A multi-scene-oriented unmanned aerial vehicle cluster cooperative control training system

Abstract

The application relates to the technical field of unmanned aerial vehicle control, and particularly discloses a kind of unmanned aerial vehicle cluster cooperative control training systems for multiple scenarios, after determining flight interval as the main training element, the same unmanned aerial vehicle coincidence screening is carried out to high-risk collision group and low boundary overlap group, target unmanned aerial vehicle is obtained, special individuals in cooperative flight that exist both high collision risk and low boundary overlap can be accurately positioned, and the flight interval range of target unmanned aerial vehicle is determined, moreover, since target unmanned aerial vehicle is the connecting point of high-risk collision group and low boundary overlap group, the cooperative ability of target unmanned aerial vehicle and adjacent unmanned aerial vehicle is improved, the flight order of local area can be improved, the collision risk can be reduced, the cooperative performance of the whole cluster can be improved, and the precision of high-rise mapping edge overlap degree is improved.

Description

Technical Field

[0001] This invention relates to the field of unmanned aerial vehicle (UAV) control technology, and more specifically to a UAV swarm collaborative control training system for various scenarios. Background Technology

[0002] With the rapid development of drone technology, drone swarms are increasingly being used in collaborative mapping, environmental monitoring, and military reconnaissance. Particularly in high-rise building mapping scenarios, deploying multiple drones to simultaneously collect image data from different directions can significantly improve mapping efficiency and expand coverage. However, when drone swarms collaboratively perform mapping tasks, effectively controlling the flight distance between drones, avoiding collision risks, and ensuring the accuracy of boundary overlap in mapping images have become important areas of current research.

[0003] Currently, existing UAV swarm collaborative control methods mostly focus on preset flight paths, GPS-based obstacle avoidance, or simple formation-keeping strategies. These methods have the following shortcomings in practical applications: First, existing technologies lack in-depth utilization of UAV mapping and imaging data, failing to effectively correlate the degree of mapping boundary overlap with flight safety risks, resulting in the inability to accurately identify key UAV combinations with high collision risks during flight. Second, when UAV swarms exhibit low mapping boundary overlap or excessively close flight spacing, existing methods struggle to determine whether flight spacing control is a major training factor affecting collaborative mapping effectiveness, thus hindering targeted optimization of training strategies and impacting the improvement of swarm collaborative control capabilities.

[0004] Furthermore, existing technologies lack effective identification methods and parameter determination methods for special drone individuals that have both high collision risks and low boundary overlap. This makes it difficult to accurately locate weak links during training, and difficult to quantify and optimize the flight distance range of key drones, thus restricting the scientific nature and effectiveness of drone swarm collaborative control training.

[0005] Therefore, this invention provides a collaborative control training system for drone swarms in various scenarios. Summary of the Invention

[0006] The purpose of this invention is to provide a collaborative control training system for drone swarms in various scenarios to solve the aforementioned background problems.

[0007] The objective of this invention can be achieved through the following technical solutions: A collaborative control training system for UAV swarms in multiple scenarios includes: Overlapping height and low analysis module: During the training process of simulating high-rise building mapping of UAVs in UAV swarm, the mapping images corresponding to each UAV track are acquired, and the boundary overlap of the mapping images corresponding to adjacent UAV tracks is analyzed to filter out low-degree boundary overlap groups and high-degree boundary overlap groups. High-overlap collision analysis module: assesses the degree of collision between two drones within the high-overlap analysis group and identifies high-risk collision groups; Training element analysis module: In the high-risk collision group, the correlation between the flight distance between adjacent UAVs and the degree of high-risk collision is evaluated. At the same time, in the low-degree boundary overlap group, the correlation between the flight distance between adjacent UAVs and the degree of boundary overlap is evaluated to determine whether the flight distance is a major training element when adjacent UAVs are conducting collaborative mapping. Training Element Control Module: After determining flight spacing as the main training element, the module performs overlap screening with UAVs on high-risk collision groups and low-degree boundary overlap groups to obtain target UAVs and determine the flight spacing range of the target UAVs.

[0008] As a further aspect of the present invention, the process of performing boundary overlap analysis on the mapping images corresponding to adjacent UAV flight paths is as follows: The drones in the drone swarm are combined to form multiple boundary overlap analysis groups. Within each boundary overlap analysis group, based on the same floor of the high-rise building, the mapping images of each drone are combined into multiple mapping image analysis pairs. Within each mapping and imaging analysis pair, the common imaging boundary of the two mapping and imaging pairs is extracted as the common basic analysis boundary. The overlap is compared and the overlap length is extracted. The ratio of the overlap length to the actual building floor edge length is calculated to obtain the boundary unit overlap length ratio. The mean of the overlap length ratio of the boundary units for each mapping and imaging analysis pair is summed to obtain the boundary overlap analysis value; If the boundary overlap analysis value is greater than or equal to the boundary overlap analysis threshold, the corresponding boundary overlap analysis group is marked as a high overlap analysis group. If the boundary overlap analysis value is less than the boundary overlap analysis threshold, the corresponding boundary overlap analysis group is marked as a low-degree overlap analysis group.

[0009] As a further aspect of the present invention, the process for evaluating the degree of collision between the two machines is as follows: The simulation surveying training cycle is divided into several simulation surveying training periods. The flight tracks of two UAVs in the same simulation surveying training period when they share the same basic analysis boundary on the same floor of the high-overlap analysis group are extracted as the same flight tracks. The simulated surveying and mapping training period is divided into several simulated surveying and mapping training time points. The flight coordinates of the UAV on the same flight track at each simulated surveying and mapping training time point are obtained as the same flight coordinates. The distance between the corresponding same flight coordinates of two UAVs in the high overlap analysis group is calculated as the unit aircraft spacing. Simulated surveying training points where the distance between aircraft units is less than or equal to the preset distance between aircraft units are considered high-risk collision points.

[0010] As a further aspect of the present invention, the process for determining high-risk collision groups is as follows: The proportion of high-risk collision points to the total number of all simulated mapping training points is used as the high-risk collision point ratio. The average value of the distance between aircraft in the unit corresponding to all simulated surveying and mapping training points is calculated to obtain the distance-collision analysis value. The high-risk collision analysis value is obtained by calculating the ratio of the number of high-risk collision times to the distance collision analysis value. If the high-risk collision analysis value is greater than the high-risk collision analysis threshold, it is marked as a high-risk collision group.

[0011] As a further aspect of the present invention, the process of analyzing the flight distance between adjacent UAVs and the degree of high-risk collision within the high-risk collision group is as follows: Extract the co-measured flight coordinates of the two UAVs at each simulated surveying and mapping training point during the simulated surveying and mapping training period, calculate and process them using the Euclidean distance formula, and calculate the ratio with the set flight track distance to obtain the high wind overlap flight distance ratio; Extract the proportion of high-risk collision points in each simulated mapping training period to the total number of all simulated mapping training periods to obtain the high-risk collision point ratio for each period; The distance between aircraft in each unit corresponding to the high-risk collision point in each simulated mapping training period is obtained and the average value is calculated to obtain the unit distance collision value for that period. The high-risk collision time point ratio of the time period is calculated by comparing it with the distance collision value of the time period unit to obtain the high-wind collision value of the time period. The high-wind collision values ​​of adjacent simulated mapping training time periods are then subtracted to obtain the high-wind overlap difference value of adjacent time periods.

[0012] As a further aspect of the present invention, the process for evaluating the flight distance and high-risk collision level between adjacent UAVs within the high-risk collision group is as follows: The difference between the high wind overlap flight distance ratios corresponding to adjacent simulated surveying and mapping training periods is obtained by subtracting the high wind unit flight distance difference between adjacent segments. The high wind overlap difference between adjacent time periods under the same adjacent simulated mapping training period is calculated by the ratio of the high wind unit flight distance difference between adjacent segments to obtain the high wind flight weight analysis value. After calculating the standard deviation and summation mean of multiple high-wind flight overlap analysis values, the coefficient of variation is used for calculation to obtain the high-wind flight distance overlap analysis value. If it is less than or equal to the high-wind flight distance overlap analysis threshold, it is displayed as a high-wind flight distance close overlap signal.

[0013] As a further aspect of the present invention, the analysis process within the low-overlap analysis group is as follows: Extract the mapping images of each UAV during the same simulated mapping training period when they share the basic analysis boundary on the same floor of the high-rise building. Extract the common basic analysis boundary of the mapping images of the two UAVs and obtain the overlap length ratio of the boundary cells; Extract the same flight coordinates of the two UAVs at each simulated surveying and mapping training point during the simulated surveying and mapping training period, and take the same flight coordinates at the same simulated surveying and mapping training point as a coordinate comparison group to obtain multiple coordinate comparison groups. The Euclidean distance formula is used to calculate and process the same flight coordinates within multiple coordinate comparison groups, and the ratio is calculated with the set flight track distance to obtain the low-overlap flight distance ratio.

[0014] As a further aspect of the present invention, the evaluation process within the low-overlap analysis group is as follows: Within the simulated mapping training cycle, the difference between the overlap length ratios of boundary units corresponding to adjacent simulated mapping training periods is obtained. The difference between the low-overlapping flight distance ratios of adjacent simulated mapping training periods is obtained by subtracting the flight distance difference of low-overlapping units in adjacent segments. The low-overlap difference between adjacent time periods under the same adjacent simulated mapping training period is calculated by the ratio of the flight distance difference of low-weight units in adjacent segments to obtain the low-weight analysis value. After calculating the standard deviation and the mean of all low-gravity flight weight analysis values, the values ​​are recalculated using the coefficient of variation formula to obtain the low-gravity flight distance overlap analysis value. If the value is less than or equal to the low-gravity flight distance overlap analysis threshold, it is displayed as a low-gravity flight distance close overlap signal.

[0015] As a further aspect of the present invention, the analysis process for determining whether flight distance is a primary training element is as follows: If, when evaluating the correlation between the flight distance and boundary overlap between adjacent UAVs in the high-risk collision group and the low-degree boundary overlap group, respectively, both signals of tight flight distance overlap under high-risk and low-degree boundary overlap are displayed, it indicates that the flight distance is the main training element when adjacent UAVs conduct collaborative mapping.

[0016] As a further aspect of the present invention, the process of acquiring the target UAV and determining the flight distance range of the target UAV is as follows: All high-risk collision groups and adjacent UAVs in low-degree boundary overlap groups are compared for overlap. High-risk collision groups and low-degree boundary overlap groups with the same UAVs are merged into a high-wind low-weight control group, and the same UAVs are marked as target UAVs. Within the low-degree boundary overlap group, the boundary unit overlap length ratio between the target UAV and the adjacent UAV in each simulated mapping training period is extracted and compared. The low-degree overlap flight distance ratio corresponding to the largest boundary unit overlap length ratio is selected and multiplied with the set flight track distance to calculate the target flight distance lower limit. In the high-risk collision group, the collision distance values ​​of the target UAV and adjacent UAVs in each time period of the simulated mapping training are extracted and compared. The high-wind overlap distance ratio corresponding to the smallest time period unit collision distance value is selected and multiplied with the set flight track distance to calculate the upper limit of the target distance. The flight distance range is determined based on the lower limit and upper limit of the target distance.

[0017] The beneficial effects of this invention are as follows: 1. In the process of training drones in a drone swarm to simulate high-rise building mapping, this invention acquires mapping images corresponding to the flight paths of each drone and performs boundary overlap analysis to screen out low-degree boundary overlap groups and high-degree boundary overlap groups. Since the mapping boundary overlap between drones in the high-degree overlap analysis group is relatively high, it means that the distance between drones during flight is relatively close, and the risk of collision is relatively high. Therefore, it is necessary to focus on real-time monitoring of the distance and flight path between drones to detect potential collision risks in advance and take timely avoidance measures. On the other hand, the mapping boundary overlap between drones in the low-degree boundary overlap group is relatively low, and it is necessary to improve the accuracy of collaborative mapping between drones.

[0018] 2. This invention assesses the collision severity of two UAVs within a high-overlap analysis group to identify high-risk collision groups. This helps to reflect weaknesses in UAV collaborative control, particularly in flight spacing control and obstacle avoidance capabilities. By strengthening the flight spacing between UAVs, obstacles can be avoided promptly and effectively. Within the high-risk collision group, the correlation between the flight spacing between adjacent UAVs and the high-risk collision severity is assessed. Simultaneously, within the low-boundary overlap group, the correlation between the flight spacing between adjacent UAVs and the boundary overlap is assessed. This determines whether flight spacing is a primary training element when adjacent UAVs are conducting collaborative mapping. When flight spacing is a primary training element, the training strategy can focus on how to accurately control flight spacing. Targeted training can significantly improve the UAV's ability to control flight spacing during collaborative flight. If flight spacing is not a primary training element, the training strategy can focus more on other aspects of collaborative control.

[0019] 3. After determining flight spacing as the main training element, this invention performs overlap screening of high-risk collision groups and low-degree boundary overlap groups with the same UAV to obtain target UAVs. It can accurately locate special individuals that have both high collision risk and low-degree boundary overlap in cooperative flight, and determine the flight spacing range of the target UAVs. Moreover, since the target UAVs serve as the connection points between high-risk collision groups and low-degree boundary overlap groups, the cooperative capability between the target UAVs and adjacent UAVs is improved. This can improve the flight order in local areas, reduce collision risk, improve the cooperative performance of the entire cluster, and improve the accuracy of the edge overlap of high-rise building mapping. Attached Figure Description

[0020] The invention will now be further described with reference to the accompanying drawings.

[0021] Figure 1 This is a block diagram of a collaborative control training system for unmanned aerial vehicle (UAV) swarms oriented towards multiple scenarios according to the present invention. Figure 2 This is a flowchart of the steps in a multi-scenario collaborative control training system for unmanned aerial vehicle (UAV) swarms in this invention. Figure 3 This is a decision flowchart within a collaborative control training system for drone swarms oriented towards multiple scenarios, as described in this invention. Detailed Implementation

[0022] To make the technical means, creative features, objectives and effects of this invention easier to understand, the invention will be further described below in conjunction with specific embodiments.

[0023] Example 1

[0024] like Figures 1-2 As shown, this embodiment of the invention provides a collaborative control training system for UAV swarms in various scenarios, including: Step 1: During the training process of simulating high-rise building mapping for drones in the drone swarm, the mapping images corresponding to each drone's flight path are acquired, and the boundary overlap of the mapping images corresponding to adjacent drone flight paths is analyzed to filter out low-degree boundary overlap groups and high-degree boundary overlap groups. It should be noted that the high-rise building being surveyed is a standard rectangle, and the drone swarm includes four surveying drones: an east-facing drone, a south-facing drone, a west-facing drone, and a north-facing drone. In some embodiments, eastward-facing UAVs and southward-facing UAVs are combined into a group of boundary overlap analysis groups; The eastward-facing and northward-facing drones are combined into a boundary overlap analysis group; The southbound and westbound UAVs are combined into a boundary overlap analysis group; The westward-bound and northward-bound UAVs are combined into a boundary overlap analysis group; For example, a set of boundary overlap analysis groups is selected, and within the boundary overlap analysis group, the mapping images of each floor of the high-rise building corresponding to each UAV are obtained; Based on the same floor of the high-rise building, the mapping images corresponding to each UAV are combined into a mapping image analysis pair, resulting in multiple mapping image analysis pairs. Each mapping image analysis pair contains only two mapping images. Within the mapping and imaging analysis pair, the common imaging boundary between the two mapping and imaging pairs is extracted as the common basic analysis boundary; It should be noted that the common imaging boundary of the two survey images refers to the two drones surveying from different directions. For example, if an eastward drone surveys the east side of a tall building, the resulting image of each floor on the east side of the tall building is called the eastward survey image. Similarly, if a southward drone surveys the south side of a tall building, the resulting image of each floor on the south side of the tall building is called the southward survey image. Under the premise of surveying the same floors of the tall building, the left edge of the eastward survey image and the right edge of the southward survey image are the same edge in the actual spatial dimension of the tall building. The two survey images are compared to overlap the common foundation analysis boundary, the overlap length of the common foundation analysis boundary is extracted, and the ratio of the overlap length of the boundary unit is calculated with the actual floor edge length of the high-rise building to obtain the overlap length ratio of the boundary unit. The boundary overlap length ratio of each mapping and imaging analysis pair is extracted, and the summation and mean value are calculated to obtain the boundary overlap analysis value; It is understandable that the boundary overlap analysis value represents the average degree of overlap of UAV mapping images from different directions on the common boundary within the boundary overlap analysis group. Specifically, the larger the boundary overlap analysis value, the higher the degree of overlap of adjacent UAV images on the common boundary within the analyzed boundary overlap analysis group; the smaller the boundary overlap analysis value, the lower the degree of overlap of adjacent UAV images on the common boundary within the analyzed boundary overlap analysis group. If the boundary overlap analysis value is greater than or equal to the boundary overlap analysis threshold, it indicates that the adjacent UAV images in the analyzed boundary overlap analysis group have a high degree of overlap on the common boundary, and are marked as a high overlap analysis group. If the boundary overlap analysis value is less than the boundary overlap analysis threshold, it indicates that the adjacent UAV images in the analyzed boundary overlap analysis group have a low degree of overlap on the common boundary, and are marked as a low-degree overlap analysis group.

[0025] The specific solution in this embodiment is as follows: During the training process of simulating high-rise building mapping for drones in a drone swarm, the mapping images corresponding to the flight paths of each drone are acquired, and boundary overlap analysis is performed to screen out low-degree boundary overlap groups and high-degree boundary overlap groups. Since the mapping boundary overlap between drones in the high-degree overlap analysis group is relatively high, it means that the distance between them during flight is relatively close, and the risk of collision is relatively high. Therefore, it is necessary to focus on real-time monitoring of the distance and flight path between drones to detect potential collision risks in advance and take timely avoidance measures. On the other hand, the mapping boundary overlap between drones in the low-degree boundary overlap group is relatively low, and it is necessary to improve the accuracy of collaborative mapping between drones.

[0026] Example 2

[0027] like Figures 1-2 As shown, this embodiment of the invention provides a collaborative control training system for UAV swarms in various scenarios, including: Step 2: Assess the degree of collision between two drones within the high-overlap analysis group to identify the high-risk collision group; In some embodiments, a simulation mapping training cycle is set and equally divided into several simulation mapping training periods, wherein adjacent simulation mapping training periods have the same duration; It should be noted that within each simulated surveying and mapping training period, adjacent UAVs within the boundary overlap analysis group collaboratively survey and map the same floor of a high-rise building, meaning that each simulated surveying and mapping training period corresponds to a boundary unit overlap length ratio. Extract the flight paths of each UAV in the high-overlap analysis group when they are in the same simulated surveying training period on the same floor of the high-rise building and share the basic analysis boundary, and use them as the same survey flight paths; The simulated surveying and mapping training period is divided into several simulated surveying and mapping training time points. The flight coordinates of the UAV on the same flight track at each simulated surveying and mapping training time point are obtained as the same flight coordinates. The distance between the corresponding flight coordinates of two UAVs in the highly overlapping analysis group is obtained as the unit aircraft spacing. If the distance between unit aircraft is less than or equal to the preset distance between unit aircraft, it means that at the same simulated surveying training point, the UAVs are flying close together when surveying the same floor of the high-rise building, and the probability of collision is high. The corresponding simulated surveying training point is then taken as a high-risk collision point. If the distance between unit aircraft is greater than the preset distance between unit aircraft, it means that at the same simulated surveying training point, the UAV flies a greater distance when surveying the same floor of the high-rise building, and the probability of collision is lower. The corresponding simulated surveying training point is taken as the low-risk collision point. The proportion of high-risk collision points to the total number of all simulated mapping training points is used as the high-risk collision point ratio. The average value of the distance between aircraft units at all simulated surveying and mapping training points in the simulated surveying and mapping training cycle is calculated to obtain the distance-collision analysis value. The high-risk collision analysis value is obtained by calculating the ratio of the number of high-risk collision times to the distance collision analysis value. Understandably, the high-wind collision analysis value reflects two factors: the frequency of high-risk collision events and the average flight distance between drones. Specifically, a larger high-wind collision analysis value indicates that in the simulated mapping training, the two drones are not only frequently in a high-collision-risk state, but also fly relatively close together overall, meaning that the probability of a collision between the two drones is relatively high. Conversely, a smaller high-wind collision analysis value indicates that in the simulated mapping training, the two drones are not only frequently in a low-collision-risk state, but also fly relatively far together overall, meaning that the probability of a collision between the two drones is relatively low. If the high-wind collision analysis value is greater than the high-wind collision analysis threshold, it means that in the simulated mapping training, the two UAVs are not only frequently in a high-collision risk state, but also fly at a relatively close distance overall, which means that the probability of the two UAVs colliding is relatively high. The highly overlapping analysis group is marked as a high-risk collision group. If the high-wind collision analysis value is less than or equal to the high-wind collision analysis threshold, it means that in the simulated mapping training, the two UAVs are not only frequently in a low-collision-risk state, but also fly at a relatively long distance overall, which means that the probability of the two UAVs colliding is small. The highly overlapping analysis group is then marked as a low-risk collision group.

[0028] Step 3: In the high-risk collision group, the correlation between the flight distance between adjacent UAVs and the degree of high-risk collision is evaluated. At the same time, in the low-degree boundary overlap group, the correlation between the flight distance between adjacent UAVs and the degree of boundary overlap is evaluated to determine whether the flight distance is a major training element when adjacent UAVs are conducting collaborative mapping. In some embodiments, within a high-risk collision group, the mapping images of each UAV during the same simulated mapping training period are extracted when they share a basic analysis boundary on the same floor of a high-rise building. Extract the common basic analysis boundary of the mapping images of the two UAVs and obtain the overlap length ratio of the boundary cells; Extract the same flight coordinates of the two UAVs at each simulated surveying and mapping training point during the simulated surveying and mapping training period, and take the same flight coordinates at the same simulated surveying and mapping training point as a coordinate comparison group to obtain multiple coordinate comparison groups. The Euclidean distance formula is used to calculate and process the same flight coordinates within multiple coordinate comparison groups, and the ratio is calculated with the set flight track distance to obtain the high wind overlap flight distance ratio. Extract the proportion of high-risk collision points in each simulated mapping training period to the total number of simulated mapping training points in all simulated mapping training periods, and obtain the high-risk collision point ratio for each period. The distance between aircraft in each unit corresponding to the high-risk collision point in each simulated mapping training period is obtained and the average value is calculated to obtain the unit distance collision value for that period. The high-risk collision time point ratio of the time period is calculated by comparing it with the flight distance collision value of the time period unit to obtain the high-risk collision value of the time period. Within the simulation surveying and mapping training cycle, the high wind collision values ​​corresponding to adjacent simulation surveying and mapping training periods are subtracted to obtain the high wind overlap difference between adjacent periods. The difference between the high wind overlap flight distance ratios corresponding to adjacent simulated surveying and mapping training periods is obtained by subtracting the high wind unit flight distance difference between adjacent segments. The high wind overlap difference between adjacent time periods under the same simulated surveying and mapping training period is calculated by ratioing the high wind unit flight distance difference between adjacent time periods to obtain the high wind flight weight analysis value. After calculating the standard deviation and the mean of the high wind flight overlap analysis values, the coefficient of variation was used to process the data to obtain the high wind flight distance overlap analysis values. It is understandable that the high-wind flight distance overlap analysis value represents the degree of correlation between the change in the degree of UAV collision risk and the degree of flight distance during different simulated mapping training periods. Specifically, if the high-wind flight distance analysis value is larger, it means that the correlation between the degree of flight distance between adjacent UAVs and the collision risk is not close during collaborative mapping training. If the high-wind flight distance analysis value is smaller, it means that the correlation between the degree of flight distance between adjacent UAVs and the collision risk is closer during collaborative mapping training. If the high-wind flight distance overlap analysis value is greater than the high-wind flight distance overlap analysis threshold, it indicates that within the high-risk collision group, the flight distance between adjacent UAVs is not closely related to the boundary overlap, which is a signal of non-close high-wind flight distance overlap. If the high-wind flight distance overlap analysis value is less than or equal to the high-wind flight distance overlap analysis threshold, it indicates that within the high-risk collision group, the flight distance between adjacent UAVs is closely related to the boundary overlap, which is a signal of tight high-wind flight distance overlap. For example, within a low-overlap analysis group, the mapping images of each UAV during the same simulated mapping training period are extracted when they share a basic analysis boundary on the same floor of a high-rise building. Extract the common basic analysis boundary of the mapping images of the two UAVs and obtain the overlap length ratio of the boundary cells; Extract the same flight coordinates of the two UAVs at each simulated surveying and mapping training point during the simulated surveying and mapping training period, and take the same flight coordinates at the same simulated surveying and mapping training point as a coordinate comparison group to obtain multiple coordinate comparison groups. The Euclidean distance formula is used to calculate and process the same flight coordinates within multiple coordinate comparison groups, and the ratio is calculated with the set flight track distance to obtain the low-overlap flight distance ratio. Within the simulated mapping training cycle, the difference between the overlap length ratios of boundary units corresponding to adjacent simulated mapping training periods is obtained. The difference between the low-overlapping flight distance ratios of adjacent simulated mapping training periods is obtained by subtracting the flight distance difference of low-overlapping units in adjacent segments. The low overlap difference between adjacent time periods under the same simulated mapping training period is calculated by the ratio of the flight distance difference between adjacent time period units to obtain the low overlap analysis value. After calculating the standard deviation and the mean of all boundary weight flight distance analysis values, and then using the coefficient of variation formula for further calculation, the low weight flight distance overlap analysis values ​​are obtained. It is understandable that the flight distance boundary overlap analysis value refers to an indicator that comprehensively considers the dynamic relationship between flight distance and boundary overlap in UAV collaborative mapping training within the low overlap analysis group. Specifically, if the low overlap flight distance overlap analysis value is larger, it indicates that the change in the degree of overlap of UAV flight boundaries and the change in the degree of deviation of flight distance are not closely related in different simulated mapping training periods. If the low overlap flight distance overlap analysis value is smaller, it indicates that the change in the degree of overlap of UAV flight boundaries and the change in the degree of deviation of flight distance are more closely related in different simulated mapping training periods. If the low-weight flight distance overlap analysis value is greater than the low-weight flight distance overlap analysis threshold, it indicates that within the low-degree boundary overlap group, the flight distance between adjacent UAVs is not closely related to the boundary overlap degree, which is displayed as a low-weight flight distance overlap non-close signal. If the low-weight flight distance overlap analysis value is less than or equal to the low-weight flight distance overlap analysis threshold, it indicates that within the low-degree boundary overlap group, the flight distance between adjacent UAVs is closely related to the boundary overlap degree, which shows a low-weight flight distance overlap signal. If the correlation evaluation between the flight distance and boundary overlap between adjacent UAVs is performed in the high-risk collision group and the low-degree boundary overlap group respectively, and both the high-risk flight distance overlap signal and the low-degree boundary overlap signal are displayed, it indicates that the flight distance is the main training element when adjacent UAVs are conducting collaborative mapping. If the correlation assessment between the flight distance and boundary overlap between adjacent UAVs is performed in the high-risk collision group and the low-degree boundary overlap group respectively, and the results show a high-risk flight distance overlap signal and a low-degree boundary overlap signal respectively, it indicates that the flight distance is not a major training element when adjacent UAVs are conducting collaborative mapping. The specific solution in this embodiment is as follows: A dual-drone collision assessment is performed on the drones within the high-overlap analysis group to identify high-risk collision groups. This helps to reflect weaknesses in the drones' collaborative control, particularly in flight spacing control and obstacle avoidance capabilities. By strengthening the flight spacing between drones, obstacles can be avoided promptly and effectively. Within the high-risk collision group, a correlation assessment is performed between the flight spacing between adjacent drones and the high-risk collision level. Simultaneously, within the low-degree boundary overlap group, a correlation assessment is performed between the flight spacing between adjacent drones and the boundary overlap degree. This determines whether flight spacing is a primary training element when adjacent drones are conducting collaborative mapping. When flight spacing is a primary training element, the training strategy can focus on how to accurately control flight spacing. Targeted training can significantly improve the drones' ability to control flight spacing during collaborative flight. If flight spacing is not a primary training element, the training strategy can focus more on other aspects of collaborative control.

[0029] Example 3

[0030] like Figures 1-2 As shown, this embodiment of the invention provides a collaborative control training system for UAV swarms in various scenarios, and further includes: Step 4: After determining flight spacing as the main training element, the high-risk collision group and the low-degree boundary overlap group are screened for overlap with the same UAV to obtain the target UAV and determine the flight spacing range of the target UAV. In some embodiments, the adjacent UAVs in all high-risk collision groups and low-degree boundary overlap groups are compared for overlap. The high-risk collision groups and low-degree boundary overlap groups with the same UAVs are merged into a high-wind-low-weight control group, and the UAVs with the same UAVs are marked as target UAVs. To help understand this, if the boundary overlap analysis group of the combination of eastward and southward drones is a high-risk collision group, and the boundary overlap analysis group of the combination of eastward and northward drones is a low-degree boundary overlap group, then the high-risk collision group and the low-degree boundary overlap group that simultaneously contain eastward drones will be merged into a high-wind-low-weight control group, that is, the eastward drones are the target drones. For example, for the low-degree boundary overlap group where the target UAV is located, the boundary cell overlap length ratio between the target UAV and the adjacent UAV in each simulated mapping training period is extracted and compared. The low-degree overlap flight distance ratio corresponding to the largest boundary cell overlap length ratio is selected and multiplied with the set flight track distance to calculate the target flight distance lower limit value. For example, for the high-risk collision group where the target UAV is located, the time unit flight distance collision value between the target UAV and the adjacent UAV in each simulated mapping training period is extracted and compared. The high-wind overlap flight distance ratio corresponding to the smallest time unit flight distance collision value is selected and multiplied with the set flight track distance to calculate the target flight distance upper limit value. Based on the lower and upper limits of the target flight distance, the range of unmanned flight distances of the target is determined.

[0031] The specific scheme of this embodiment is as follows: After determining the flight spacing as the main training element, the high-risk collision group and the low-degree boundary overlap group are screened for overlap with the same UAV to obtain the target UAV. It can accurately locate special individuals that have both high collision risk and low-degree boundary overlap in cooperative flight, and determine the flight spacing range of the target UAV. Moreover, since the target UAV is the connection point between the high-risk collision group and the low-degree boundary overlap group, it can improve the cooperative ability of the target UAV and the adjacent UAV, improve the flight order in the local area, reduce the collision risk, improve the cooperative performance of the entire cluster, and improve the accuracy of the edge overlap of high-rise building mapping.

[0032] The foregoing has provided a detailed description of one embodiment of the present invention, but this description is merely a preferred embodiment and should not be construed as limiting the scope of the invention. All equivalent variations and modifications made within the scope of the present invention should still fall within the scope of the present invention.

Claims

1. A collaborative control training system for UAV swarms in multiple scenarios, characterized in that: Includes the following modules: Overlapping height and low analysis module: During the training process of simulating high-rise building mapping for drones in the drone swarm, the mapping images corresponding to each drone's flight path are acquired, and the boundary overlap of the mapping images corresponding to adjacent drone flight paths is analyzed to filter out low-degree boundary overlap groups and high-degree boundary overlap groups. High-overlap collision analysis module: assesses the degree of collision between two drones within the high-overlap analysis group and identifies high-risk collision groups; Training Element Analysis Module: In the high-risk collision group, the correlation between the flight distance between adjacent UAVs and the degree of high-risk collision is evaluated. At the same time, in the low-degree boundary overlap group, the correlation between the flight distance between adjacent UAVs and the degree of boundary overlap is evaluated to determine whether the flight distance is a major training element when adjacent UAVs are conducting collaborative mapping. Training Element Control Module: After determining flight spacing as the main training element, the module performs overlap screening with UAVs on high-risk collision groups and low-degree boundary overlap groups to obtain target UAVs and determine the flight spacing range of the target UAVs.

2. The UAV swarm collaborative control training system for multiple scenarios according to claim 1, characterized in that: The process of performing boundary overlap analysis on the mapping images corresponding to adjacent UAV flight paths is as follows: The drones in the drone swarm are combined to form multiple boundary overlap analysis groups. Within each boundary overlap analysis group, based on the same floor of the high-rise building, the mapping images of each drone are combined into multiple mapping image analysis pairs. Within each mapping and imaging analysis pair, the common imaging boundary of the two mapping and imaging pairs is extracted as the common basic analysis boundary. The overlap is compared and the overlap length is extracted. The ratio of the overlap length to the actual building floor edge length is calculated to obtain the boundary unit overlap length ratio. The mean of the overlap length ratio of the boundary units for each mapping and imaging analysis pair is summed to obtain the boundary overlap analysis value; If the boundary overlap analysis value is greater than or equal to the boundary overlap analysis threshold, the corresponding boundary overlap analysis group is marked as a high overlap analysis group. If the boundary overlap analysis value is less than the boundary overlap analysis threshold, the corresponding boundary overlap analysis group is marked as a low-degree overlap analysis group.

3. The UAV swarm collaborative control training system for multiple scenarios according to claim 1, characterized in that: The process for assessing the severity of a two-engine collision is as follows: The simulation surveying training cycle is divided into several simulation surveying training periods. The flight tracks of two UAVs in the same simulation surveying training period when they share the same basic analysis boundary on the same floor of a high-rise building are extracted and used as the same flight tracks. The simulated surveying and mapping training period is divided into several simulated surveying and mapping training time points. The flight coordinates of the UAV on the same flight track at each simulated surveying and mapping training time point are obtained as the same flight coordinates. The distance between the corresponding same flight coordinates of two UAVs in the high overlap analysis group is calculated as the unit aircraft spacing. Simulated surveying training points where the distance between aircraft units is less than or equal to the preset distance between aircraft units are considered high-risk collision points.

4. The UAV swarm collaborative control training system for multiple scenarios according to claim 3, characterized in that: The process for identifying high-risk collision groups is as follows: The proportion of high-risk collision points to the total number of all simulated mapping training points is used as the high-risk collision point ratio. The average value of the distance between aircraft in the unit corresponding to all simulated surveying and mapping training points is calculated to obtain the distance-collision analysis value. The high-risk collision analysis value is obtained by calculating the ratio of the number of high-risk collision times to the distance collision analysis value. If the high-risk collision analysis value is greater than the high-risk collision analysis threshold, it is marked as a high-risk collision group.

5. A drone swarm collaborative control training system for multiple scenarios according to claim 1, characterized in that: The process of analyzing the flight distance between adjacent UAVs and the degree of high-risk collision within the high-risk collision group is as follows: Extract the co-measured flight coordinates of the two UAVs at each simulated surveying and mapping training point during the simulated surveying and mapping training period, calculate and process them using the Euclidean distance formula, and calculate the ratio with the set flight track distance to obtain the high wind overlap flight distance ratio; Extract the proportion of high-risk collision points in each simulated mapping training period to the total number of all simulated mapping training periods to obtain the high-risk collision point ratio for each period; The distance between aircraft in each unit corresponding to the high-risk collision point in each simulated mapping training period is obtained and the average value is calculated to obtain the unit distance collision value for that period. The high-risk collision time point ratio of the time period is calculated by comparing it with the distance collision value of the time period unit to obtain the high-wind collision value of the time period. The high-wind collision values ​​of adjacent simulated surveying and mapping training time periods are then subtracted to obtain the high-wind overlap difference value of adjacent time periods.

6. The UAV swarm collaborative control training system for multiple scenarios according to claim 5, characterized in that: The process for assessing the flight distance between adjacent drones and the degree of high-risk collision within the high-risk collision group is as follows: The difference between the high wind overlap flight distance ratios corresponding to adjacent simulated surveying and mapping training periods is obtained by subtracting the high wind unit flight distance difference between adjacent segments. The high wind overlap difference between adjacent time periods under the same adjacent simulated mapping training period is calculated by the ratio of the high wind unit flight distance difference between adjacent segments to obtain the high wind flight weight analysis value. After calculating the standard deviation and summation mean of multiple high-wind flight overlap analysis values, the coefficient of variation is used for calculation to obtain the high-wind flight distance overlap analysis value. If it is less than or equal to the high-wind flight distance overlap analysis threshold, it is displayed as a high-wind flight distance close overlap signal.

7. The UAV swarm collaborative control training system for multiple scenarios according to claim 1, characterized in that: The analysis process within the low overlap analysis group is as follows: Extract the mapping images of each UAV during the same simulated mapping training period when they share the basic analysis boundary on the same floor of the high-rise building. Extract the common basic analysis boundary of the mapping images of the two UAVs and obtain the overlap length ratio of the boundary cells; Extract the same flight coordinates of the two UAVs at each simulated surveying and mapping training point during the simulated surveying and mapping training period, and take the same flight coordinates at the same simulated surveying and mapping training point as a coordinate comparison group to obtain multiple coordinate comparison groups. The Euclidean distance formula is used to calculate and process the same flight coordinates within multiple coordinate comparison groups, and the ratio is calculated with the set flight track distance to obtain the low-overlap flight distance ratio.

8. The UAV swarm collaborative control training system for multiple scenarios according to claim 7, characterized in that: The evaluation process within the low-overlap analysis group is as follows: Within the simulated mapping training cycle, the difference between the overlap length ratios of boundary units corresponding to adjacent simulated mapping training periods is obtained. The difference between the low-overlapping flight distance ratios of adjacent simulated mapping training periods is obtained by subtracting the flight distance difference of low-overlapping units in adjacent segments. The low-overlap difference between adjacent time periods under the same adjacent simulated mapping training period is calculated by the ratio of the flight distance difference of low-weight units in adjacent segments to obtain the low-weight analysis value. After calculating the standard deviation and the mean of all low-gravity flight weight analysis values, the values ​​are recalculated using the coefficient of variation formula to obtain the low-gravity flight distance overlap analysis values. If the values ​​are less than or equal to the low-gravity flight distance overlap analysis threshold, they are displayed as a low-gravity flight distance close overlap signal.

9. A drone swarm collaborative control training system for multiple scenarios according to claim 1, characterized in that: The analysis process for determining whether flight spacing is a primary training element is as follows: If, when evaluating the correlation between the flight distance and boundary overlap between adjacent UAVs in the high-risk collision group and the low-degree boundary overlap group, respectively, both signals of tight flight distance overlap under high-risk and low-degree boundary overlap are displayed, it indicates that the flight distance is the main training element when adjacent UAVs conduct collaborative mapping.

10. A drone swarm collaborative control training system for multiple scenarios according to claim 1, characterized in that: The process of acquiring the target drone and determining its flight distance range is as follows: All high-risk collision groups and adjacent UAVs in low-degree boundary overlap groups are compared for overlap. High-risk collision groups and low-degree boundary overlap groups with the same UAVs are merged into a high-wind low-weight control group, and the same UAVs are marked as target UAVs. Within the low-degree boundary overlap group, the boundary unit overlap length ratio between the target UAV and the adjacent UAV in each simulated mapping training period is extracted and compared. The low-degree overlap flight distance ratio corresponding to the largest boundary unit overlap length ratio is selected and multiplied with the set flight track distance to calculate the target flight distance lower limit. In the high-risk collision group, the collision distance values ​​of the target UAV and adjacent UAVs in each time period of the simulated mapping training are extracted and compared. The high-wind overlap distance ratio corresponding to the smallest time period unit collision distance value is selected and multiplied with the set flight track distance to calculate the upper limit of the target distance. The flight distance range is determined based on the lower limit of the target distance and the upper limit of the target distance.

Images