Unmanned aerial vehicle flight early warning method and system in repeated area

Abstract

The invention discloses an unmanned aerial vehicle flight early warning method and system in a repeated area, and belongs to the technical field of unmanned aerial vehicle intelligent flight control. The objective of the invention is to solve the problem of accurate early warning of unmanned aerial vehicle flight. The method comprises the following steps: completely converting flight track segments of the unmanned aerial vehicle into track segment structure vectors; constructing a multi-scale similarity evaluation method to obtain a fusion similarity score of the trajectory segment structure vector; calculating repetition frequency of a track section; repeated track segments and flight concentration areas are obtained; the flight density of the unmanned aerial vehicle in the unit area in the high-heat area is calculated; designing a high-heat area to estimate a maximum safe flight density threshold value; calculating a comprehensive conflict risk score between the two unmanned aerial vehicles in the high-heat region by combining a direction difference, a speed difference and a time difference through Gaussian attenuation function modeling spatial distance influence; and generating a risk grade comprehensive score for grading response and strategy output. According to the invention, dynamic and adaptive maximum density threshold evaluation is realized, and the reliability of early warning is improved.

Description

Technical Field

[0001] This invention belongs to the field of intelligent flight control technology for unmanned aerial vehicles (UAVs), specifically relating to a method and system for early warning of UAV flight in a repetitive area. Background Technology

[0002] In urban low-altitude airspace, drone flight paths often exhibit significant repetition, especially during inspection, logistics, and security missions, creating high-frequency flight corridors. As flight density increases, airspace resources within these overlapping areas become highly congested, increasing the risk of flight conflicts, navigation interference, and even collisions. Therefore, implementing flight warnings in these overlapping areas can proactively identify potential density exceedances or flight conflict risks, providing real-time references for drone scheduling and path planning. This helps avoid safety issues caused by localized congestion or inconsistent flight directions, thereby ensuring airspace safety and operational efficiency in complex environments.

[0003] Repeated areas are typically the most concentrated air corridors for UAV missions, exhibiting high-frequency, regular, yet complex flight behavior. Providing flight warnings for these areas not only enables timely detection of high-risk factors such as abnormal density and conflicting flight directions, but also allows for dynamic assessment of the area's safety status, achieving refined airspace management. By integrating historical trajectory distribution with real-time flight data, the warning mechanism can provide risk level indications to the automatic control system, supporting tiered response mechanisms such as trajectory replanning, delayed takeoff, and priority passage, thereby effectively improving the collaborative capabilities and safe operation level of multiple UAV systems.

[0004] Existing methods are mostly based on trajectory overlap or density statistics, lacking in-depth modeling of the structural characteristics of flight behavior, and are unable to cope with complex situations such as trajectory deformation, directional differences, and local repetition. In addition, the risk assessment methods are simplistic and usually fail to comprehensively consider multiple factors such as flight direction consistency, spatial heat distribution, information entropy, and flight conflict scores, making it difficult to accurately characterize the regional risk level and affecting the accuracy of early warning. Summary of the Invention

[0005] The problem this invention aims to solve is to provide accurate early warning of unmanned aerial vehicle (UAV) flights, and proposes a method and system for early warning of UAV flights in repetitive areas.

[0006] To achieve the above objectives, the present invention provides the following technical solution:

[0007] A method for early warning of unmanned aerial vehicle (UAV) flights within a repetitive area includes the following steps:

[0008] S1. Collect UAV flight trajectory data, resample each collected UAV flight trajectory data, and then use a sliding window to divide it into UAV flight trajectory segments to form a set of UAV flight trajectory segments;

[0009] S2. Convert all the UAV flight trajectory segments obtained in step S1 into trajectory segment structure vectors;

[0010] S3. Construct a multi-scale similarity evaluation method to score and evaluate the similarity of trajectory segment structure vectors from two aspects: structural similarity and point matching similarity, and obtain the fusion similarity score of trajectory segment structure vectors;

[0011] S4. Calculate the trajectory segment repetition frequency using the fusion similarity score of the trajectory segment structure vector obtained in step S3, and count the total number of similar segments between each trajectory segment and all other trajectory segments based on a set threshold, and extract the set of repetitive trajectory segments.

[0012] S5. For the repeated trajectory segments obtained in step S4, introduce a kernel function to construct a trajectory heat function, then calculate the information entropy of the trajectory region based on the heat function, establish a heat map of the trajectory region, and obtain the flight concentration area;

[0013] S6. Extract high-heat regions from the heat map of the trajectory area obtained in step S5, and then calculate the flight density of UAVs per unit area within the high-heat regions; design an adaptive modeling method that combines information entropy, directional consistency and historical trajectory density to estimate the maximum safe flight density threshold for each high-heat region;

[0014] S7. For each high-heat region obtained in step S6, model the spatial distance influence using a Gaussian decay function, and calculate the comprehensive conflict risk score between two UAVs within the high-heat region by combining the direction difference, velocity difference, and time difference.

[0015] S8. Calculate the comprehensive conflict score for all UAVs in the high-heat area obtained in step S7, and sum the conflict risk score of the area. Then, normalize the flight density obtained in step S6 and the comprehensive conflict score of all UAVs respectively, and then weight and fuse them to generate a comprehensive risk level score, which is used for graded response and strategy output.

[0016] Furthermore, the specific implementation method of step S1 includes the following steps:

[0017] S1.1. Use GPS equipment to record the latitude and longitude data of the UAV flight trajectory, and convert the latitude and longitude data into coordinates in a plane coordinate system to obtain the collected UAV flight trajectory data;

[0018] S1.2. The UAV flight trajectory data obtained in step S1.1 is spatially uniformly resampled using linear interpolation to obtain the resampled UAV flight trajectory data;

[0019] S1.3. The resampled UAV flight trajectory data obtained in step S1.2 is divided into several small segments using a sliding window method. Each UAV flight trajectory segment includes a fixed number of segments. From a series of points, we obtain:

[0020]

[0021] in, Let j be the trajectory segment starting from the j-th trajectory point. for The number of points in the trajectory, p j Let j be the coordinates of the j-th trajectory point. This represents the number of trajectory points in the resampled trajectory.

[0022] All trajectory segments constitute the set of drone flight trajectory segments. .

[0023] Furthermore, in step S2, each UAV flight trajectory segment is transformed into a structure vector, including the initial direction angle, average curvature, total length, and the difference vector between all adjacent points, resulting in:

[0024]

[0025] in, For trajectory segment structure vectors, This is the starting direction angle; For the mean curvature, This represents the total length of the trajectory segment; Let x be the difference vector between the l-th trajectory point and the (l-1)-th trajectory point. It is the difference vector of the y-axis between the l-th trajectory point and the (l-1)-th trajectory point.

[0026] Furthermore, the specific implementation method of step S3 includes the following steps:

[0027] S3.1. Structural similarity, from a holistic perspective, measures the similarity in the morphological structure of trajectory segments through their direction, curvature, and length, thus obtaining a structural similarity score. ;

[0028] S3.2. Point matching similarity is calculated using a focused micro-alignment method. ;

[0029] S3.3. Calculate the fusion similarity score of the trajectory segment structure vectors, expressed as:

[0030]

[0031] in, To incorporate similarity weighting coefficients; The fusion similarity score is used to evaluate the structural vectors of the trajectory segments.

[0032] Furthermore, the specific implementation method of step S5 includes the following steps:

[0033] S5.1. Consider forming a spatial heat map for each trajectory segment determined to be repeated in step S4. Use a two-dimensional Gaussian kernel function to map the trajectory segment to a set of spatial heat functions. By superimposing the trajectory segment kernel functions of all repeated segments, construct the heat map of the entire region.

[0034] The kernel function K is calculated as follows:

[0035]

[0036] in, For point With trajectory segment The minimum Euclidean distance; To control the nuclear bandwidth of the heat diffusion range;

[0037] S5.2. Spatially superimpose the kernel functions of all repeating trajectory segments to form a heat function. This is used to represent the density of a location repeatedly traversed by a trajectory segment, and its expression is:

[0038]

[0039] in, A set of repeating trajectory segments;

[0040] S5.3. Divide the candidate region into multiple sub-regions c, calculate the probability of heat distribution, and then calculate the information entropy of the candidate region;

[0041]

[0042] H

[0043] in, Let c be the heat function of the c-th sub-region; This is the sum of the heat functions of all sub-regions; Let H be the heat distribution probability of the c-th sub-region; H is the information entropy, used to measure the dispersion of the heat distribution.

[0044] Furthermore, the specific implementation method of step S6 includes the following steps:

[0045] S6.1. Based on the heatmap, extract all locations where the heat value exceeds the threshold, and perform spatial connectivity partitioning on the extracted regions to obtain high heat regions;

[0046] S6.2. Use coordinate mapping to count the number of drones falling into the high-heat area obtained in step S6.1, and calculate the drone flight density per unit area. The expression is:

[0047] ,

[0048] in, High-temperature area Flight density; High-temperature area A collection of drones within the area; High-temperature area The number of drones within; High-temperature area The area; For the set of all drone states; Let g be the set of states of the g-th drone. Let G be the coordinates of the g-th UAV;

[0049] S6.3. Combining information entropy, directional consistency, and historical trajectory density, estimate the maximum safe flight density threshold for each high-heat region. The expression is:

[0050]

[0051] in, High-temperature area Maximum permissible flight density; Information entropy for the heat distribution in high-heat areas; High-temperature area Directional consistency score; High-temperature area The total length of the historical trajectory within; , , These are the weighting coefficients for information entropy, directional consistency score, and historical trajectory density, respectively, used to adjust for differences between dimensions.

[0052] Furthermore, the expression for the comprehensive conflict risk score between the two drones in the high-heat area in step S7 is as follows:

[0053]

[0054] in, The conflict risk score between the h1th and h2th drones; The distance between the h1th and h2th drones is obtained by the drone monitoring system. The parameters for controlling the effect of distance on the degree of attenuation are determined by expert experience; , , These are the weighting coefficients for the direction difference, velocity difference, and time difference, used to adjust for differences between dimensions, and determined by expert experience; , The flight directions of the h1th and h2th drones are obtained by the drone monitoring system. , The flight speeds of the h1th and h2th drones are obtained by the drone monitoring system, respectively. The time difference between the h1th and h2th drones entering the same area is obtained by the drone monitoring system. To prevent small constants from being divided by zero, they are determined by expert experience.

[0055] Furthermore, the specific implementation method of step S8 includes the following steps:

[0056] S8.1. Calculate and sum the comprehensive conflict scores of all UAVs within the high-heat area obtained in step S7 to obtain the total conflict risk score for the area, expressed as:

[0057]

[0058] in, High-temperature area Overall conflict risk score; For any pair of different drone combinations within the region; High-temperature area A collection of drones within the area;

[0059] S8.2. Generate a comprehensive risk level score, expressed as follows:

[0060]

[0061] in, High-temperature area Overall score; High-temperature area Flight density; High-temperature area The number of drones within; , These are the weighting coefficients for density risk and conflict risk, used to adjust for differences between dimensions.

[0062] A drone flight warning system for a repetitive area includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed, it implements the steps of the drone flight warning method for a repetitive area as described above.

[0063] The beneficial effects of this invention are:

[0064] The present invention discloses a method for early warning of UAV flights in repetitive areas. First, by segmenting the trajectory into local segments and constructing structural vectors, the accuracy and robustness of identifying repetitive paths are improved, adapting to diverse flight patterns. Second, by integrating structural similarity and point matching, the accuracy of trajectory segment comparison is improved. Third, the region division method based on heat and information entropy modeling can accurately identify trajectory clustering areas. In addition, by comprehensively considering directional consistency, historical density, and real-time status, a dynamic and adaptive maximum density threshold assessment is achieved, and a conflict scoring model and risk fusion mechanism are introduced to improve the reliability of the early warning. Attached Figure Description

[0065] Figure 1 This is a flowchart of a method for early warning of unmanned aerial vehicle (UAV) flights within a repetitive area, as described in this invention.

[0066] Figure 2 This is a bar chart showing the conflict risk score between any two drones in an embodiment of the present invention. Detailed Implementation

[0067] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to the accompanying drawings and specific embodiments. It should be understood that the specific embodiments described herein are only for explaining the invention and are not intended to limit the invention; that is, the described specific embodiments are merely a part of the embodiments of the invention, and not all of them. The components of the specific embodiments of the invention described and shown in the accompanying drawings can generally be arranged and designed in various different configurations, and the invention may also have other embodiments.

[0068] Therefore, the following detailed description of specific embodiments of the invention provided in the accompanying drawings is not intended to limit the scope of the claimed invention, but merely to illustrate selected specific embodiments of the invention. All other specific embodiments obtained by those skilled in the art based on these specific embodiments without inventive effort are within the scope of protection of this invention.

[0069] To further understand the invention's content, features, and effects, the following specific embodiments are provided, along with accompanying drawings. Figure 1 and attached Figure 2 Detailed explanation is as follows:

[0070] Example 1:

[0071] A method for early warning of unmanned aerial vehicle (UAV) flights within a repetitive area includes the following steps:

[0072] S1. Collect UAV flight trajectory data, resample each collected UAV flight trajectory data, and then use a sliding window to divide it into UAV flight trajectory segments to form a set of UAV flight trajectory segments;

[0073] Furthermore, the specific implementation method of step S1 includes the following steps:

[0074] S1.1. Use GPS equipment to record the latitude and longitude data of the UAV flight trajectory, and convert the latitude and longitude data into coordinates in a plane coordinate system to obtain the collected UAV flight trajectory data;

[0075] S1.2. The UAV flight trajectory data obtained in step S1.1 is spatially uniformly resampled using linear interpolation to obtain the resampled UAV flight trajectory data;

[0076] Furthermore, due to the different sampling frequencies of different devices, the spacing between trajectory points may be uneven, leading to significant structural differences between similar trajectories. Therefore, it is necessary to perform spatially uniform resampling on each trajectory to ensure that the distance between adjacent points is a fixed value. This improves the comparability between trajectory segments and avoids interference from trajectory point density in subsequent similarity calculations. Linear interpolation can be used for resampling.

[0077] ,

[0078] in, The trajectory after resampling; This represents the number of resampled trajectory points; Let be the coordinates of the i-th trajectory point. The planar coordinates of the i-th trajectory point are denoted as . ,in, Let x and y be the x and y coordinates of the i-th trajectory point, respectively. i = 1, 2, ..., n.

[0079] S1.3. The resampled UAV flight trajectory data obtained in step S1.2 is divided into several small segments using a sliding window method. Each UAV flight trajectory segment includes a fixed number of segments. From a series of points, we obtain:

[0080]

[0081] in, Let j be the trajectory segment starting from the j-th trajectory point. for The number of points in the trajectory, p j Let j be the coordinates of the j-th trajectory point. This represents the number of trajectory points in the resampled trajectory.

[0082] All trajectory segments constitute the set of drone flight trajectory segments. , .

[0083] Furthermore, to uncover recurring patterns in the trajectories, we divided each trajectory into several smaller segments using a sliding window approach. Each trajectory segment contains a fixed number of... Each continuous point represents a local shape of the trajectory. In this way, we can transform complex, long trajectories into multiple local segments, allowing similarity calculations to focus more on local structural features and better identify repeating paths. The sliding window typically uses a step size of 1 to ensure the complete coverage of the trajectory segments.

[0084] S2. Convert all the UAV flight trajectory segments obtained in step S1 into trajectory segment structure vectors;

[0085] Furthermore, in step S2, each UAV flight trajectory segment is transformed into a structure vector, including the initial direction angle, average curvature, total length, and the difference vector between all adjacent points, resulting in:

[0086]

[0087] in, For trajectory segment structure vectors, This is the starting direction angle; For the mean curvature, This represents the total length of the trajectory segment; Let x be the difference vector between the l-th trajectory point and the (l-1)-th trajectory point. It is the difference vector of the y-axis between the l-th trajectory point and the (l-1)-th trajectory point.

[0088] Furthermore, , refers to the difference vector Quantity; , refers to the difference vector Quantity.

[0089] in, , , These are the coordinates of the first two points of the trajectory segment; It is a two-parameter arctangent function. , It is the dot product of vectors; The number of trajectory points; Let be the coordinates of the j-th trajectory point. , It represents the magnitude of the vector.

[0090] Furthermore, this structured representation allows trajectory segments to be transformed into points in a high-dimensional space, enabling the use of vector distance to measure their similarity. This structural embedding method preserves both global and local features of the trajectory.

[0091] S3. Construct a multi-scale similarity evaluation method to score and evaluate the similarity of trajectory segment structure vectors from two aspects: structural similarity and point matching similarity, and obtain the fusion similarity score of trajectory segment structure vectors;

[0092] Furthermore, the specific implementation method of step S3 includes the following steps:

[0093] S3.1. Structural similarity, from a holistic perspective, measures the similarity in the morphological structure of trajectory segments through their direction, curvature, and length. The formula for calculating the structural similarity score is as follows:

[0094]

[0095] in, For the a-th trajectory segment The b-th trajectory segment The structural similarity score of the structural vectors; , These are the a-th trajectory segments. The b-th trajectory segment The structure vector; , They are respectively , The initial direction angle; , They are respectively , The average curvature; , They are respectively , The total length of the trajectory segment; , , , These are the initial direction angle, average curvature, total length of the trajectory segment, and weight coefficients of the structure vector, respectively.

[0096] S3.2. Point matching similarity is calculated using a focused micro-alignment method;

[0097] set up For trajectory segment The number of trajectory points; For trajectory segment Number of trajectory points; trajectory segment The coordinates of the trajectory of any point on the trajectory are denoted as ; For trajectory segment The number of trajectory points; For trajectory segment Number of trajectory points; trajectory segment The coordinates of the trajectory of any point on the trajectory are denoted as ;

[0098] S3.2.1. For trajectory segments , Collection trajectory segment and On Key points, including the start point, end point, midpoint, and points with equal length differences, are used to construct a set of point pairs. ,in, Where is the number of key points, and ji is the key point number. Representing trajectory segments and trajectory segment The jith key point; each point represents the corresponding point of two trajectory segments at the same progress position.

[0099] S3.2.2. Calculate the spatial difference between each pair of matched keypoints. The expression is:

[0100]

[0101] Then calculate the average difference between each pair of matched keypoints. ;

[0102] S3.2.3. Introduce the length of the trajectory segment as a normalization reference to calculate the point matching similarity score. The expression is:

[0103]

[0104] in, The similarity score is used to measure the location matching.

[0105] S3.3. Calculate the fusion similarity score of the trajectory segment structure vectors, expressed as:

[0106]

[0107] in, To incorporate similarity weighting coefficients; The fusion similarity score is used to evaluate the structural vectors of the trajectory segments.

[0108] Furthermore, geometric features such as direction, curvature, and length of trajectory segments are used to measure morphological similarity. This method is computationally efficient and robust, suitable for quickly filtering trajectory segments with similar structural features. However, structural features alone often fail to capture the subtle differences in spatial location between trajectory points. Therefore, point-location matching similarity is introduced. By comparing sequences of trajectory points, it can more precisely reflect the local differences in the spatial distribution of trajectory segments. The two complement each other: structural similarity focuses on macroscopic morphology, while point-location matching similarity focuses on microscopic alignment. Their combined use can effectively improve the accuracy and stability of trajectory comparison.

[0109] S4. Calculate the trajectory segment repetition frequency using the fusion similarity score of the trajectory segment structure vector obtained in step S3, and count the total number of similar segments between each trajectory segment and all other trajectory segments based on a set threshold, and extract the set of repetitive trajectory segments.

[0110] Furthermore, the repetition of trajectory segments is a key indicator for determining whether they belong to a "frequently used path". This involves counting the total number of similar segments between each trajectory segment and all other trajectory segments (i.e., the number of segments with a similarity greater than a similarity threshold). (Number of repetitions). If the repetition frequency of a certain segment exceeds the set repetition frequency threshold. We consider these as "repeating segments". This method can effectively filter out randomly similar noise segments.

[0111] The repetition frequency is calculated as follows:

[0112] in, This represents the set consisting of the entire trajectory segment. In addition to the trajectory segment Other trajectory segments besides; The similarity threshold is determined by expert experience; The repetition frequency.

[0113] The above formula means that for a certain trajectory segment Find the set formed by the entire trajectory segment. Other trajectory segments Other trajectory segments whose similarity to each other is greater than a similarity threshold And count the number of such trajectory segments, i.e., the repetition frequency. .

[0114] The set of repeating segments is defined as follows: , A set of repeating trajectory segments; The repetition frequency threshold is determined by expert experience; the above formula means that all cases with a repetition frequency ≥ the threshold are included. Select the trajectory segments that are "" to form a set of repeated trajectory segments. .

[0115] S5. For the repeated trajectory segments obtained in step S4, introduce a kernel function to construct a trajectory heat function, and then calculate the information entropy of the trajectory region based on the heat function to establish a heat map of the trajectory region;

[0116] Furthermore, each trajectory segment identified as repeating will spread its heat influence in the surrounding area, forming a spatial heat region. A two-dimensional Gaussian kernel function is used to map the trajectory segment to a set of spatial heat functions, representing its probability or influence of propagation in space. The bandwidth of the kernel function... This determines the spatial diffusion range. By superimposing the kernel functions of all repeating segments, a heatmap of the entire region can be constructed.

[0117] Furthermore, the specific implementation method of step S5 includes the following steps:

[0118] S5.1. Consider forming a spatial heat map for each trajectory segment determined to be repeated in step S4. Use a two-dimensional Gaussian kernel function to map the trajectory segment to a set of spatial heat functions. By superimposing the trajectory segment kernel functions of all repeated segments, construct the heat map of the entire region.

[0119] The kernel function K is calculated as follows:

[0120]

[0121] in, For point With trajectory segment The minimum Euclidean distance; To control the nuclear bandwidth of the heat diffusion range;

[0122] S5.2. Spatially superimpose the kernel functions of all repeating trajectory segments to form a heat function. This is used to represent the density of a location repeatedly traversed by a trajectory segment, and its expression is:

[0123]

[0124] in, A set of repeating trajectory segments;

[0125] S5.3. Divide the candidate region AR into multiple sub-regions c, calculate the heat distribution probability, and then calculate the information entropy of the candidate region;

[0126]

[0127]

[0128] in, Let c be the heat function of the c-th sub-region; This is the sum of the heat functions of all sub-regions; Let be the heat distribution probability of the c-th sub-region; Information entropy is used to measure the degree of dispersion in heat distribution.

[0129] S6. Extract high-heat regions from the heat map of the trajectory area obtained in step S5, and then calculate the flight density of UAVs per unit area within the high-heat regions; design an adaptive modeling method that combines information entropy, directional consistency and historical trajectory density to estimate the maximum safe flight density threshold for each high-heat region;

[0130] Furthermore, the specific implementation method of step S6 includes the following steps:

[0131] S6.1. Based on the heatmap, extract all locations where the heat value exceeds the threshold, and perform spatial connectivity partitioning on the extracted regions to obtain high heat regions;

[0132] S6.2. Use coordinate mapping to count the number of drones falling into the high-heat area obtained in step S6.1, and calculate the drone flight density per unit area. The expression is:

[0133] ,

[0134] in, High-temperature area Flight density; High-temperature area A collection of drones within the area; High-temperature area The number of drones within; High-temperature area The area; For the set of all drone states; Let g be the set of states of the g-th drone. Let G be the coordinates of the g-th UAV;

[0135] S6.3. Combining information entropy, directional consistency, and historical trajectory density, estimate the maximum safe flight density threshold for each high-heat region. The expression is:

[0136]

[0137] in, High-temperature area Maximum permissible flight density; Information entropy for the heat distribution in high-heat areas; High-temperature area Directional consistency score; High-temperature area The total length of the historical trajectory within; , , These are the weighting coefficients for information entropy, directional consistency score, and historical trajectory density, respectively, used to adjust for differences between dimensions.

[0138] Furthermore, the specific calculation method for the directional consistency score includes the following steps:

[0139] S6.3.1. Extract the region where the heat value reaches the set threshold according to the heat function, and regard the region boundary as the critical position where the heat value changes significantly, representing the potential structural region traversed by the trajectory set;

[0140] The location of the regional boundary is denoted as The definition is as follows:

[0141]

[0142] in, This is the heat threshold; The heat value is equal to The boundary;

[0143] S6.3.2. Analyze whether the trajectory directions within the region boundary obtained in step S6.3.1 are consistent. Calculate the gradient direction of the heatmap at each trajectory point and statistically analyze the consistency of directions on the region boundary.

[0144] First, using the spatial gradient of the heatmap, the gradient direction angle of each trajectory point is calculated, expressed as:

[0145]

[0146] in, The gradient direction angle; This is a two-dimensional arctangent function that returns the angle of the vector direction. These are the partial derivatives of the heat function in the x and y directions, respectively;

[0147] Then, the directional consistency score is defined as the calculation region boundary. The average angle between the orientation of all trajectory points and the gradient direction of the region. The consistency is obtained by taking the average of the cosine similarities:

[0148]

[0149] in, Scoring is given based on regional directional consistency.

[0150] S7. For each high-heat region obtained in step S6, model the spatial distance influence using a Gaussian decay function, and calculate the comprehensive conflict risk score between two UAVs within the high-heat region by combining the direction difference, velocity difference, and time difference.

[0151] Furthermore, the expression for the comprehensive conflict risk score between the two drones in the high-heat area in step S7 is as follows:

[0152]

[0153] in, The conflict risk score between the h1th and h2th drones; The distance between the h1th and h2th drones is obtained by the drone monitoring system. The parameters for controlling the effect of distance on the degree of attenuation are determined by expert experience; , , These are the weighting coefficients for the direction difference, velocity difference, and time difference, used to adjust for differences between dimensions, and determined by expert experience; , The flight directions of the h1th and h2th drones are obtained by the drone monitoring system. , The flight speeds of the h1th and h2th drones are obtained by the drone monitoring system, respectively. The time difference between the h1th and h2th drones entering the same area is obtained by the drone monitoring system. To prevent small constants from being divided by zero, they are determined by expert experience.

[0154] S8. Calculate the comprehensive conflict score for all UAVs in the high-heat area obtained in step S7, and sum the conflict risk score of the area. Then, normalize the flight density obtained in step S6 and the comprehensive conflict score of all UAVs respectively, and then weight and fuse them to generate a comprehensive risk level score, which is used for graded response and strategy output.

[0155] Furthermore, the specific implementation method of step S8 includes the following steps:

[0156] S8.1. Calculate and sum the comprehensive conflict scores of all UAVs within the high-heat area obtained in step S7 to obtain the total conflict risk score for the area, expressed as:

[0157]

[0158] in, High-temperature area Overall conflict risk score; For any pair of different drone combinations within the region; High-temperature area A collection of drones within the area;

[0159] S8.2. Generate a comprehensive risk level score, expressed as follows:

[0160]

[0161] in, High-temperature area Overall score; High-temperature area Flight density; High-temperature area The number of drones within; , These are the weighting coefficients for density risk and conflict risk, used to adjust for differences between dimensions.

[0162] The method in this embodiment was verified through the following practical experiment:

[0163] Four drone flight paths, A1, A2, A3, and A4, are defined for a given area. The trajectory points for path A1 are: (0, 0), (20, 28), (40, 55), (60, 76), (80, 92), (100, 101), (120, 102), (140, 96), (160, 84), (180, 66); the trajectory points for path A2 are: (2, 30), (22, 58), (42, 82), (62, 100), (82, 109), (102, 109), (122, 115), (142, 107), (162, 92), (182, 73); and the trajectory points for path A3 are: (4, 60), (24, 84), (44, 102). (64, 112), (84, 113), (104, 115), (124, 123), (144, 116), (164, 100), (184, 81), and the trajectory points corresponding to trajectory A4 are: (6, 90), (26, 114), (46, 131), (66, 138), (86, 135), (106, 131), (126, 141), (146, 126), (166, 107), (186, 89); using the method of this embodiment, the conflict risk score between any two UAVs is calculated as shown in Table 1 and Figure 2 As shown, the final overall score is 3.187.

[0164] Table 1

[0165]

[0166] This embodiment addresses the reliability issue of identifying repetitive regions and providing flight warnings in UAV trajectory data. The method employs structured trajectory segment modeling, similarity fusion calculation, heatmap construction, information entropy analysis, and comprehensive conflict risk scoring to characterize flight behavior and risk profiles from multiple dimensions, thereby achieving high-precision identification and effective warnings for flight activities within repetitive regions.

[0167] Example 2:

[0168] A drone flight warning system for a repetitive area includes a processor, a memory, and a computer program stored in the memory and executable on the processor. When the computer program is executed, it implements the steps of a drone flight warning method for a repetitive area as described in Embodiment 1.

[0169] It should be noted that relational terms such as "first" and "second" are used merely to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or apparatus. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or apparatus that includes said element.

[0170] Although this application has been described above with reference to specific embodiments, various modifications can be made and components can be replaced with equivalents without departing from the scope of this application. In particular, as long as there is no structural conflict, the features in the specific embodiments disclosed in this application can be combined with each other in any way. The lack of an exhaustive description of these combinations in this specification is merely for the sake of brevity and resource conservation. Therefore, this application is not limited to the specific embodiments disclosed herein, but includes all technical solutions falling within the scope of the claims.

Claims

1. A method for repeated regional UAV flight early warning, characterized in that, Comprising the following steps: S1. Collecting unmanned aerial vehicle flight trajectory data, resampling each collected unmanned aerial vehicle flight trajectory data, and then dividing into unmanned aerial vehicle flight trajectory segments using a sliding window to form a set of unmanned aerial vehicle flight trajectory segments; S2. Converting all unmanned aerial vehicle flight trajectory segments obtained in step S1 into trajectory segment structure vectors; S3. Building a multi-scale similarity evaluation method, scoring and evaluating the similarity of the trajectory segment structure vectors from two aspects of structural similarity and point matching similarity, and obtaining a fusion similarity score of the trajectory segment structure vectors; S4. Calculating the trajectory segment repetition frequency using the fusion similarity score of the trajectory segment structure vectors obtained in step S3, and based on a set threshold, counting the total number of similar segments between each trajectory segment and all other trajectory segments to extract a set of repeated trajectory segments; S5. For the repeated trajectory segments obtained in step S4, introducing a kernel function to build a trajectory heat function, and then calculating the information entropy of the trajectory region based on the heat function to establish a heat map of the trajectory region and obtain a concentrated flight region; S6. For the heat map of the trajectory region obtained in step S5, extracting a high-heat region, and then calculating the flight density of the unmanned aerial vehicle per unit area in the high-heat region; designing an adaptive modeling method, combining information entropy, direction consistency, and historical trajectory density to estimate the maximum safe flight density threshold for each high-heat region; S7. For each high-heat region obtained in step S6, modeling the influence of spatial distance through a Gaussian decay function, combining direction difference, speed difference, and time difference to calculate the comprehensive conflict risk score between two unmanned aerial vehicles in the high-heat region; S8. Calculating the comprehensive conflict score for all unmanned aerial vehicles in the high-heat region obtained in step S7, and then summing up the region's conflict risk score; then normalizing the flight density obtained in step S6 and the comprehensive conflict score of all unmanned aerial vehicles, and then weighting and fusing to generate a risk level comprehensive score for hierarchical response and strategy output. 2.The method of claim 1, wherein, The specific implementation method of step S1 comprises the following steps: S1.

1. Record the latitude and longitude data of the unmanned aerial vehicle flight trajectory using a GPS device, and convert the latitude and longitude data into coordinates in a plane coordinate system to obtain the collected unmanned aerial vehicle flight trajectory data; S1.

2. Perform spatial uniform resampling on the unmanned aerial vehicle flight trajectory data obtained in step S1.1 using linear interpolation to obtain resampled unmanned aerial vehicle flight trajectory data; S1.

3. The resampled UAV flight trajectory data obtained in step S1.2 is divided into several small segments using a sliding window method. Each UAV flight trajectory segment includes a fixed number of segments. From a series of points, we obtain: ； wherein, is a trajectory segment starting at the jth trajectory point, is is the number of trajectory points in the trajectory, p j is the jth trajectory point coordinate, is the number of trajectory points in the resampled trajectory; All trajectory segments form a set of UAV flight trajectory segments .

3. The method of claim 1 or 2, wherein, In step S2, each unmanned aerial vehicle flight trajectory segment is converted into a structure vector, including a starting direction angle, an average curvature, a total length, and a difference vector between all adjacent points, to obtain: ； wherein, is a trajectory segment structure vector, is a start direction angle; is an average curvature, is a total length of the trajectory segment; is a differential vector of the x-axis of the lth trajectory point and the (l-1)th trajectory point, is a differential vector of the y-axis of the lth trajectory point and the (l-1)th trajectory point.

4. The method of claim 3, wherein the method further comprises: The specific implementation method of step S3 comprises the following steps: S3.

1. Structural similarity From the overall perspective, the structural similarity score is obtained by measuring the similarity of the trajectory segment structure vector form through the direction, curvature, length of the trajectory segment ; S3.

2. Point matching similarity uses a focused micro-alignment method to calculate the point matching similarity score ; S3.

3. Calculate the fusion similarity score of the trajectory segment structure vector, which is expressed as: ； wherein, is a fusion similarity weighting coefficient; is a fusion similarity score of the trajectory segment structure vector.

5. The method of claim 4, wherein the method further comprises: The specific implementation method of step S5 comprises the following steps: S5.

1. Consider each trajectory segment determined to be repeated in step S4 to form a spatial heat region, map the trajectory segment to a set of spatial heat functions using a two-dimensional Gaussian kernel function, and construct a heat map of the entire region by superimposing the trajectory segment kernel functions of all repeated segments; The calculation method of the kernel function K is as follows: ； wherein, is a point with a minimum Euclidean distance from the trajectory segment ; and is a core bandwidth for controlling a heat diffusion range; S5.

2. The kernel functions of all repeated trajectory segments are superimposed in space to form the heat function , which is used to represent the density of a certain position being repeatedly passed by the trajectory segments, and the expression is as follows: ； wherein, is a set of repeating trajectory segments; S5.

3. Divide the candidate region into multiple sub-regions c, count the heat distribution probability, and then calculate the information entropy of the candidate region; ； H ； wherein, is the heat function of the cth sub-region; is the sum of heat functions of all sub-regions; is the heat distribution probability of the cth sub-region; H is the information entropy, used to measure the discrete degree of the heat distribution.

6. The method of claim 5, wherein the method further comprises: The specific implementation method of step S6 includes the following steps: S6.

1. On the basis of the heat map, extract the positions where the heat value exceeds the threshold value, perform spatial connected component segmentation on the extracted region, and obtain a high heat region; S6.

2. The number of unmanned aerial vehicles falling into the high heat region obtained in step S6.1 is counted by using the coordinate mapping method, and the flight density of the unmanned aerial vehicles in the unit area is calculated, and the expression is: ， ； wherein, is a flight density of the high-heat region ; is a set of drones within the high-heat region ; is a number of drones within the high-heat region ; is an area of the high-heat region ; is a set of all drone states; is a set of states of the gth drone, is a coordinate of the gth drone; S6.

3. Combine the information entropy, the direction consistency and the historical trajectory density to estimate the maximum safe flight density threshold value for each high heat region. The expression is: ； wherein, is the maximum allowed flight density of the high-heat region; is the information entropy of the heat distribution of the high-heat region; is the direction consistency score of the high-heat region; is the total length of historical trajectories within the high-heat region; are the weight coefficients of the information entropy, the direction consistency score, and the historical trajectory density, respectively, for adjusting the dimensional differences.​​​​​ 7. The method of claim 6, wherein the method further comprises: The expression of the comprehensive conflict risk score between two unmanned aerial vehicles in the high heat region of step S7 is: ； wherein, is the conflict risk score between the h1th and h2th UAVs; is the distance between the h1th and h2th UAVs, obtained by the UAV monitoring system; is a parameter that controls the degree of attenuation of the distance, determined by expert experience; , , are weight coefficients of the direction difference, speed difference, and time difference, respectively, used to adjust the differences between dimensions, determined by expert experience; , are the flight directions of the h1th and h2th UAVs, respectively, obtained by the UAV monitoring system; , are the flight speeds of the h1th and h2th UAVs, respectively, obtained by the UAV monitoring system; is the time difference between the h1th and h2th UAVs entering the same region, obtained by the UAV monitoring system; is a small constant to prevent division by zero, determined by expert experience.

8. The method of claim 7, wherein the method further comprises: The specific implementation method of step S8 includes the following steps: S8.

1. Accumulate the comprehensive conflict scores of all unmanned aerial vehicles in the high heat region obtained in step S7 to obtain the total conflict risk score of the region, and the expression is: ； wherein, is a high heat area total conflict risk score; is any pair of different drone combinations within the area; is a high heat area within the area; S8.

2. Generate a risk level comprehensive score, and the expression is: ； wherein, is a comprehensive score of high-heat areas ; is a flight density of high-heat areas ; is a number of drones within high-heat areas ; , are weight coefficients of density risk and conflict risk, respectively, for adjusting the dimensional difference.

9. A repeated regional UAV flight early warning system, characterized in that, A computer program product including a processor, a memory and a computer program stored in the memory and executable on the processor, wherein the computer program implements the steps of the repeated region unmanned aerial vehicle flight warning method according to any one of claims 1-8 when executed.

Images