analyze the echo signal data of UAVs based on the fusion network model constructed by convolutional neural network and long short-term memory neural network, thereby improving the accuracy of the obtained target altitude and improving the positioning accuracy of UAVs.

[0011] In some embodiments of this application, the echo signal data includes echo signal data from the main base station and echo signal data from the slave base station; the step of determining the initial target altitude and first horizontal distance of each UAV based on the echo signal data of each UAV and the corresponding cooperative sensing group specifically includes: The echo signal data of each UAV and the corresponding cooperative sensing group's main base station are processed by time-domain filtering and frame synchronization to determine the signal delay of the echo signal data of each UAV's main base station. Based on the signal delay, the initial target altitude of each UAV and the corresponding cooperative sensing group's main base station is determined. Based on the time difference of arrival (TDOA) positioning algorithm, the echo signal data of the main base station and the echo signal data of the slave base station of each UAV are analyzed to determine the first horizontal distance between each UAV and the main base station of the corresponding cooperative sensing group.

[0012] This application first performs time-domain filtering and frame synchronization processing on the echo signal data of the corresponding main base station of each UAV to determine the signal delay, and then determines the initial target altitude of each UAV and the main base station of the corresponding cooperative sensing group. Then, based on the time difference of arrival positioning algorithm, it analyzes the echo signal data of the main base station and the echo signal data of the slave base station of each UAV to determine the first horizontal distance. This can provide a data basis for further determining the target altitude of the UAV based on the initial target altitude and the first horizontal distance, thereby meeting the actual needs of UAV monitoring network.

[0013] In some embodiments of this application, the step of correcting the initial target altitude of each UAV based on the fusion network model using echo signal data and a first horizontal distance, and determining the target altitude of each UAV, specifically includes: In-phase orthogonal sequences are extracted from the echo signal data of each UAV to extract time-domain characteristics and obtain the time-domain features of each UAV. Based on the Fast Fourier Transform, the in-phase orthogonal sequences corresponding to each UAV are converted into Doppler spectrum signals to extract frequency domain characteristics and obtain the frequency domain features of each UAV. Based on the first horizontal distance and initial target altitude of each UAV, a two-dimensional feature vector of the airspace of each UAV is constructed to characterize the airspace characteristics and obtain the airspace characteristics of each UAV. Based on the fusion network model, the initial target altitude of each UAV is corrected and predicted by using the temporal, frequency, and spatial characteristics of each UAV, thereby determining the target altitude of each UAV.

[0014] This application first extracts the in-phase orthogonal sequence from the echo signal data of each UAV to extract its time-domain characteristics, and then extracts its frequency-domain characteristics based on Fast Fourier Transform. Next, it constructs a two-dimensional feature vector in the airspace to represent the airspace characteristics based on the first horizontal distance and the initial target altitude of the UAV. Then, based on the fusion network model, it makes correction predictions by using the time-domain, frequency-domain, and airspace characteristics of each UAV to determine the target altitude of each UAV. The fusion network model, based on convolutional neural networks and long short-term memory neural networks, can fully analyze the echo signal data of the UAV, thereby improving the accuracy of the obtained target altitude and the positioning accuracy of the UAV.

[0015] In some embodiments of this application, the step of calculating the theoretical beam tilt angle of each sector of each base station based on the first horizontal distance and target altitude of each UAV specifically includes: Based on the positional relationship between each UAV and the corresponding main base station of the cooperative sensing group, a solution triangle is constructed. Then, based on the first horizontal distance and target height of each UAV, combined with the base station antenna height, the theoretical beam tilt angle of each sector of each base station is obtained by solving the solution triangle.

[0016] This application first constructs a solution triangle based on the positional relationship between each UAV and the corresponding main base station of the cooperative sensing group. Then, it combines the first horizontal distance of each UAV, the target height, and the distance of the base station antenna to obtain the theoretical beam tilt angle of each sector of each base station. This provides a data basis for subsequent optimization of the theoretical beam tilt angle through the tilt angle optimization function, thereby meeting the actual needs of UAV monitoring network.

[0017] In some embodiments of this application, the step of constructing a tilt angle optimization function with the goal of minimizing the overlap range and blind zone range of sector coverage, in order to optimize the theoretical beam tilt angle of each sector of each base station in each cooperative sensing group, and obtain the actual execution tilt angle, specifically includes: A tilt angle optimization function is constructed with the goal of minimizing the overlap range and blind zone range of the sector coverage area; Based on the signal design and performance design of each base station, the first constraint of the tilt angle optimization function is determined; To solve the tilt angle optimization function, the first constraint is integrated into the tilt angle optimization function using a penalty function method. The theoretical beam tilt angle of each sector of each base station is substituted into the function, and the solution is obtained based on the gradient descent algorithm. This optimizes the theoretical beam tilt angle of each sector of each base station in each cooperative sensing group, and yields the actual execution tilt angle of each sector of each base station.

[0018] This application first constructs a tilt angle optimization function with the goal of minimizing the overlap range and blind zone range of the sector coverage area. Then, it determines the first constraint based on the signal design and performance design of each base station. The first constraint is then integrated into the tilt angle optimization function using the penalty function method, and solved using the gradient descent algorithm to optimize the theoretical beam tilt angle and obtain the actual execution tilt angle. This can optimize the overlapping idle part and the blind zone omission part in the sector coverage area determined by the theoretical beam tilt angle, thereby improving the airspace coverage and resource utilization rate of the base station to meet the actual needs of UAV monitoring network.

[0019] According to a second aspect of the embodiments of this application, a method for locating a drone is provided, comprising: Based on the multi-base station networking method for unmanned aerial vehicles (UAVs) described in this application, feature data of each base station in multiple cooperative sensing groups within a region to be located are acquired, and real-time data acquisition is performed on the multiple cooperative sensing groups to obtain echo signal data of each base station in each cooperative sensing group; wherein, the feature data includes topological features and coordinate features; Based on the topological characteristics of each base station in each collaborative sensing group, the corresponding echo signal data is processed to construct the observation vector of each collaborative sensing group, and based on the coordinate characteristics of each base station in each collaborative sensing group, the sensing matrix of each collaborative sensing group is constructed. Based on the pre-set compressed sensing joint reconstruction algorithm, the spatial point cloud optimization equation corresponding to each collaborative sensing group is constructed and solved through the observation vector and sensing matrix of each collaborative sensing group, so as to obtain the spatial point cloud corresponding to each collaborative sensing group. Based on the airspace point cloud corresponding to each collaborative sensing group, the real-time flight trajectories of several UAVs within each collaborative sensing group are generated to complete the positioning of each UAV within the area to be located.

[0020] This application first acquires feature data of each base station in multiple cooperative sensing groups within the target area using a multi-base station networking method. Real-time data acquisition is then performed on these cooperative sensing groups to obtain echo signal data from each base station within each group. Next, using the feature data of each base station within each cooperative sensing group, the echo signal data is processed to construct the observation vector and perception matrix for each cooperative sensing group. The topological features in the feature data of each base station are used to process the echo signal data to obtain the observation vector of the cooperative sensing group to which each base station belongs. The coordinate features in the feature data of each base station are used to construct the perception matrix of the cooperative sensing group to which each base station belongs. This allows for observation and perception modeling of the spatial domain corresponding to each cooperative sensing group, providing a data foundation for subsequent spatial point cloud solving. Then, based on a compressed sensing joint reconstruction algorithm, the spatial point cloud optimization equation for each cooperative sensing group is constructed and solved to obtain the corresponding spatial point cloud. This generates real-time flight trajectories for several UAVs within each cooperative sensing group to complete UAV positioning. The compressed sensing joint reconstruction algorithm enables accurate solving of the spatial point cloud, resulting in more accurate real-time flight trajectories for UAVs, thereby improving the positioning accuracy of UAVs to meet practical needs.

[0021] According to a third aspect of the embodiments of this application, a multi-base station networking device for unmanned aerial vehicles is provided, including a base station sector division module, a target height calculation module, a tilt angle calculation and optimization module, and a regional networking execution module; The base station sector division module is used to divide each base station into multiple collaborative sensing groups according to the location of each base station in the area to be networked, and to determine multiple sectors of each base station in each collaborative sensing group according to the preset spatial layering requirements. The target height calculation module is used to acquire echo signal data of each UAV in the area to be networked based on each cooperative sensing group, determine the first horizontal distance of each UAV based on the echo signal data of each UAV, and determine the target height of each UAV based on a preset fusion network model and the first horizontal distance of each UAV; wherein, the fusion network model is constructed based on a convolutional neural network and a long short-term memory neural network. The tilt angle calculation and optimization module is used to calculate the theoretical beam tilt angle of each sector of each base station based on the first horizontal distance and target height of each UAV, and construct a tilt angle optimization function with the minimum of the overlap range and blind zone range of the sector coverage as the optimization objective, so as to optimize the theoretical beam tilt angle of each sector of each base station and obtain the actual execution tilt angle. The regional networking execution module is used to construct a network for each base station and each UAV within the area to be networked, based on the actual execution tilt angle of each sector of each base station.

[0022] In some embodiments of this application, the base station sector division module includes a base station distance calculation unit, a base station group division unit, and a base station sector determination unit; The base station distance calculation unit is used to calculate the actual distance between any two base stations based on the great circle distance formula, according to the location coordinates of each base station in the area to be networked. The base station group division unit is used to divide each base station into multiple cooperative sensing groups in a triangular distribution manner; wherein, the base stations in each cooperative sensing group are adjacent and the actual distance between them meets the preset spacing requirements. The base station sector determination unit is used to determine the sector coverage range and initial tilt angle of each sector of each base station in each cooperative sensing group based on the base station antenna height, according to the preset spatial layering requirements, so as to determine multiple sectors of each base station in each cooperative sensing group.

[0023] In some embodiments of this application, the target altitude solving module includes a UAV data acquisition unit, a UAV data measurement unit, and a target altitude determination unit; The UAV data acquisition unit is used to determine the corresponding cooperative sensing group based on the flight trajectory of each UAV in the area to be networked, and to acquire the echo signal data of the corresponding UAV based on the cooperative sensing group. The UAV data measurement unit is used to determine the initial target altitude and first horizontal distance of each UAV based on the echo signal data of each UAV and the corresponding cooperative sensing group. The target altitude determination unit is used to determine the target altitude of each UAV by correcting the initial target altitude of each UAV based on the fusion network model, using the echo signal data of each UAV and the first horizontal distance.

[0024] In some embodiments of this application, the echo signal data includes echo signal data from the main base station and echo signal data from the slave base station; the UAV data measurement unit includes an initial altitude determination subunit and a horizontal distance determination subunit; The initial height determination subunit is used to perform time-domain filtering and frame synchronization processing on the echo signal data of each UAV and the corresponding main base station of the cooperative sensing group, determine the signal delay of the echo signal data of each UAV's main base station, and determine the initial target height of each UAV and the corresponding main base station of the cooperative sensing group based on the signal delay. The horizontal distance determination subunit is used to analyze the echo signal data of the main base station and the echo signal data of the slave base station of each UAV based on the time difference of arrival positioning algorithm, and determine the first horizontal distance between each UAV and the main base station of the corresponding cooperative sensing group.

[0025] In some embodiments of this application, the target altitude determination unit includes a time-domain feature extraction subunit, a frequency-domain feature extraction subunit, a spatial-domain feature extraction subunit, and an altitude correction prediction subunit; The time-domain feature extraction subunit is used to extract in-phase orthogonal sequences from the echo signal data of each UAV in order to extract time-domain features and obtain the time-domain characteristics of each UAV. The frequency domain feature extraction subunit is used to convert the in-phase orthogonal sequences corresponding to each UAV into Doppler spectrum signals based on the fast Fourier transform in order to extract frequency domain features and obtain the frequency domain characteristics of each UAV. The airspace feature extraction subunit is used to construct a two-dimensional airspace feature vector for each UAV based on the first horizontal distance and initial target altitude of each UAV, so as to characterize the airspace features and obtain the airspace features of each UAV. The altitude correction prediction subunit is used to correct and predict the initial target altitude of each UAV based on the fusion network model and through the temporal, frequency, and spatial characteristics of each UAV, thereby determining the target altitude of each UAV.

[0026] In some embodiments of this application, the tilt angle solution optimization module includes a theoretical tilt angle solution unit; the theoretical tilt angle solution unit is used to construct a solution triangle based on the positional relationship between each UAV and the main base station of the corresponding cooperative sensing group, and to solve the solution triangle to obtain the theoretical beam tilt angle of each sector of each base station based on the first horizontal distance and target height of each UAV and the base station antenna height.

[0027] In some embodiments of this application, the tilt angle solution optimization module includes an optimization function construction unit, a first constraint determination unit, and an optimization function solution unit; The optimization function construction unit is used to construct a tilt angle optimization function with the minimum of the overlap range and blind zone range of the sector coverage as the optimization objective. The first constraint determination unit is used to determine the first constraint of the tilt angle optimization function based on the signal design and performance design of each base station; The optimization function solving unit is used to solve the tilt angle optimization function. During the solution process, the first constraint is integrated into the tilt angle optimization function using a penalty function method. The theoretical beam tilt angle of each sector of each base station is substituted into the function, and the solution is performed based on the gradient descent algorithm to optimize the theoretical beam tilt angle of each sector of each base station in each cooperative sensing group, thereby obtaining the actual execution tilt angle of each sector of each base station.

[0028] This application first divides each base station in the area to be networked into multiple cooperative sensing groups. Based on the spatial layering requirements, it determines multiple sectors of each base station in each cooperative sensing group. Then, based on the echo signal data of each UAV acquired by each cooperative sensing group, it determines the first horizontal distance of each UAV and the target altitude determined by the fusion network model. This enables cooperative positioning of UAVs through multiple base stations in the cooperative sensing groups, avoiding the problems of small sensing range and weak anti-interference capability caused by existing single-base station positioning, thereby improving the positioning accuracy of UAVs. Simultaneously, through the fusion network model constructed based on convolutional neural networks and long short-term memory neural networks, it can fully analyze the echo signal data of UAVs. Improving the accuracy of the obtained target altitude, i.e., improving the positioning accuracy of drones, is necessary to meet the actual needs of drone surveillance networking. Then, based on the first horizontal distance and target altitude of each drone, the theoretical beam tilt angle of each sector of each base station is calculated. With the goal of minimizing the overlap and blind zone ranges of the sector coverage, a tilt angle optimization function is constructed to optimize the theoretical beam tilt angle of each sector of each base station in each collaborative sensing group, obtaining the actual execution tilt angle. This optimizes the overlapping idle parts and blind zone omissions in the sector coverage determined by the theoretical beam tilt angle, thereby improving the airspace coverage and resource utilization of the base stations to meet the actual needs of drone surveillance networking.

[0029] According to a fourth aspect of the embodiments of this application, a positioning device for an unmanned aerial vehicle (UAV) is provided, including a base station data acquisition module, a data observation and perception processing module, an airspace point cloud optimization solution module, and a UAV real-time positioning module. The base station data acquisition module is used to acquire feature data of each base station of multiple cooperative sensing groups in the area to be located, based on the multi-base station networking method of UAV described in this application, and to perform real-time data acquisition on the multiple cooperative sensing groups to obtain echo signal data of each base station in each cooperative sensing group; wherein, the feature data includes topological features and coordinate features. The data observation and sensing processing module is used to process the corresponding echo signal data according to the topological characteristics of each base station in each collaborative sensing group, construct the observation vector of each collaborative sensing group, and construct the sensing matrix of each collaborative sensing group according to the coordinate characteristics of each base station in each collaborative sensing group. The spatial point cloud optimization solution module is used to construct and solve the spatial point cloud optimization equation corresponding to each collaborative sensing group based on the preset compressed sensing joint reconstruction algorithm, through the observation vector and sensing matrix of each collaborative sensing group, to obtain the spatial point cloud corresponding to each collaborative sensing group. The real-time UAV positioning module is used to generate real-time flight trajectories of several UAVs within each collaborative sensing group based on the airspace point cloud corresponding to each collaborative sensing group, so as to complete the positioning of each UAV in the area to be located.

[0030] This application first acquires feature data of each base station in multiple cooperative sensing groups within the target area using a multi-base station networking method. Real-time data acquisition is then performed on these cooperative sensing groups to obtain echo signal data from each base station within each group. Next, using the feature data of each base station within each cooperative sensing group, the echo signal data is processed to construct the observation vector and perception matrix for each cooperative sensing group. The topological features in the feature data of each base station are used to process the echo signal data to obtain the observation vector of the cooperative sensing group to which each base station belongs. The coordinate features in the feature data of each base station are used to construct the perception matrix of the cooperative sensing group to which each base station belongs. This allows for observation and perception modeling of the spatial domain corresponding to each cooperative sensing group, providing a data foundation for subsequent spatial point cloud solving. Then, based on a compressed sensing joint reconstruction algorithm, the spatial point cloud optimization equation for each cooperative sensing group is constructed and solved to obtain the corresponding spatial point cloud. This generates real-time flight trajectories for several UAVs within each cooperative sensing group to complete UAV positioning. The compressed sensing joint reconstruction algorithm enables accurate solving of the spatial point cloud, resulting in more accurate real-time flight trajectories for UAVs, thereby improving the positioning accuracy of UAVs to meet practical needs. Attached Figure Description

[0031] Figure 1 This is a flowchart illustrating a multi-base station networking method for unmanned aerial vehicles (UAVs) according to certain embodiments of this application. Figure 2 This is a flowchart illustrating a method for locating a drone according to certain embodiments of this application. Figure 3 This is a modular structure diagram of a multi-base station networking device for an unmanned aerial vehicle (UAV) according to certain embodiments of this application. Figure 4 This is a modular structure diagram of a positioning device for a drone shown in some embodiments of this application. Detailed Implementation

[0032] The embodiments of this application are described in detail below. Examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals denote the same or similar elements or elements having the same or similar functions throughout. The embodiments described below in conjunction with the accompanying drawings are exemplary and are only used to explain some embodiments of this application, and should not be construed as limiting the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments shown in this application without inventive effort are within the protection scope of this application.

[0033] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be construed as indicating or implying relative importance or implicitly specifying the number of indicated technical features. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of that feature. In the description of this application, unless otherwise explicitly specified, "a plurality of" or "several" means two or more.

[0034] In existing drone surveillance network solutions, single-base station sensing technology relies solely on the echo signal of a single base station. When drones are exposed to interference, the calculation deviation increases significantly. Furthermore, the coverage of a single base station is limited, and it cannot effectively detect drones in signal blind spots, resulting in low positioning accuracy and weak anti-interference capabilities, failing to meet actual surveillance needs. Fixed-beam coverage technology cannot dynamically adjust with the drone's real-time location. When drones fly at airspace layer boundaries, multiple base station beam overlap can cause signal interference, or beam coverage gaps can lead to lost sensing. Moreover, because fixed beams must be configured according to maximum coverage requirements, a large amount of idle beam resources are generated in low-density drone airspace, while insufficient beam coverage in high-density drone airspace leads to surveillance omissions. This results in poor airspace coverage flexibility and low airspace resource utilization, also failing to meet actual surveillance needs. Therefore, improving positioning accuracy and optimizing airspace resource utilization during drone surveillance to meet actual surveillance requirements remains a pressing technical problem that needs to be solved.

[0035] Based on the above technical background, please refer to Figure 1 This application provides a method for multi-base station networking of unmanned aerial vehicles (UAVs), including steps S101 to S104, the specific steps of which are as follows: Step S101: Based on the location of each base station in the area to be networked, divide each base station into multiple cooperative sensing groups, and determine multiple sectors of each base station in each cooperative sensing group according to the preset spatial layering requirements.

[0036] In some embodiments of this application, the step of dividing each base station into multiple cooperative sensing groups based on the location of each base station in the area to be networked, and determining multiple sectors of each base station in each cooperative sensing group according to preset spatial layering requirements, specifically includes: Based on the location coordinates of each base station in the area to be networked, the actual distance between any two base stations is calculated using the great circle distance formula. Each base station is divided into multiple collaborative sensing groups in a triangular distribution; within each collaborative sensing group, the base stations are adjacent and the actual distance between them meets the preset spacing requirements. Based on the preset spatial layering requirements and the base station antenna height, the sector coverage range and initial tilt angle of each sector of each base station in each cooperative sensing group are obtained, so as to determine multiple sectors of each base station in each cooperative sensing group.

[0037] Specifically, the preferred embodiment of the great circle distance formula is the Haversine formula; the preferred value of the preset spacing requirement is 1 km.

[0038] The reason for considering dividing the base stations using a triangular distribution is that the triangular distribution of the cooperative sensing groups can maximize the coverage of the airspace and reduce signal overlap and interference between the cooperative sensing groups.

[0039] Specifically, after dividing into cooperative sensing groups, timestamp synchronization needs to be performed within each cooperative sensing group. Specifically, within the same cooperative sensing group, any base station is used as the reference base station, and signals with timestamps are sent to the other two base stations through the reference base station to calculate the synchronization distance error. The clock phase is adjusted based on the synchronization distance error to ensure that the synchronization distance error within the same cooperative sensing group does not exceed the preset synchronization distance threshold. The preferred value of the synchronization distance threshold is 0.3m.

[0040] Specifically, the number of sectors for each base station is generally limited to three, and the three sectors of the same base station are responsible for airspace at different altitudes (i.e., sector coverage ranges), such as 0-300m for low altitude, 300-600m for mid-altitude, and 600-1200m for high altitude. After the sector coverage range of a certain sector is determined, the initial tilt angle of that sector is determined based on the base station antenna height and the sector coverage range, specifically: ; in, This is the initial tilt angle; This represents the lower limit height of the sector coverage area; This refers to the height of the base station antenna. For reference horizontal distance, the preferred value is 500m.

[0041] This application first calculates the actual distance between any two base stations based on the location coordinates of each base station in the area to be networked, using the great circle distance formula. Based on the location and actual distance, the base stations are divided into multiple cooperative sensing groups in a triangular distribution manner. This triangular distribution method can ensure the sensing coverage as much as possible. Then, based on the spatial layering requirements and the base station antenna height, the sector coverage range and initial tilt angle of each sector of each base station in each cooperative sensing group are obtained. This provides an implementation basis for subsequent UAV cooperative positioning through multiple base stations in each cooperative sensing group, thereby meeting the actual needs of UAV monitoring network construction.

[0042] Step S102: Based on each cooperative sensing group, acquire the echo signal data of each UAV in the area to be networked, determine the first horizontal distance of each UAV based on the echo signal data of each UAV, and determine the target altitude of each UAV based on a preset fusion network model and the first horizontal distance of each UAV; wherein, the fusion network model is constructed based on a convolutional neural network and a long short-term memory neural network.

[0043] In some embodiments of this application, the step of acquiring echo signal data of each UAV within the area to be networked based on each cooperative sensing group, determining the first horizontal distance of each UAV based on the echo signal data of each UAV, and determining the target altitude of each UAV based on a preset fusion network model and the first horizontal distance of each UAV, specifically includes: Based on the flight trajectories of each UAV in the area to be networked, the corresponding cooperative sensing group is determined, and the echo signal data of the corresponding UAV is obtained based on the cooperative sensing group. Based on the echo signal data of each UAV, and based on the corresponding cooperative sensing group, the initial target altitude and first horizontal distance of each UAV are determined. Based on the fusion network model, the initial target altitude of each UAV is corrected by using the echo signal data of each UAV and the first horizontal distance, thereby determining the target altitude of each UAV.

[0044] This application first determines the corresponding cooperative sensing group based on the flight trajectory of each UAV to obtain the UAV echo signal data. It can perform cooperative positioning of UAVs through multiple base stations in the cooperative sensing group, avoiding the problems of small sensing range and weak anti-interference capability caused by existing single base station positioning, thereby improving the positioning accuracy of UAVs. Then, it determines the initial target altitude and first horizontal distance of each UAV, and corrects the initial target altitude based on the fusion network model to obtain the target altitude of each UAV. It can fully analyze the echo signal data of UAVs based on the fusion network model constructed by convolutional neural network and long short-term memory neural network, thereby improving the accuracy of the obtained target altitude and improving the positioning accuracy of UAVs.

[0045] In some embodiments of this application, the echo signal data includes echo signal data from the main base station and echo signal data from the slave base station; the step of determining the initial target altitude and first horizontal distance of each UAV based on the echo signal data of each UAV and the corresponding cooperative sensing group specifically includes: The echo signal data of each UAV and the corresponding cooperative sensing group's main base station are processed by time-domain filtering and frame synchronization to determine the signal delay of the echo signal data of each UAV's main base station. Based on the signal delay, the initial target altitude of each UAV and the corresponding cooperative sensing group's main base station is determined. Based on the time difference of arrival (TDOA) positioning algorithm, the echo signal data of the main base station and the echo signal data of the slave base station of each UAV are analyzed to determine the first horizontal distance between each UAV and the main base station of the corresponding cooperative sensing group.

[0046] Specifically, for the same cooperative sensing group, including one master base station and two slave base stations, the preferred master base station is the one that provides the highest coverage of the triangular distribution area formed by the three base stations in the cooperative sensing group and has the best communication link stability. It is easy to understand that the determination of the master base station is irrelevant to the UAV. Therefore, when acquiring the UAV's echo signal data, the GNSS data actively reported by the UAV can be obtained first based on the communication link of the master base station as the UAV's echo signal data. If no UAV actively reports, the master base station can transmit RSS reference signals directionally and periodically to the target airspace (i.e., the triangular distribution area), and the UAV's echo signal data can be obtained by receiving and capturing the reflected echo signals from the UAV.

[0047] Specifically, when performing time-domain filtering on the echo signal data from the main base station, bandpass filtering is preferred to remove background noise outside the 3.5GHz band and interference signals from other devices in the same frequency band. After filtering, the start time of the echo signal data is located using a frame synchronization algorithm, thereby calculating the total delay of a single communication between the main base station and the UAV. (i.e., the total time delay from signal transmission to echo reception) ), and calculate the initial target height. ( (Speed ​​of light). The initial target altitude signal is assumed to propagate vertically and is affected by the horizontal offset of the drone and multipath reflections such as those from buildings or trees, resulting in an error of 10-20m.

[0048] Specifically, after the main base station acquires the echo signal data of a certain drone, two slave base stations in the same cooperative sensing group will initiate signal acquisition of the drone to obtain the corresponding echo signal data of the slave base station; subsequently, based on the time difference between the two slave base stations and the main base station in receiving the echo, Based on the preset distance requirement (1km) between the main base station and the slave base station, the first horizontal distance between the main base station and the UAV is derived using a time difference of arrival (TDOA) positioning algorithm. .

[0049] This application first performs time-domain filtering and frame synchronization processing on the echo signal data of the corresponding main base station of each UAV to determine the signal delay, and then determines the initial target altitude of each UAV and the main base station of the corresponding cooperative sensing group. Then, based on the time difference of arrival positioning algorithm, it analyzes the echo signal data of the main base station and the echo signal data of the slave base station of each UAV to determine the first horizontal distance. This can provide a data basis for further determining the target altitude of the UAV based on the initial target altitude and the first horizontal distance, thereby meeting the actual needs of UAV monitoring network.

[0050] In some embodiments of this application, the step of correcting the initial target altitude of each UAV based on the fusion network model using echo signal data and a first horizontal distance, and determining the target altitude of each UAV, specifically includes: In-phase orthogonal sequences are extracted from the echo signal data of each UAV to extract time-domain characteristics and obtain the time-domain features of each UAV. Based on the Fast Fourier Transform, the in-phase orthogonal sequences corresponding to each UAV are converted into Doppler spectrum signals to extract frequency domain characteristics and obtain the frequency domain features of each UAV. Based on the first horizontal distance and initial target altitude of each UAV, a two-dimensional feature vector of the airspace of each UAV is constructed to characterize the airspace characteristics and obtain the airspace characteristics of each UAV. Based on the fusion network model, the initial target altitude of each UAV is corrected and predicted by using the temporal, frequency, and spatial characteristics of each UAV, thereby determining the target altitude of each UAV.

[0051] Specifically, when correcting and predicting the initial target altitude of a UAV using a fusion network model, the UAV's temporal, frequency, and spatial features are first input into the model sequentially. These features are then extracted and integrated through a convolutional neural network module and a long short-term memory neural network module. Finally, a fully connected layer fuses the extracted features to output the UAV's target altitude. Among them, the convolutional neural network module extracts local features through 3 layers of convolutional kernels, and the long short-term memory neural network module captures dynamic changes through 2 layers of temporal units.

[0052] This application first extracts the in-phase orthogonal sequence from the echo signal data of each UAV to extract its time-domain characteristics, and then extracts its frequency-domain characteristics based on Fast Fourier Transform. Next, it constructs a two-dimensional feature vector in the airspace to represent the airspace characteristics based on the first horizontal distance and the initial target altitude of the UAV. Then, based on the fusion network model, it makes correction predictions by using the time-domain, frequency-domain, and airspace characteristics of each UAV to determine the target altitude of each UAV. The fusion network model, based on convolutional neural networks and long short-term memory neural networks, can fully analyze the echo signal data of the UAV, thereby improving the accuracy of the obtained target altitude and the positioning accuracy of the UAV.

[0053] Step S103: Based on the first horizontal distance and target altitude of each UAV, calculate the theoretical beam tilt angle of each sector of each base station, and construct a tilt angle optimization function with the minimum overlap range and blind zone range of the sector coverage as the optimization objective, so as to optimize the theoretical beam tilt angle of each sector of each base station in each cooperative sensing group and obtain the actual execution tilt angle.

[0054] In some embodiments of this application, the step of calculating the theoretical beam tilt angle of each sector of each base station based on the first horizontal distance and target altitude of each UAV specifically includes: Based on the positional relationship between each UAV and the corresponding main base station of the cooperative sensing group, a solution triangle is constructed. Then, based on the first horizontal distance and target height of each UAV, combined with the base station antenna height, the theoretical beam tilt angle of each sector of each base station is obtained by solving the solution triangle.

[0055] Specifically, for each UAV, a right-angled triangle is constructed with the base station antenna of the corresponding main base station as the origin and the UAV's position as the target point. The horizontal side of this right-angled triangle is the first horizontal distance. The vertical side represents the target height of the drone. Base station antenna height corresponding to the main base station The difference is the theoretical beam tilt angle. .

[0056] This application first constructs a solution triangle based on the positional relationship between each UAV and the corresponding main base station of the cooperative sensing group. Then, it combines the first horizontal distance of each UAV, the target height, and the distance of the base station antenna to obtain the theoretical beam tilt angle of each sector of each base station. This provides a data basis for subsequent optimization of the theoretical beam tilt angle through the tilt angle optimization function, thereby meeting the actual needs of UAV monitoring network.

[0057] In some embodiments of this application, the step of constructing a tilt angle optimization function with the goal of minimizing the overlap range and blind zone range of sector coverage, in order to optimize the theoretical beam tilt angle of each sector of each base station in each cooperative sensing group, and obtain the actual execution tilt angle, specifically includes: A tilt angle optimization function is constructed with the goal of minimizing the overlap range and blind zone range of the sector coverage area; Based on the signal design and performance design of each base station, the first constraint of the tilt angle optimization function is determined; To solve the tilt angle optimization function, the first constraint is integrated into the tilt angle optimization function using a penalty function method. The theoretical beam tilt angle of each sector of each base station is substituted into the function, and the solution is obtained based on the gradient descent algorithm. This optimizes the theoretical beam tilt angle of each sector of each base station in each cooperative sensing group, and yields the actual execution tilt angle of each sector of each base station.

[0058] Specifically, the tilt angle optimization function is as follows: ; in, These are the actual tilt angles of the corresponding sectors of each base station in the collaborative sensing group; The data is obtained by calculating the sector coverage range and actual tilt angle of each base station in the corresponding sector within the collaborative sensing group. This is the weighting coefficient, with a preferred value of 0.8; It is calculated by the difference between the total regulatory airspace and the sector coverage of each base station in the collaborative sensing group across all sectors.

[0059] Specifically, the first constraint is: (1) single sector beamwidth ≥ 10°; (2) tilt adjustment rate ≤ 5° / ms, which is used to characterize the mechanical performance limitations of base station beam control.

[0060] Specifically, when integrating the first constraint into the tilt angle optimization function using the penalty function method, it is necessary to... Transformed into ; This refers to the first constraint, characterized by a penalty term. More specifically, ; That is, the single-sector beamwidth constraint characterized by a penalty term. That is, the tilt adjustment rate constraint represented by the penalty term; This represents the weight of the penalty term. Generally, All are large positive numbers, used to increase the significance of the error of the corresponding tilt angle optimization scheme when the penalty term is violated, so that the corresponding tilt angle optimization scheme is eliminated in the optimization, thereby ensuring that the final solution satisfies the two hardware physical constraints in the first constraint.

[0061] Specifically, when solving the problem using the gradient descent algorithm, the theoretical beam tilt angle is iteratively calculated. In each iteration, the calculated tilt angle for the current round is output, and a first check is performed. If the calculated tilt angle meets the first check condition, it is used as the actual calculated tilt angle; otherwise, the next iteration begins, and the weight coefficients in the tilt angle optimization function are adjusted in the next iteration. Or penalty coefficient The calculation continues until the solution tilt angle that satisfies the first check is obtained, and the corresponding solution tilt angle is taken as the actual execution tilt angle. More specifically, the first check is simulated by a beam coverage range simulation model. The conditions for the first check include: (1) whether the target height of the UAV is covered by the beams of at least 2 base stations in the cooperative sensing group; (2) whether the "tower black" area (i.e., the area within 100m directly above the top of the base station) of each base station in the cooperative sensing group is covered by the beams of at least 1 base station.

[0062] This application first constructs a tilt angle optimization function with the goal of minimizing the overlap range and blind zone range of the sector coverage area. Then, it determines the first constraint based on the signal design and performance design of each base station. The first constraint is then integrated into the tilt angle optimization function using the penalty function method, and solved using the gradient descent algorithm to optimize the theoretical beam tilt angle and obtain the actual execution tilt angle. This can optimize the overlapping idle part and the blind zone omission part in the sector coverage area determined by the theoretical beam tilt angle, thereby improving the airspace coverage and resource utilization rate of the base station to meet the actual needs of UAV monitoring network.

[0063] Step S104: Based on the actual execution tilt angle of each sector of each base station, construct a network for each base station and each UAV within the area to be networked.

[0064] This application first divides each base station in the area to be networked into multiple cooperative sensing groups. Based on the spatial layering requirements, it determines multiple sectors of each base station in each cooperative sensing group. Then, based on the echo signal data of each UAV acquired by each cooperative sensing group, it determines the first horizontal distance of each UAV and the target altitude determined by the fusion network model. This enables cooperative positioning of UAVs through multiple base stations in the cooperative sensing groups, avoiding the problems of small sensing range and weak anti-interference capability caused by existing single-base station positioning, thereby improving the positioning accuracy of UAVs. Simultaneously, through the fusion network model constructed based on convolutional neural networks and long short-term memory neural networks, it can fully analyze the echo signal data of UAVs. Improving the accuracy of the obtained target altitude, i.e., improving the positioning accuracy of drones, is necessary to meet the actual needs of drone surveillance networking. Then, based on the first horizontal distance and target altitude of each drone, the theoretical beam tilt angle of each sector of each base station is calculated. With the goal of minimizing the overlap and blind zone ranges of the sector coverage, a tilt angle optimization function is constructed to optimize the theoretical beam tilt angle of each sector of each base station in each collaborative sensing group, obtaining the actual execution tilt angle. This optimizes the overlapping idle parts and blind zone omissions in the sector coverage determined by the theoretical beam tilt angle, thereby improving the airspace coverage and resource utilization of the base stations to meet the actual needs of drone surveillance networking.

[0065] After completing the network construction based on the multi-base station networking method for UAVs provided in this application, this application also provides a positioning method for UAVs, including steps S201 to S204, the specific steps of which are as follows: Step S201: Based on the multi-base station networking method for UAVs according to any one of the claims of this application, acquire the feature data of each base station of multiple cooperative sensing groups in the area to be located, and perform real-time data acquisition on the multiple cooperative sensing groups to obtain the echo signal data of each base station in each cooperative sensing group; wherein, the feature data includes topological features and coordinate features.

[0066] Specifically, when collecting real-time data in each collaborative sensing group, it is necessary to align the timestamps of the echo signal data from different base stations in the same collaborative sensing group.

[0067] Step S202: Based on the topological characteristics of each base station in each cooperative sensing group, process the corresponding echo signal data, construct the observation vector of each cooperative sensing group, and construct the sensing matrix of each cooperative sensing group based on the coordinate characteristics of each base station in each cooperative sensing group.

[0068] Specifically, for the same collaborative sensing group, the echo signal data of the three base stations are first constructed into an observation vector based on the topological characteristics of the three base stations. Then, based on the coordinate characteristics of these three base stations, specifically latitude and longitude coordinates, an observation and sensing matrix is ​​constructed. This is used to reflect the mapping relationship between the echo signal and the spatial location.

[0069] Step S203: Based on the preset compressed sensing joint reconstruction algorithm, construct and solve the spatial point cloud optimization equation corresponding to each collaborative sensing group through the observation vector and sensing matrix of each collaborative sensing group, and obtain the spatial point cloud corresponding to each collaborative sensing group.

[0070] Specifically, using the compressed sensing joint reconstruction algorithm, the spatial point cloud optimization equation for any cooperative sensing group can be expressed as: ; in, Represents the L2 norm; Denotes the norm 1; This represents the regularization coefficient, with a preferred value of 0.01. The problem involves solving for the spatial point cloud. As will be readily understood by those skilled in the art, the L2 norm term ensures consistency between the reconstructed result and the echo signal data in the spatial point cloud optimization equation, while the regularization term leverages the sparsity of the UAV as a single effective target to improve accuracy, thereby guaranteeing the accuracy of the spatial point cloud solution.

[0071] Step S204: Based on the airspace point cloud corresponding to each cooperative sensing group, generate the real-time flight trajectories of several UAVs within each cooperative sensing group to complete the positioning of each UAV in the area to be located.

[0072] Specifically, after obtaining the airspace point cloud of a certain collaborative sensing group, the latitude, longitude, and altitude information of the UAV are extracted from the airspace point cloud, and the speed is calculated in combination with the time series to generate the corresponding real-time flight trajectory, thereby completing the positioning of the corresponding UAV.

[0073] Compared to existing technologies, this application first acquires feature data of each base station in multiple cooperative sensing groups within the area to be located using a multi-base station networking method. Then, it performs real-time data acquisition on multiple cooperative sensing groups to obtain echo signal data from each base station within each group. Next, using the feature data of each base station within each cooperative sensing group, it processes the echo signal data to construct the observation vector for each cooperative sensing group and builds the sensing matrix for each group. Finally, it processes the echo signal data using the topological features in the feature data of each base station to obtain the observation vector for each cooperative sensing group, and constructs the sensing matrix for each group using the coordinate features in the feature data of each base station. The perception matrix of the cooperative perception group to which the base station belongs can perform observation modeling and perception modeling of the airspace corresponding to each cooperative perception group, providing a data foundation for subsequent solution of airspace point cloud. Then, based on the compressed sensing joint reconstruction algorithm, the airspace point cloud optimization equation of each cooperative perception group is constructed and solved to obtain the corresponding airspace point cloud, thereby generating the real-time flight trajectory of several UAVs in each cooperative perception group to complete UAV positioning. The compressed sensing joint reconstruction algorithm can accurately solve the airspace point cloud, and thus generate more accurate real-time flight trajectory of UAVs, thereby improving the positioning accuracy of UAVs to meet actual needs.

[0074] For a method corresponding to the one described above, please refer to [link / reference]. Figure 3 This application provides a multi-base station networking device for unmanned aerial vehicles, including a base station sector division module 310, a target height calculation module 320, an tilt angle calculation and optimization module 330, and a regional networking execution module 340. The base station sector division module 310 is used to divide each base station into multiple cooperative sensing groups according to the location of each base station in the area to be networked, and to determine multiple sectors of each base station in each cooperative sensing group according to the preset spatial layering requirements. The target height calculation module 320 is used to acquire echo signal data of each UAV in the area to be networked based on each cooperative sensing group, determine the first horizontal distance of each UAV based on the echo signal data of each UAV, and determine the target height of each UAV based on a preset fusion network model and the first horizontal distance of each UAV; wherein, the fusion network model is constructed based on a convolutional neural network and a long short-term memory neural network. The tilt angle calculation and optimization module 330 is used to calculate the theoretical beam tilt angle of each sector of each base station based on the first horizontal distance and target height of each UAV, and construct a tilt angle optimization function with the minimum of the overlap range and blind zone range of the sector coverage as the optimization objective, so as to optimize the theoretical beam tilt angle of each sector of each base station and obtain the actual execution tilt angle. The regional networking execution module 340 is used to construct a network for each base station and each UAV within the area to be networked, based on the actual execution tilt angle of each sector of each base station.

[0075] In some embodiments of this application, the base station sector division module 310 includes a base station distance calculation unit, a base station group division unit, and a base station sector determination unit; The base station distance calculation unit is used to calculate the actual distance between any two base stations based on the great circle distance formula, according to the location coordinates of each base station in the area to be networked. The base station group division unit is used to divide each base station into multiple cooperative sensing groups in a triangular distribution manner; wherein, the base stations in each cooperative sensing group are adjacent and the actual distance between them meets the preset spacing requirements. The base station sector determination unit is used to determine the sector coverage range and initial tilt angle of each sector of each base station in each cooperative sensing group based on the base station antenna height, according to the preset spatial layering requirements, so as to determine multiple sectors of each base station in each cooperative sensing group.

[0076] In some embodiments of this application, the target altitude solving module 320 includes a UAV data acquisition unit, a UAV data measurement unit, and a target altitude determination unit; The UAV data acquisition unit is used to determine the corresponding cooperative sensing group based on the flight trajectory of each UAV in the area to be networked, and to acquire the echo signal data of the corresponding UAV based on the cooperative sensing group. The UAV data measurement unit is used to determine the initial target altitude and first horizontal distance of each UAV based on the echo signal data of each UAV and the corresponding cooperative sensing group. The target altitude determination unit is used to determine the target altitude of each UAV by correcting the initial target altitude of each UAV based on the fusion network model, using the echo signal data of each UAV and the first horizontal distance.

[0077] In some embodiments of this application, the echo signal data includes echo signal data from the main base station and echo signal data from the slave base station; the UAV data measurement unit includes an initial altitude determination subunit and a horizontal distance determination subunit; The initial height determination subunit is used to perform time-domain filtering and frame synchronization processing on the echo signal data of each UAV and the corresponding main base station of the cooperative sensing group, determine the signal delay of the echo signal data of each UAV's main base station, and determine the initial target height of each UAV and the corresponding main base station of the cooperative sensing group based on the signal delay. The horizontal distance determination subunit is used to analyze the echo signal data of the main base station and the echo signal data of the slave base station of each UAV based on the time difference of arrival positioning algorithm, and determine the first horizontal distance between each UAV and the main base station of the corresponding cooperative sensing group.

[0078] In some embodiments of this application, the target altitude determination unit includes a time-domain feature extraction subunit, a frequency-domain feature extraction subunit, a spatial-domain feature extraction subunit, and an altitude correction prediction subunit; The time-domain feature extraction subunit is used to extract in-phase orthogonal sequences from the echo signal data of each UAV in order to extract time-domain features and obtain the time-domain characteristics of each UAV. The frequency domain feature extraction subunit is used to convert the in-phase orthogonal sequences corresponding to each UAV into Doppler spectrum signals based on the fast Fourier transform in order to extract frequency domain features and obtain the frequency domain characteristics of each UAV. The airspace feature extraction subunit is used to construct a two-dimensional airspace feature vector for each UAV based on the first horizontal distance and initial target altitude of each UAV, so as to characterize the airspace features and obtain the airspace features of each UAV. The altitude correction prediction subunit is used to correct and predict the initial target altitude of each UAV based on the fusion network model and through the temporal, frequency, and spatial characteristics of each UAV, thereby determining the target altitude of each UAV.

[0079] In some embodiments of this application, the tilt angle calculation and optimization module 330 includes a theoretical tilt angle calculation unit; the theoretical tilt angle calculation unit is used to construct a solution triangle based on the positional relationship between each UAV and the main base station of the corresponding cooperative sensing group, and to solve the solution triangle to obtain the theoretical beam tilt angle of each sector of each base station based on the first horizontal distance and target height of each UAV and the base station antenna height.

[0080] In some embodiments of this application, the tilt angle solving optimization module 330 includes an optimization function construction unit, a first constraint determination unit, and an optimization function solving unit; The optimization function construction unit is used to construct a tilt angle optimization function with the minimum of the overlap range and blind zone range of the sector coverage as the optimization objective. The first constraint determination unit is used to determine the first constraint of the tilt angle optimization function based on the signal design and performance design of each base station; The optimization function solving unit is used to solve the tilt angle optimization function. During the solution process, the first constraint is integrated into the tilt angle optimization function using a penalty function method. The theoretical beam tilt angle of each sector of each base station is substituted into the function, and the solution is performed based on the gradient descent algorithm to optimize the theoretical beam tilt angle of each sector of each base station in each cooperative sensing group, thereby obtaining the actual execution tilt angle of each sector of each base station.

[0081] This application first divides each base station in the area to be networked into multiple cooperative sensing groups. Based on the spatial layering requirements, it determines multiple sectors of each base station in each cooperative sensing group. Then, based on the echo signal data of each UAV acquired by each cooperative sensing group, it determines the first horizontal distance of each UAV and the target altitude determined by the fusion network model. This enables cooperative positioning of UAVs through multiple base stations in the cooperative sensing groups, avoiding the problems of small sensing range and weak anti-interference capability caused by existing single-base station positioning, thereby improving the positioning accuracy of UAVs. Simultaneously, through the fusion network model constructed based on convolutional neural networks and long short-term memory neural networks, it can fully analyze the echo signal data of UAVs. Improving the accuracy of the obtained target altitude, i.e., improving the positioning accuracy of drones, is necessary to meet the actual needs of drone surveillance networking. Then, based on the first horizontal distance and target altitude of each drone, the theoretical beam tilt angle of each sector of each base station is calculated. With the goal of minimizing the overlap and blind zone ranges of the sector coverage, a tilt angle optimization function is constructed to optimize the theoretical beam tilt angle of each sector of each base station in each collaborative sensing group, obtaining the actual execution tilt angle. This optimizes the overlapping idle parts and blind zone omissions in the sector coverage determined by the theoretical beam tilt angle, thereby improving the airspace coverage and resource utilization of the base stations to meet the actual needs of drone surveillance networking.

[0082] For a method corresponding to the one described above, please refer to [link / reference]. Figure 4 The present application provides a positioning device for a drone, including a base station data acquisition module 410, a data observation and perception processing module 420, an airspace point cloud optimization solution module 430, and a drone real-time positioning module 440. The base station data acquisition module 410 is used to acquire feature data of each base station of multiple cooperative sensing groups in the area to be located based on the multi-base station networking method of UAV described in this application, and to perform real-time data acquisition on the multiple cooperative sensing groups to obtain echo signal data of each base station in each cooperative sensing group; wherein, the feature data includes topological features and coordinate features. The data observation and sensing processing module 420 is used to process the corresponding echo signal data according to the topological characteristics of each base station in each collaborative sensing group, construct the observation vector of each collaborative sensing group, and construct the sensing matrix of each collaborative sensing group according to the coordinate characteristics of each base station in each collaborative sensing group. The spatial point cloud optimization solution module 430 is used to construct and solve the spatial point cloud optimization equation corresponding to each collaborative sensing group based on the preset compressed sensing joint reconstruction algorithm, through the observation vector and sensing matrix of each collaborative sensing group, to obtain the spatial point cloud corresponding to each collaborative sensing group. The UAV real-time positioning module 440 is used to generate real-time flight trajectories of several UAVs in each collaborative sensing group based on the airspace point cloud corresponding to each collaborative sensing group, so as to complete the positioning of each UAV in the area to be positioned.

[0083] This application first acquires feature data of each base station in multiple cooperative sensing groups within the target area using a multi-base station networking method. Real-time data acquisition is then performed on these cooperative sensing groups to obtain echo signal data from each base station within each group. Next, using the feature data of each base station within each cooperative sensing group, the echo signal data is processed to construct the observation vector and perception matrix for each cooperative sensing group. The topological features in the feature data of each base station are used to process the echo signal data to obtain the observation vector of the cooperative sensing group to which each base station belongs. The coordinate features in the feature data of each base station are used to construct the perception matrix of the cooperative sensing group to which each base station belongs. This allows for observation and perception modeling of the spatial domain corresponding to each cooperative sensing group, providing a data foundation for subsequent spatial point cloud solving. Then, based on a compressed sensing joint reconstruction algorithm, the spatial point cloud optimization equation for each cooperative sensing group is constructed and solved to obtain the corresponding spatial point cloud. This generates real-time flight trajectories for several UAVs within each cooperative sensing group to complete UAV positioning. The compressed sensing joint reconstruction algorithm enables accurate solving of the spatial point cloud, resulting in more accurate real-time flight trajectories for UAVs, thereby improving the positioning accuracy of UAVs to meet practical needs.

[0084] It should be understood that the apparatus provided in the embodiments of this application corresponds to the aforementioned method. The multi-base station networking apparatus or the positioning apparatus for a UAV provided in the embodiments of this application can correspondingly implement the multi-base station networking method or the positioning method for a UAV provided in any embodiment of this application.

[0085] Adaptively, embodiments of this application also provide a computer device and a computer-readable storage medium.

[0086] The computer device includes: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor; When the processor executes the computer program, it implements a multi-base station networking method for unmanned aerial vehicles (UAVs) or a positioning method for UAVs according to this application.

[0087] The computer-readable storage medium stores multiple instructions, which are adapted for a processor to load and execute a multi-base station networking method for a drone or a positioning method for a drone according to this application.

[0088] The above description represents some embodiments of this application, providing a further detailed explanation of the purpose, technical solution, and beneficial effects of this application. It should be understood that the above-described embodiments of this application should not be construed as limiting this application. In particular, any changes, modifications, equivalent substitutions, and variations made by those skilled in the art within the spirit and principles of this application should be included within the scope of protection of this application.

Claims

1. A method for multi-base station networking of unmanned aerial vehicles (UAVs), characterized in that, include: Based on the location of each base station in the area to be networked, each base station is divided into multiple cooperative sensing groups, and based on the preset spatial layering requirements, multiple sectors of each base station in each cooperative sensing group are determined. Based on each cooperative sensing group, the echo signal data of each UAV in the area to be networked is acquired. Based on the echo signal data of each UAV, the first horizontal distance of each UAV is determined. Based on a preset fusion network model and combined with the first horizontal distance of each UAV, the target altitude of each UAV is determined. The fusion network model is constructed based on a convolutional neural network and a long short-term memory neural network. Based on the first horizontal distance and target altitude of each UAV, the theoretical beam tilt angle of each sector of each base station is calculated. With the minimum of the overlap range and blind zone range of the sector coverage as the optimization objective, a tilt angle optimization function is constructed to optimize the theoretical beam tilt angle of each sector of each base station in each cooperative sensing group, so as to obtain the actual execution tilt angle. Based on the actual execution tilt angle of each sector of each base station, network construction is carried out for each base station and each UAV in the area to be networked.

2. The multi-base station networking method for unmanned aerial vehicles according to claim 1, characterized in that, The process involves dividing each base station into multiple cooperative sensing groups based on their location within the area to be networked, and determining multiple sectors for each base station in each cooperative sensing group according to preset spatial layering requirements. Specifically, this includes: Based on the location coordinates of each base station in the area to be networked, the actual distance between any two base stations is calculated using the great circle distance formula. Each base station is divided into multiple collaborative sensing groups in a triangular distribution; within each collaborative sensing group, the base stations are adjacent and the actual distance between them meets the preset spacing requirements. Based on the preset spatial layering requirements and the base station antenna height, the sector coverage range and initial tilt angle of each sector of each base station in each cooperative sensing group are obtained, so as to determine multiple sectors of each base station in each cooperative sensing group.

3. The multi-base station networking method for unmanned aerial vehicles according to claim 1, characterized in that, The process involves acquiring echo signal data of each UAV within the network-to-be-networked area based on each cooperative sensing group, determining the first horizontal distance of each UAV based on the echo signal data, and determining the target altitude of each UAV based on a preset fusion network model and the first horizontal distance of each UAV. Specifically, this includes: Based on the flight trajectories of each UAV in the area to be networked, the corresponding cooperative sensing group is determined, and the echo signal data of the corresponding UAV is obtained based on the cooperative sensing group. Based on the echo signal data of each UAV, and based on the corresponding cooperative sensing group, the initial target altitude and first horizontal distance of each UAV are determined. Based on the fusion network model, the initial target altitude of each UAV is corrected by using the echo signal data of each UAV and the first horizontal distance, thereby determining the target altitude of each UAV.

4. A multi-base station networking method for unmanned aerial vehicles according to claim 3, characterized in that, The echo signal data includes echo signal data from the main base station and echo signal data from the slave base station; the step of determining the initial target altitude and first horizontal distance of each UAV based on the echo signal data of each UAV and the corresponding cooperative sensing group specifically includes: The echo signal data of each UAV and the corresponding cooperative sensing group's main base station are processed by time-domain filtering and frame synchronization to determine the signal delay of the echo signal data of each UAV's main base station. Based on the signal delay, the initial target altitude of each UAV and the corresponding cooperative sensing group's main base station is determined. Based on the time difference of arrival (TDOA) positioning algorithm, the echo signal data of the main base station and the echo signal data of the slave base station of each UAV are analyzed to determine the first horizontal distance between each UAV and the main base station of the corresponding cooperative sensing group.

5. A multi-base station networking method for unmanned aerial vehicles according to claim 3, characterized in that, Based on the fusion network model, the initial target altitude of each UAV is corrected using echo signal data and a first horizontal distance to determine the target altitude of each UAV. Specifically, this includes: In-phase orthogonal sequences are extracted from the echo signal data of each UAV to extract time-domain characteristics and obtain the time-domain features of each UAV. Based on the Fast Fourier Transform, the in-phase orthogonal sequences corresponding to each UAV are converted into Doppler spectrum signals to extract frequency domain characteristics and obtain the frequency domain features of each UAV. Based on the first horizontal distance and initial target altitude of each UAV, a two-dimensional feature vector of the airspace of each UAV is constructed to characterize the airspace characteristics and obtain the airspace characteristics of each UAV. Based on the fusion network model, the initial target altitude of each UAV is corrected and predicted by using the temporal, frequency, and spatial characteristics of each UAV, thereby determining the target altitude of each UAV.

6. A multi-base station networking method for unmanned aerial vehicles according to claim 1, characterized in that, The calculation of the theoretical beam tilt angle for each sector of each base station based on the first horizontal distance and target altitude of each UAV specifically includes: Based on the positional relationship between each UAV and the corresponding main base station of the cooperative sensing group, a solution triangle is constructed. Then, based on the first horizontal distance and target height of each UAV, combined with the base station antenna height, the solution triangle is used to obtain the theoretical beam tilt angle of each sector of each base station.

7. A multi-base station networking method for unmanned aerial vehicles according to claim 1, characterized in that, The optimization objective is to minimize the overlap and blind zone of the sector coverage area. A tilt angle optimization function is constructed to optimize the theoretical beam tilt angle of each sector of each base station in each cooperative sensing group, yielding the actual execution tilt angle. Specifically, this includes: A tilt angle optimization function is constructed with the goal of minimizing the overlap range and blind zone range of the sector coverage area; Based on the signal design and performance design of each base station, the first constraint of the tilt angle optimization function is determined; To solve the tilt angle optimization function, the first constraint is integrated into the tilt angle optimization function using a penalty function method. The theoretical beam tilt angle of each sector of each base station is substituted into the function, and the solution is obtained based on the gradient descent algorithm. This optimizes the theoretical beam tilt angle of each sector of each base station in each cooperative sensing group, and yields the actual execution tilt angle of each sector of each base station.

8. A positioning method for an unmanned aerial vehicle (UAV), characterized in that, include: Based on the multi-base station networking method for unmanned aerial vehicles as described in any one of claims 1 to 7, the feature data of each base station of multiple cooperative sensing groups in the area to be located is obtained, and real-time data acquisition is performed on the multiple cooperative sensing groups to obtain the echo signal data of each base station in each cooperative sensing group; wherein, the feature data includes topological features and coordinate features; Based on the topological characteristics of each base station in each collaborative sensing group, the corresponding echo signal data is processed to construct the observation vector of each collaborative sensing group, and based on the coordinate characteristics of each base station in each collaborative sensing group, the sensing matrix of each collaborative sensing group is constructed. Based on the pre-set compressed sensing joint reconstruction algorithm, the spatial point cloud optimization equation corresponding to each collaborative sensing group is constructed and solved through the observation vector and sensing matrix of each collaborative sensing group, so as to obtain the spatial point cloud corresponding to each collaborative sensing group. Based on the airspace point cloud corresponding to each collaborative sensing group, the real-time flight trajectories of several UAVs within each collaborative sensing group are generated to complete the positioning of each UAV in the area to be located.

9. A multi-base station networking device for unmanned aerial vehicles (UAVs), characterized in that, It includes a base station sector division module, a target height calculation module, a tilt angle calculation and optimization module, and a regional networking execution module; The base station sector division module is used to divide each base station into multiple collaborative sensing groups according to the location of each base station in the area to be networked, and to determine multiple sectors of each base station in each collaborative sensing group according to the preset spatial layering requirements. The target height calculation module is used to acquire echo signal data of each UAV in the area to be networked based on each cooperative sensing group, determine the first horizontal distance of each UAV based on the echo signal data of each UAV, and determine the target height of each UAV based on a preset fusion network model and the first horizontal distance of each UAV; wherein, the fusion network model is constructed based on a convolutional neural network and a long short-term memory neural network. The tilt angle calculation and optimization module is used to calculate the theoretical beam tilt angle of each sector of each base station based on the first horizontal distance and target height of each UAV, and construct a tilt angle optimization function with the minimum of the overlap range and blind zone range of the sector coverage as the optimization objective, so as to optimize the theoretical beam tilt angle of each sector of each base station and obtain the actual execution tilt angle. The regional networking execution module is used to construct a network for each base station and each UAV within the area to be networked, based on the actual execution tilt angle of each sector of each base station.

10. A positioning device for a drone, characterized in that, It includes a base station data acquisition module, a data observation and processing module, a spatial point cloud optimization and solution module, and a UAV real-time positioning module; The base station data acquisition module is used to acquire feature data of each base station of multiple cooperative sensing groups in the area to be located based on the multi-base station networking method of UAV as described in any one of claims 1 to 7, and to perform real-time data acquisition on the multiple cooperative sensing groups to obtain echo signal data of each base station in each cooperative sensing group; wherein, the feature data includes topological features and coordinate features. The data observation and sensing processing module is used to process the corresponding echo signal data according to the topological characteristics of each base station in each collaborative sensing group, construct the observation vector of each collaborative sensing group, and construct the sensing matrix of each collaborative sensing group according to the coordinate characteristics of each base station in each collaborative sensing group. The spatial point cloud optimization solution module is used to construct and solve the spatial point cloud optimization equation corresponding to each collaborative sensing group based on the preset compressed sensing joint reconstruction algorithm, through the observation vector and sensing matrix of each collaborative sensing group, to obtain the spatial point cloud corresponding to each collaborative sensing group. The real-time UAV positioning module is used to generate real-time flight trajectories of several UAVs within each collaborative sensing group based on the airspace point cloud corresponding to each collaborative sensing group, so as to complete the positioning of each UAV in the area to be located.

Images