Unmanned aerial vehicle group cooperation path planning method and system, product and medium

Abstract

The invention discloses an unmanned aerial vehicle group collaborative path planning method and system, a product and a medium, and relates to the field of non-electrical variable control or adjustment systems.The method comprises the steps that environmental data such as terrain height, obstacle distribution and dynamic targets in an unmanned aerial vehicle detection range are collected and converted into a grid unit space occupation state, a search space is divided into three-dimensional grids, and the three-dimensional grids are divided into three-dimensional data; and calculating an unexplored probability value of each unit, extracting time dimension change data of the value and gradient information between the units, identifying an environment dynamic change area, determining boundary features and generating path node data. Node data is compressed into lightweight path information containing coordinates, priority and time sequence constraints, and a track sequence is generated and distributed by combining the position of the unmanned aerial vehicle. By implementing the method, the detection efficiency of the unmanned aerial vehicle can be improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of non-electric variable control or regulation system, and particularly relates to a method, system, product and medium for path planning of unmanned aerial vehicle (UAV) group cooperation. BACKGROUND

[0002] With the increasing demand for large-scale environmental perception in the fields of natural disaster rescue and urban security monitoring, it has become an important development trend to use multiple UAVs to cooperatively complete complex environmental detection tasks. UAV group cooperative detection can significantly improve the efficiency and coverage of environmental information acquisition, and has important significance for improving the response capability to emergencies.

[0003] In the prior art, a path planning method based on preset waypoints is mainly used for UAV group cooperative detection. This scheme first divides multiple sub-regions according to a task region, and assigns a fixed sequence of detection waypoints to each UAV. Each UAV completes the regional detection task according to the preset waypoints, and transmits the collected environmental data to the ground control station for processing and analysis through a communication network. The ground control station determines whether manual intervention is needed to adjust the waypoint setting based on the received data.

[0004] However, in a complex dynamic environment, the environmental features such as terrain and obstacles often show the characteristics of continuous change. The preset waypoint method needs frequent manual path adjustment, and the static allocation of waypoints makes it difficult for the UAV to flexibly adjust the detection strategy according to the real-time environmental changes. At the same time, the transmission and processing of environmental data also causes system response delay, affecting the detection efficiency. SUMMARY

[0005] The present application provides a method, system, product and medium for path planning of UAV group cooperation, which is used to improve the detection efficiency of UAV.

[0006] In a first aspect, the present application provides a method for cooperative path planning of a UAV group, which is applied to a cooperative path planning system of the UAV group. The method comprises: collecting environmental data within a detection range of each UAV, determining a spatial occupancy state of a plurality of grid cells in a probability map according to the environmental data, wherein the environmental data comprises terrain height data, obstacle distribution data and dynamic target data; dividing a search space into three-dimensional grid cells, calculating an unexplored probability value of each grid cell according to the spatial occupancy state, wherein the unexplored probability value represents a possibility of an unsearched area in each grid cell; extracting variation data of the unexplored probability value of the grid cell in a time dimension, calculating a probability value difference of the unexplored probability value of the grid cell at adjacent time instants, identifying a dynamic change region of the environment according to the variation data; extracting gradient information of the unexplored probability value between the grid cells, determining boundary features of the environmental change region according to the probability value difference and the gradient information, and generating path node data based on the boundary features; compressing the path node data into lightweight path information, wherein the lightweight path information comprises node coordinates, node priority calculated based on a probability distribution, and a timing constraint relationship determined based on an environmental change trend; generating a flight path sequence satisfying the timing constraint based on the lightweight path information and current position information of each UAV, and distributing the flight path sequence to the corresponding UAV.

[0007] In the above embodiment, the unexplored probability value variation data and gradient information of the grid cell are extracted based on the environmental data, the dynamic change region of the environment is identified and the boundary features are determined. The generated lightweight path information contains node priority and timing constraint relationship, which guides the UAV group to adaptively adjust the detection strategy according to the environmental change trend. The UAV can dynamically plan the flight path according to the real-time environmental features, eliminating the delay of manual path adjustment and improving the response speed and detection efficiency of the system to environmental changes.

[0008] In combination with some embodiments of the first aspect, in some embodiments, the steps of collecting the environmental data within the detection range of each UAV, converting the environmental data into the spatial occupancy state of the grid cell, and establishing the probability map according to the spatial occupancy state specifically comprise: receiving ranging data and image data collected by a laser radar and a camera carried by each UAV; performing noise filtering on the ranging data to generate three-dimensional point cloud data; performing coordinate alignment and splicing on the three-dimensional point cloud data, extracting ground points and obstacle points in the point cloud data to obtain the terrain height data and the obstacle distribution data; comparing the position difference of targets in images at different time instants and marking moving targets in the images to obtain the dynamic target data; mapping the terrain height data, the obstacle distribution data and the dynamic target data to a three-dimensional grid map to obtain the spatial occupancy state of each grid cell; and generating the probability map according to the spatial occupancy state of each grid cell.

[0009] In the above embodiment, the laser radar ranging data is filtered and a three-dimensional point cloud is generated, the moving target is identified in combination with the image data, and a complete environment representation model is established. The terrain, obstacles and dynamic target data are uniformly mapped into the grid map, and the fusion of multi-source environment information is realized. The establishment of the probability map provides an accurate environment structure description for subsequent path planning, and enhances the perception and understanding ability of the UAV group to the complex environment.

[0010] In combination with some embodiments of the first aspect, in some embodiments, the step of extracting the change data of the unexplored probability value of the grid cell in the time dimension and calculating the probability value difference of the unexplored probability value of the grid cell at adjacent time points, specifically comprises: collecting and recording the sequence of the unexplored probability value of each grid cell within a set time window, and calculating the difference of the unexplored probability value of the grid cell at adjacent sampling time points within the time window; screening out the grid cells whose probability value changes exceed a preset change threshold, and counting the change frequency of each significant change grid cell within the time window; clustering the grid cells based on spatial proximity and change pattern similarity, extracting the spatial range and change characteristics of each cluster region, and outputting the environmental dynamic change region marked with the change frequency.

[0011] In the above embodiment, the probability value sequence is recorded within the time window and the difference is calculated, and the grid cells with significant changes are screened out. The clustering is performed based on spatial proximity and change pattern similarity, and the spatiotemporal characteristics of the environmental change region are extracted. The output change frequency quantitatively describes the dynamic characteristics of the environment, provides a quantitative basis for the environment change for path planning, and optimizes the detection strategy of the UAV group to the dynamic region.

[0012] In combination with some embodiments of the first aspect, in some embodiments, the step of extracting the gradient information of the unexplored probability value between the grid cells, determining the boundary characteristics of the environmental change region according to the probability value difference and the gradient information, and generating the path node data based on the boundary characteristics, specifically comprises: calculating the gradient of the unexplored probability value between adjacent grid cells, extracting the grid cells with sudden change of the probability value according to the gradient size; connecting the spatially adjacent high-gradient grid cells to construct a boundary contour, analyzing the characteristics of the boundary contour and extracting key nodes, and calculating the environmental change intensity at each key node; generating a priority according to the node position and the change intensity, and outputting the path node data containing the node position and the priority.

[0013] In the above embodiment, the probability value gradient is used to extract the boundary characteristics of the environment, the continuous contour of the high-gradient cell is constructed, the environmental change intensity of the key node is calculated, and the path priority is generated. Based on the quantitative analysis of the gradient information and the change intensity, an automatic mapping mechanism from the environmental characteristics to the path nodes is formed, which reduces the workload of manual path point setting and improves the sensitivity and adaptability of the path planning to the environmental change.

[0014] In some embodiments of the first aspect, in the step of generating a flight path sequence meeting the timing constraints based on the lightweight path information and the current position information of each unmanned aerial vehicle, and distributing the flight path sequence to the corresponding unmanned aerial vehicle, the step specifically comprises: obtaining real-time position and motion state parameters of each unmanned aerial vehicle, calculating the shortest sailing time of each unmanned aerial vehicle to each path node; assigning exploration tasks to the unmanned aerial vehicle based on the sailing time and node priority, generating an initial flight path meeting the task allocation, verifying the feasibility of the flight path according to the performance parameters of the unmanned aerial vehicle; optimizing the flight path segment in the initial flight path that does not meet the constraints, establishing a timing relationship of the flight path of multiple unmanned aerial vehicles and generating a flight path sequence, and distributing the flight path sequence to the corresponding unmanned aerial vehicle.

[0015] In the above embodiment, the sailing time between nodes is calculated based on the real-time state of the unmanned aerial vehicle, the task allocation is combined with the node priority, and the feasibility of the flight path is verified, and the flight path sequence is optimized under the timing constraints of multiple machines. Through the unified processing of flight path planning and task allocation, a collaborative planning framework is established that takes into account the task efficiency and the performance of the machine, effectively improving the overall detection efficiency and task completion quality of the unmanned aerial vehicle group.

[0016] In some embodiments of the first aspect, after the step of generating a flight path sequence meeting the timing constraints based on the lightweight path information and the current position information of each unmanned aerial vehicle, and distributing the flight path sequence to the corresponding unmanned aerial vehicle, the method further comprises: obtaining dynamic obstacle feature information in the environment data, updating the obstacle avoidance strategy of the flight path sequence according to the feature information, and obtaining an adjusted flight path sequence; generating an optimized flight path sequence based on the obstacle avoidance strategy, judging the execution state of the optimized flight path sequence, and realizing task control in response to the execution state, and distributing the adjusted flight path sequence to the corresponding unmanned aerial vehicle for execution.

[0017] In the above embodiment, the dynamic obstacle features and the obstacle avoidance strategy are updated in real time, the flight path sequence is optimized and the execution state is monitored, and the dynamic adjustment of the task process is realized. A closed-loop control mechanism is established from environment perception to flight path optimization, ensuring the continuous detection capability of the unmanned aerial vehicle group in dynamic environment and the stability of task execution.

[0018] In some embodiments of the first aspect, after the step of generating a flight path sequence meeting the timing constraints based on the lightweight path information and the current position information of each unmanned aerial vehicle, and distributing the flight path sequence to the corresponding unmanned aerial vehicle, the method further comprises: obtaining the remaining power state parameters of each unmanned aerial vehicle, calculating the available flight time corresponding to the remaining power state parameters; determining the task allocation weight according to the available flight time, adjusting the flight parameters of the unmanned aerial vehicle according to the task allocation weight, and generating a return path planning meeting the power constraint.

[0019] In the above embodiments, the available flight time is calculated based on the residual power state parameter, the task allocation weight and the flight parameter are adjusted according to the flight time, and the return path meeting the power constraint is generated. An active management mechanism from power monitoring to task adjustment is established, the power constraint is integrated into the task planning process, the continuous operation capability and safety of the UAV group in the long-time detection task are ensured, and the task interruption or equipment loss caused by power depletion is avoided.

[0020] In a second aspect, the embodiments of the present application provide a UAV group cooperative path planning system, comprising: one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program codes, the computer program codes comprise computer instructions, and the one or more processors invoke the computer instructions to enable the UAV group cooperative path planning system to perform the method described in the first aspect and any possible implementation manner of the first aspect.

[0021] In a third aspect, the embodiments of the present application provide a computer program product comprising instructions, which, when executed on a UAV group cooperative path planning system, enable the UAV group cooperative path planning system to perform the method described in the first aspect and any possible implementation manner of the first aspect.

[0022] In a fourth aspect, the embodiments of the present application provide a computer readable storage medium comprising instructions, which, when executed on a UAV group cooperative path planning system, enable the UAV group cooperative path planning system to perform the method described in the first aspect and any possible implementation manner of the first aspect.

[0023] It can be understood that the UAV group cooperative path planning system provided in the second aspect, the computer program product provided in the third aspect and the computer storage medium provided in the fourth aspect are all used to execute the method provided in the embodiments of the present application. Therefore, the beneficial effects that can be achieved are referred to the beneficial effects in the corresponding method, which will not be described here.

[0024] The one or more technical solutions provided in the embodiments of the present application have at least the following technical effects or advantages: 1. In the embodiments of the present application, the probability map is constructed based on environmental data, the unexplored probability value change data and gradient information of the grid element are extracted, the dynamic change region of the environment is identified and the boundary feature is determined. The generated lightweight path information contains node priority and timing constraint relationship, which guides the UAV group to adaptively adjust the detection strategy according to the environmental change trend. The UAV can dynamically plan the flight path according to the real-time environmental characteristics, eliminates the delay of manual path adjustment, and improves the response speed and detection efficiency of the system to environmental changes.

[0025] 2、The application filters noise of laser radar ranging data and generates a three-dimensional point cloud, identifies moving targets in combination with image data, and establishes a complete environment representation model. The terrain, obstacles and dynamic target data are uniformly mapped into a grid map, realizing the fusion of multi-source environment information. The establishment of the probability map provides an accurate environment structure description for subsequent path planning, enhancing the perception and understanding ability of the UAV group to complex environments.

[0026] 3、The application records a probability value sequence within a time window and calculates a difference value to screen out grid cells with significant changes. Based on spatial proximity and change pattern similarity, clustering is performed to extract the spatio-temporal features of the environment change area. The output change frequency quantitatively describes the dynamic characteristics of the environment, providing a quantitative basis for environmental changes for path planning and optimizing the detection strategy of the UAV group to dynamic areas. BRIEF DESCRIPTION OF DRAWINGS

[0027] Figure 1 is a flowchart of a UAV group cooperative path planning method in an embodiment of the application; Figure 2 is another flowchart of a UAV group cooperative path planning method in an embodiment of the application; Figure 3 is a schematic diagram of an entity device structure of a UAV group cooperative path planning system in an embodiment of the application. DETAILED DESCRIPTION

[0028] The terms used in the following embodiments of the application are only for the purpose of describing specific embodiments and are not intended to be limiting on the application. As used in the specification, the singular forms "a," "an" and "the" are intended to include the plural forms as well, unless the context clearly indicates otherwise. It will be further understood that the terms "and / or" used in the application, means any or all possible combinations of one or more of the listed items.

[0029] Hereinafter, the terms "first" and "second" are only for the purpose of description, and cannot be understood as implying or suggesting relative importance or implicitly indicating the number of indicated technical features. Therefore, the features defined with "first" and "second" can explicitly or implicitly include one or more of the features, and in the description of the embodiments of the application, the meaning of "multiple" is two or more, unless otherwise specified.

[0030] To facilitate understanding, the application scenarios of the embodiments of the application are introduced as follows.

[0031] In large-scale natural disaster rescue operations, search and rescue teams need to quickly obtain dynamic information of the disaster area. The traditional single unmanned aerial vehicle reconnaissance method has limited coverage and cannot respond to complex and changing terrain environment in a timely manner. For example, in a landslide site, areas with a risk of continuous collapse need to be monitored, and possible trapped personnel also need to be searched. The terrain height, vegetation coverage and weather conditions are constantly changing, bringing great challenges to search and rescue planning. At this time, a technical solution is needed that can coordinate multiple unmanned aerial vehicles, adapt to environmental changes and optimize search paths in real time. The key problem is how to let the unmanned aerial vehicle group dynamically adjust the search strategy based on real-time environmental perception data and maximize search efficiency within the limited endurance time.

[0032] The existing unmanned aerial vehicle group coordination technology mainly adopts the preset route method. For example, in a forest fire monitoring task, multiple unmanned aerial vehicles are assigned to different areas to perform fixed route cruising tasks. Each unmanned aerial vehicle flies independently and can only obtain local environmental information. When new fire is found, manual intervention is needed to re-plan the route. This method has obvious defects: first, the preset route cannot adapt to the dynamically changing environment, such as changes in no-fly areas caused by the spread of fire; second, there is a lack of information sharing and task coordination mechanism among unmanned aerial vehicles, leading to repeated coverage or missed detection; third, resource constraints such as battery endurance are not considered, which can easily cause task interruption. These problems seriously affect the search efficiency and task completion quality of the unmanned aerial vehicle group.

[0033] After adopting the path planning method of the present application, in a search and rescue task in an earthquake disaster area, multiple unmanned aerial vehicles can work efficiently. The system first obtains environmental data through laser radar and camera and establishes an accurate three-dimensional grid probability map. When a unmanned aerial vehicle finds new signs of building collapse, this information will be updated to the probability map in real time. The system automatically calculates the boundary features of the environmental change area and generates new path nodes. Other unmanned aerial vehicles will automatically adjust the route according to the node priority and their own position to focus on high-risk areas. At the same time, the system will consider the remaining power of each unmanned aerial vehicle to ensure its safe return. For example, when new displacement occurs in area A, the system will prioritize the closest and most power-efficient unmanned aerial vehicle to detect, while other unmanned aerial vehicles continue to monitor the original area. This intelligent coordination method significantly improves search efficiency and avoids delays caused by manual intervention.

[0034] After adopting the path planning method of the application, multiple unmanned aerial vehicles can work efficiently in a search and rescue task in an earthquake disaster area. The system first acquires environmental data through a laser radar and a camera, and establishes an accurate three-dimensional grid probability map. When a building is found to have new signs of collapse by an unmanned aerial vehicle, this information will be updated in real time to the probability map. The system automatically calculates the boundary features of the changed area of the environment, and generates new path nodes. Other unmanned aerial vehicles will automatically adjust the flight path according to the node priority and the position of the unmanned aerial vehicle itself, and focus on the high-risk area. At the same time, the system will consider the remaining power of each unmanned aerial vehicle to ensure that it can safely return. For example, when new displacement occurs in area A, the system will preferentially dispatch the nearest unmanned aerial vehicle with sufficient power to detect, while other unmanned aerial vehicles continue to monitor the original area. This intelligent collaborative method significantly improves the search and rescue efficiency and avoids the delay of manual intervention.

[0035] For ease of understanding, the method provided by the embodiment is described in the following flow. Please refer to Figure 1 , which is a flowchart of the unmanned aerial vehicle group cooperative path planning method in the embodiment of the application.

[0036] S101, collect environmental data in the detection range of each unmanned aerial vehicle, and determine the space occupation state of a plurality of grid units in the probability map according to the environmental data, the environmental data including terrain height data, obstacle distribution data and dynamic target data.

[0037] Among them, the environmental data represents the surrounding environment information obtained by the unmanned aerial vehicle in the detection process, including the digital representation of elements such as terrain, obstacles and dynamic targets; the space occupation state is a state description of whether a specific grid unit is occupied by an environmental element, usually represented by a value between 0 and 1 to represent the occupation degree; the probability map represents a spatial distribution map that represents the environmental data in the form of probability, used to quantitatively describe the uncertainty of environmental characteristics. The terrain height data is the elevation information of the ground undulation; the obstacle distribution data represents the spatial position information of the static obstacles; the dynamic target data is the position and speed of the moving object and other time-varying characteristic data.

[0038] This step is performed when the unmanned aerial vehicle starts to perform the detection task, and is used to establish an initial probability model of the environment. Specifically, the system first acquires the original environmental data in the detection range through the sensors carried by each unmanned aerial vehicle, including the ranging data of the laser radar, the image data of the camera, etc. Then the data is converted to a unified three-dimensional space representation through coordinate transformation and data fusion. For each grid unit, the corresponding space occupation probability value is calculated according to the characteristics of the environmental elements contained therein. Finally, based on the occupation state of all grid units, a complete environmental probability map is constructed.

[0039] In some embodiments, the environment data collection and probability map construction can be implemented in various ways: optionally, using a laser radar to scan the environment, obtaining point cloud data, performing noise reduction and registration processing on the point cloud data, extracting ground points and obstacle points, calculating the spatial distribution characteristics of each point, and finally estimating the grid occupancy probability according to the density distribution of the points; optionally, using a binocular camera to obtain environment images, reconstructing a three-dimensional scene through a stereo vision algorithm, identifying target objects in the images, tracking the motion trajectories of dynamic targets, and calculating the grid occupancy probability in combination with the spatial position and motion characteristics of the targets. It can be understood that other sensors and processing methods can also be used to realize environment perception and probability modeling, which are not limited here.

[0040] S102, dividing the search space into three-dimensional grid cells, calculating the unexplored probability value of each grid cell according to the space occupancy state, the unexplored probability value representing the possibility of the existence of unsearched areas in each grid cell.

[0041] Wherein, the search space represents a three-dimensional space region that needs to be detected; the three-dimensional grid cell refers to a cubic cell formed by uniformly dividing the search space; the unexplored probability value represents the possibility of a specific grid cell not being detected, with a value range of 0-1; the unsearched area refers to a space region that has not been effectively detected by the unmanned aerial vehicle.

[0042] This step is executed after the completion of the environment probability map construction, for quantitatively describing the distribution of the to-be-detected region. Specifically, first, the entire task space is divided into three-dimensional grid cells of equal size. For each grid cell, based on its current space occupancy state, in combination with the effective detection range of the detection sensor and the historical detection records, the unexplored probability value of the cell is calculated. The greater the probability value, the more likely it is that the region contains important information that has not been detected.

[0043] In some embodiments, the calculation of the unexplored probability value can be implemented in various ways: optionally, a probability calculation method based on the sensor detection model is established, considering the detection range, resolution and noise characteristics of the sensor, in combination with factors such as detection distance and incident angle, to calculate the effective detection probability of each grid cell, and the unexplored probability value is obtained by subtracting the effective detection probability from 1; optionally, an information entropy-based probability calculation method is used to analyze the information uncertainty of the grid cell, considering the environmental change frequency and detection time interval, evaluating the timeliness of the information, and comprehensively determining the unexplored probability value. It can be understood that other probability calculation methods can also be used to quantify the distribution characteristics of the unexplored region, which are not limited here.

[0044] S103, extracting the change data of the grid cell unexplored probability value in the time dimension, calculating the probability value difference of the grid cell unexplored probability value at adjacent time points, and identifying the dynamic change region of the environment according to the change data.

[0045] wherein the time dimension change data represents a history record sequence of the grid cell unexplored probability value change over time; the probability value difference refers to the numerical difference of the unexplored probability value of the same grid cell at adjacent two time points; the dynamic change region represents the spatial range where the environmental feature changes significantly. The adjacent time points refer to two consecutive data sampling time points; the change data is used to represent the quantitative description of the dynamic characteristics of the environment.

[0046] This step is executed after obtaining the environmental data at consecutive multiple time points, and is used to identify the dynamic change characteristics in the environment. Specifically, the system first establishes a time series database to record the unexplored probability value of each grid cell at different time points. For each grid cell, the probability value difference at adjacent sampling time points is calculated to form a probability value change sequence. By analyzing the amplitude and frequency of the probability value change, the regions where significant changes occur in the environment are identified. These regions may contain important dynamic targets or environmental change information and need to be focused on.

[0047] In some embodiments, the identification of the environmental dynamic change region can be achieved in various ways: optionally, a sliding time window method is used, a fixed size observation window is set, the statistical characteristics of the probability value in the window are calculated, including mean, variance and change rate, the regions where changes are significant are determined according to the threshold judgment of these characteristics, and finally the judgment results of consecutive multiple time windows are fused to obtain stable dynamic region identification results; optionally, a time series pattern analysis method is used, the probability value change sequence is subjected to Fourier transform, the frequency characteristics of the change are extracted, the energy distribution of different frequency components is calculated, and the regions of periodic change and burst change are identified according to the energy distribution characteristics, and finally the complete dynamic change region is determined in combination with the spatial correlation analysis. It can be understood that other time series analysis methods can also be used to identify the environmental dynamic characteristics, which are not limited here.

[0048] S104, extract the unexplored probability value gradient information between the grid cells, determine the boundary characteristics of the environmental change region according to the probability value difference and the gradient information, and generate path node data based on the boundary characteristics.

[0049] wherein the gradient information represents the change rate of the unexplored probability value between adjacent grid cells; the boundary characteristics refer to the spatial profile and morphological characteristics of the dynamic change region; the path node data is used to represent the key path point information when planning the route. The probability value difference refers to the probability change amount at different time points; the boundary characteristics include the position, direction and curvature of the boundary and other geometric characteristics.

[0050] This step is executed after the dynamic change area is identified, for determining the specific range of the change area and generating the path points. Specifically, first, the unexplored probability value gradient between adjacent grid cells in the space is calculated to obtain the spatial variation characteristics of the probability value. In combination with the probability value difference in the time dimension, the spatial distribution law of the environmental change is analyzed. By connecting the grid cells with large probability gradients, the boundary contour of the change area is constructed. Finally, based on the geometric features of the boundary contour, key positions are selected as path nodes for subsequent route planning.

[0051] In some embodiments, the boundary feature extraction and path node generation can be implemented in various ways: optionally, a morphological processing method is used to perform edge detection and contour extraction on the probability gradient image, the continuity of the boundary contour is optimized through morphological operation, the curvature and inflection point features of the contour are calculated, the path nodes are selected at the feature significant positions, and finally the node is optimized for spatial uniformity to ensure reasonable distribution of the path points; optionally, a region growing algorithm is used to start from a seed point with significant probability value change, continuously expand the region range through similarity criteria, extract the complete boundary of the change area, calculate the main direction and branch points of the boundary, determine the path node positions according to the spatial distribution of these feature points, and finally delete redundant nodes and adjust the positions. It can be understood that other boundary detection and feature extraction methods can also be used to generate path nodes, which are not limited here.

[0052] S105, compress the path node data into lightweight path information, the lightweight path information including node coordinates, node priority calculated based on probability distribution, and timing constraint relationship determined based on environmental change trend.

[0053] Wherein, the lightweight path information represents the simplified path description after data compression and optimization; the node coordinates refer to the three-dimensional spatial position information of the path key points; the node priority represents the detection importance of the path nodes; the timing constraint relationship refers to the access order requirement between the path nodes. The probability distribution refers to the spatial distribution characteristics of the unexplored probability value of the grid cell; the environmental change trend represents the evolution law of the dynamic area characteristics over time; the data compression refers to the process of reducing data redundancy while maintaining key information.

[0054] This step is executed after the initial path nodes are generated, for optimizing the data structure of the path description. Specifically, first, the original path nodes are spatially clustered and screened to remove redundant nodes and retain representative key nodes. Then, based on the unexplored probability distribution characteristics of the grid cell where the node is located, the detection priority of each node is calculated. At the same time, the time characteristics of the environmental change are analyzed to determine the access timing constraints between the nodes. Finally, the node coordinates, priority, and timing constraints are integrated into a unified lightweight data structure.

[0055] In some embodiments, the lightweight processing of path information can be achieved in various ways: optionally, a compression method based on spatial features is used, first the distance and direction features between adjacent path nodes are calculated, nodes with close distance and similar direction are merged, then a multi-level grid structure is used for spatial indexing of the nodes, finally the compression ratio is determined based on the environmental complexity of the area where the node is located, and a simplified path description is generated; optionally, a compression method based on importance score is used, first an importance evaluation model of the node is established, considering factors such as unexplored probability value, environmental change frequency and position coverage, the importance score of each node is calculated, then the nodes are sorted and filtered according to the score, finally the spatial distribution of the reserved nodes is optimized to ensure the detection effect of the key area. It can be understood that other data compression and optimization methods can also be used to realize the lightweight representation of path information, which is not limited here.

[0056] S106, based on the lightweight path information and the current position information of each unmanned aerial vehicle, a flight path sequence meeting the timing constraint is generated, and the flight path sequence is distributed to the corresponding unmanned aerial vehicle.

[0057] Wherein, the flight path sequence represents the complete flight path of the unmanned aerial vehicle performing the detection task; the timing constraint refers to the access time requirement of the path node; the current position information represents the real-time spatial coordinates and attitude data of each unmanned aerial vehicle. Flight path distribution refers to sending the planned flight path information to the corresponding unmanned aerial vehicle; meeting the timing constraint means that the flight path planning needs to consider the order requirement of node access.

[0058] This step is executed after the lightweight processing of path information is completed, and is used to generate the specific flight path of the unmanned aerial vehicle. Specifically, first, the real-time position and state information of each unmanned aerial vehicle is obtained, and the node distribution in the lightweight path information is combined to calculate the flight cost of each unmanned aerial vehicle to each path node. Then, based on the node priority and the timing constraint, each unmanned aerial vehicle is assigned a suitable detection task to generate an initial flight path. The generated flight path is verified for feasibility and optimized, to ensure that the unmanned aerial vehicle performance constraint is met. Finally, the optimized flight path sequence is distributed to each unmanned aerial vehicle for execution.

[0059] In some embodiments, the path planning and distribution can be implemented in various ways: optionally, a hierarchical planning method is used, coarse path planning is first performed at a global level based on task priority and timing constraints to determine the approximate task area of each UAV, then the specific path points of the path are refined at a local level considering the motion characteristics and energy constraints of the UAV, and finally a coordination mechanism is established among multiple UAVs to ensure conflict avoidance during path execution; optionally, a dynamic programming method is used, a path cost model considering multiple constraints is first established, including flight distance, time window and energy consumption, etc., then a heuristic search algorithm is used to find the optimal path that meets the constraints, and finally the path parameters are dynamically adjusted according to the environmental changes to realize real-time path optimization. It can be understood that other path planning and task allocation methods can also be used to generate and distribute paths, which are not limited here.

[0060] The method provided by the embodiment will be further described in more detail below. Please refer to Figure 2 , which is another flowchart of the UAV group cooperative path planning method in the embodiment of the present application.

[0061] S201, receiving ranging data and image data collected by the laser radar and camera carried by each UAV; performing noise filtering on the ranging data to generate three-dimensional point cloud data.

[0062] Among them, the laser radar ranging data represents the distance information of the space point obtained by using the laser ranging principle; the image data refers to the two-dimensional image of the environment collected by the camera; the noise filtering refers to removing the abnormal values and interference information in the ranging data; the three-dimensional point cloud data represents a three-dimensional scene representation composed of a large number of spatial scattered points.

[0063] In the environment data collection stage, each UAV scans the environment by emitting laser beams through the carried laser radar, records the time difference from emission to reception of the laser beam to calculate the target distance, and at the same time obtains the angle information of the reflection point, thereby obtaining the three-dimensional coordinates of the space point. Due to the influence of factors such as atmosphere and target surface characteristics, there are noise points in the original data. Through statistical analysis of these data, abnormal points are removed by setting distance threshold and density threshold, and then the data is smoothed by using Gaussian filtering or median filtering method, finally the three-dimensional point cloud expressing the environmental characteristics is generated. Each point in the point cloud contains spatial position and reflection intensity information, which is used for subsequent scene reconstruction.

[0064] S202, performing coordinate alignment and splicing on the three-dimensional point cloud data, extracting ground points and obstacle points in the point cloud data, and obtaining terrain height data and obstacle distribution data.

[0065] Wherein, the coordinate alignment refers to converting the point cloud data acquired at different positions to a unified coordinate system; the point cloud splicing refers to combining multiple point cloud data into a complete scene description; the ground points refer to a set of spatial points belonging to the ground; and the obstacle points refer to a set of spatial points belonging to obstacles.

[0066] When processing the point cloud data collected at multiple positions, first, feature points in the point cloud are extracted, a matching relationship between the feature points is calculated, a rigid body transformation matrix is established to realize accurate alignment between different point clouds. Then, the aligned point cloud data is spliced into a unified data set. The complete point cloud is segmented into ground, a ground plane is fitted using the RANSAC algorithm, a set of ground points is obtained, and terrain height information is extracted. The remaining non-ground points are identified as independent obstacles through clustering analysis, and the position, size and other characteristic parameters of the obstacles are recorded to construct an obstacle distribution map.

[0067] S203, comparing the target position difference in the images at different times and marking the moving targets in the images to obtain dynamic target data; mapping the terrain height data, obstacle distribution data and dynamic target data to a three-dimensional grid map to obtain the space occupation state of each grid cell.

[0068] Wherein, the target position difference refers to the position change of the same target in the images at different times; the dynamic target data represents the position and speed information of the moving object; and the space occupation state refers to the state quantity of whether there is an environmental element in the grid cell.

[0069] Processing the image sequence collected continuously, the motion of the target in the image is tracked through feature matching and optical flow calculation and other methods. The displacement of the target in the image plane is calculated, the plane displacement is converted into actual space displacement combined with the camera parameters to obtain the motion trajectory of the target. The terrain, obstacle and dynamic target data obtained are uniformly mapped to a three-dimensional grid map. For each grid cell, the occupation state value is calculated according to the type and number of environmental elements contained. The occupation state can be binary (occupied / unoccupied) or probabilistic (occupation probability between 0 and 1), which is used to represent the spatial structure characteristics of the environment.

[0070] S204, generating a probability map according to the space occupation state of each grid cell, the environment data including terrain height data, obstacle distribution data and dynamic target data.

[0071] Wherein, the probability map represents a probability distribution model of environmental characteristics; the space occupation state refers to the existence state of environmental elements in the grid cell; the terrain height data represents a numerical description of the ground undulation; the obstacle distribution data refers to the spatial position information of static obstacles; and the dynamic target data represents the space-time characteristics of moving objects.

[0072] When converting the spatial occupancy state of each grid cell into a probability representation, the characteristics of different types of environmental data need to be considered comprehensively. For terrain height data, the terrain probability is calculated according to the continuity of the height value; for obstacle data, the occupancy probability is determined based on the density and distribution range of the obstacles; for dynamic targets, the probability distribution of the influence range is calculated in combination with the motion characteristics of the targets. The probabilities of various environmental characteristics are fused using a Bayesian probability model to generate a unified probability map. For each grid cell, its probability value reflects the confidence of the presence of environmental elements at that location, and the larger the probability value, the more attention needs to be paid to that area.

[0073] S205, dividing the search space into three-dimensional grid cells, calculating the unexplored probability value of each grid cell according to the spatial occupancy state, the unexplored probability value representing the possibility of the existence of unsearched areas in each grid cell.

[0074] Among them, the search space refers to the three-dimensional range of the region that needs to be detected; the grid cell represents the basic cubic cell after spatial division; the unexplored probability value refers to the degree of insufficient detection of a specific location; the unsearched area refers to the spatial area that has not been detected.

[0075] The task area is divided into a regular three-dimensional grid structure, and the grid size is determined according to environmental characteristics and detection accuracy requirements. For each grid cell, the unexplored probability value is calculated based on its current spatial occupancy state. The calculation process takes the following factors into account: the detection range and accuracy characteristics of the sensor, assigning a higher unexplored probability to the area outside the detection range; historical detection records, adjusting the probability value according to the number of detections and time intervals; environmental complexity, increasing the unexplored probability of areas with high obstacle density or frequent dynamic changes. In this way, a complete unexplored probability distribution map is constructed to guide subsequent path planning.

[0076] S206, collecting and recording the unexplored probability value sequence of each grid cell within a set time window, calculating the unexplored probability value difference of the grid cell at adjacent sampling time within the time window.

[0077] Among them, the time window represents the time period of continuous observation; the unexplored probability value sequence refers to the time sequence of the probability value of a specific grid cell; the probability value difference represents the change amount of the probability value at adjacent time; the sampling time refers to the discrete time point of data collection.

[0078] The probability value change of each grid cell is recorded in a fixed size time window at a set sampling frequency. The sampling frequency needs to meet the sampling theorem requirement of environmental change characteristics, ensuring that important dynamic information is not lost. For each grid cell, the probability value in the time window is formed into a time series. The probability value difference between adjacent sampling times is calculated, with the difference calculated by subtracting the previous time probability value from the current time probability value. These differences reflect the change rate of environmental characteristics, and a larger difference indicates that the region has undergone significant changes and needs to be detected. Statistical analysis of the probability value differences of all grid cells is performed to establish a time series feature model of environmental changes.

[0079] S207, screen out grid cells with probability value changes exceeding a preset change threshold, and count the change frequency of each significant change grid cell in the time window.

[0080] Wherein, the preset change threshold represents the standard value for judging the significance of the probability value change; the significant change grid cell refers to the cell whose probability value change amplitude exceeds the threshold; the change frequency represents the number of times the probability value changes significantly; and the time window refers to the time range of the observation data.

[0081] The probability value difference sequence calculated for each grid cell is analyzed, and a uniform change threshold is set as the judgment standard. When the absolute value of the probability value difference at a certain time exceeds the threshold, it is recorded as a significant change. The setting of the change threshold is based on the statistical analysis of environmental characteristics, and is usually 2-3 times the standard deviation of the probability value. For the identified significant change grid cells, the total number of significant changes is counted in the entire time window. The counting process uses a counter to record the number of times the threshold is exceeded, and also records the time stamp of the change. These statistical data reflect the time distribution characteristics of environmental changes, and the higher the change frequency, the more significant the dynamic characteristics of the region.

[0082] S208, cluster the grid cells based on spatial proximity and change pattern similarity, extract the spatial range and change characteristics of each cluster region, and output the environmental dynamic change region marked with change frequency.

[0083] Wherein, the spatial proximity refers to the distance relationship of the grid cells in the spatial position; the change pattern similarity represents the similarity of the probability value change characteristics; the cluster region refers to the set of grid cells with similar change characteristics; and the change frequency represents the time frequency characteristics of the environmental feature changes.

[0084] The grid cells with significant changes are grouped according to spatial positions and change characteristics. Spatial proximity is determined by calculating the Euclidean distance between grid cells and setting a distance threshold. Change pattern similarity is determined by calculating the correlation coefficient or distance measure of the probability value change sequence. Using a density clustering algorithm such as DBSCAN, spatially adjacent and change pattern similar grid cells are aggregated into independent regions. For each cluster region, the boundary coordinates of its spatial range are calculated, and the average change frequency and change amplitude of the grid cells in the region are extracted as characteristic parameters. The final output of the dynamic change region contains a complete description of the location range and change characteristics.

[0085] S209, calculate the unexplored probability value gradient between adjacent grid cells, and extract the grid cells with probability value mutations according to the gradient size.

[0086] Wherein, the unexplored probability value gradient refers to the change rate of the probability value between adjacent grid cells; the gradient size represents the degree of probability value spatial change; the probability value mutation grid cell refers to the grid cell with a larger gradient value.

[0087] The probability value gradient of adjacent grid cells in three-dimensional space is calculated. The first-order difference of the probability value is calculated in each of the six adjacent directions (up, down, left, right, front, and back) of each grid cell to obtain the gradient component in each direction. The central difference format is used to calculate the gradient, that is, the probability value difference between the current grid cell and the adjacent cell is divided by the grid spacing. The gradient components in each direction are combined to obtain the gradient vector, and the modulus of the gradient vector is calculated as the gradient size. A gradient threshold is set, and the grid cells with a gradient size exceeding the threshold are extracted as the probability value mutation positions. These positions usually correspond to the boundaries of environmental features or regions with sharp changes, which need to be considered in path planning.

[0088] S210, connect the spatially adjacent high-gradient grid cells to construct a boundary contour, perform feature analysis on the boundary contour and extract key nodes, and calculate the environmental change intensity at each key node.

[0089] Wherein, the high-gradient grid cell refers to the grid cell with a gradient value exceeding the threshold; the boundary contour represents the spatial boundary line of the environmental change region; the key node refers to a position with significant features on the boundary contour; and the environmental change intensity represents the degree of change of the environmental features at a specific position.

[0090] The high gradient grid cells adjacent in spatial position are connected into complete boundary contours by a boundary tracking algorithm. The tracking process starts from any high gradient cell, and then searches for adjacent high gradient cells in turn according to the eight-neighborhood search method and establishes a connection relationship. Geometric features of the formed boundary contours are calculated, including curvature, corner, and other parameters. Key nodes are extracted at positions with large curvature or obvious corners. For each key node, the gradient value and the amplitude of the probability value change in the neighborhood range are counted, and the environmental change strength is calculated. The change strength calculation comprehensively considers three factors of gradient size, probability value difference, and change duration, and obtains a normalized strength value by using weighted summation.

[0091] S211, generating a priority according to the node position and the change strength, and outputting path node data containing the node position and the priority.

[0092] Among them, the node position refers to the three-dimensional spatial coordinates of the key node; the change strength represents the degree of change of the environmental features at the node; the priority refers to the importance of the detection of the path node; and the path node data contains the complete attribute information of the node.

[0093] The detection priority is determined based on the position and change strength information of the key node. The priority calculation adopts a multi-factor scoring method, including the following calculation steps: first, the coverage score is calculated according to the spatial distribution of the node, and the node with dispersed position obtains a higher score; then, the importance score is calculated based on the change strength, and the larger the strength, the higher the score; finally, the two scores are weighted and summed according to the set weight to obtain the comprehensive priority score of the node. The position coordinates, change strength, and priority score of the node are organized into a unified data structure as the basic data for path planning. The priority score is represented by a normalized value of 0-1, which is convenient for subsequent processing.

[0094] S212, compressing the path node data into lightweight path information, which includes node coordinates, node priority calculated based on probability distribution, and time sequence constraint relationship determined based on environmental change trend.

[0095] Among them, the lightweight path information refers to the optimized and compressed path description data; the node coordinates represent the spatial position of the path point; the node priority refers to the importance of the detection order; and the time sequence constraint relationship represents the time sequence requirement of the node access.

[0096] The original path node data is compressed and structurally optimized to generate a more efficient data representation. First, the nodes are spatially clustered, merging nodes that are close in distance and retaining higher priority nodes as representative points. Then, the node priorities are adjusted based on probability distribution characteristics. The calculation method is as follows: the probability value of the node's location is combined with the node's original priority to obtain an updated priority value. Finally, the time characteristics of environmental changes are analyzed, and access order constraints between nodes are established. The timing constraints are represented by a directed graph structure, and the weights of the edges in the graph reflect the access order between nodes. The optimized node coordinates, priorities, and timing constraints are packaged into a compact data structure for subsequent flight path planning.

[0097] S213, obtain the real-time position and motion state parameters of each unmanned aerial vehicle, and calculate the shortest flight time of each unmanned aerial vehicle to each path node.

[0098] wherein the real-time position refers to the current three-dimensional spatial coordinates of the unmanned aerial vehicle; the motion state parameters include velocity, acceleration, and attitude angle, etc. dynamic characteristics; the shortest flight time represents the minimum time required for the unmanned aerial vehicle to reach the target node from the current position; the path node refers to the key path point in the planned flight path.

[0099] The real-time position information of the unmanned aerial vehicle is obtained through the positioning system carried by the unmanned aerial vehicle, including latitude, longitude and height data, as well as recording the motion parameters such as current speed vector, acceleration and Euler angle. For each unmanned aerial vehicle and each path node, the shortest flight time is calculated. The calculation process is based on the A* algorithm: first, a three-dimensional path planning space considering obstacles is established, and the kinematic constraints of the unmanned aerial vehicle are set; then the heuristic distance estimate from the starting point to the ending point is calculated, and the distance is divided by the maximum cruising speed of the unmanned aerial vehicle to obtain the time estimate value; finally, the shortest path that meets the constraint conditions is obtained through iterative search, and the actual flight time on this path is calculated. The calculation of flight time takes into account the acceleration and deceleration process, and the accurate time value is obtained by solving the kinematic equation.

[0100] S214, based on the flight time and node priority, assign exploration tasks to the unmanned aerial vehicle, generate an initial flight path that meets the task allocation, and verify the feasibility of the flight path according to the performance parameters of the unmanned aerial vehicle.

[0101] wherein the flight time refers to the time required for the unmanned aerial vehicle to complete the path flight; the node priority represents the detection importance of the path point; the initial flight path refers to the preliminary planned flight path; the performance parameters include maximum speed, turning radius, climb rate, etc. flight performance indicators.

[0102] Based on the calculated flight time matrix and node priorities, a task allocation algorithm is used to assign reconnaissance tasks to UAVs. The allocation process is implemented using the Hungarian algorithm: first, a cost matrix is ​​constructed, with each element being a weighted combination of flight time and node priority; then, the optimal matching scheme is found through matrix transformation, determining the set of path nodes responsible for each UAV. Initial tracks are generated for each UAV based on the allocation results. Cubic spline curves are used for track smoothing to ensure path continuity. The generated tracks are then validated for feasibility, checking whether they meet performance constraints such as the UAV's minimum turning radius and maximum climb angle. The validation process involves simulating and calculating characteristic parameters such as track curvature and slope, comparing them with performance limitations to ensure the tracks meet flight requirements.

[0103] S215. Optimize the track segments in the initial track that do not meet the constraints, establish the time sequence relationship of multiple UAV tracks and generate track sequences, and distribute the track sequences to the corresponding UAVs.

[0104] Among them, a track segment refers to a local path segment in a track; a constraint refers to the kinematic and dynamic restrictions imposed on the UAV; a track timing relationship indicates the sequential requirements for the execution of multiple UAV tracks; and a track sequence refers to a complete list of flight path points.

[0105] Flight path segments that do not meet the constraints are optimized and adjusted. The optimization process uses an iterative method: first, flight path segments that violate the constraints are identified, including path segments with excessively small turning radii, excessively large climb angles, or excessive speeds; then, these flight path segments are refitted using Bézier curves, and the control point positions are adjusted to make the curves meet the constraint requirements; finally, the performance indicators of the optimized flight path are verified. A temporal relationship is established for the flight paths of multiple UAVs, and the execution order is calculated based on task priority and spatial location. The establishment of the temporal relationship uses a graph theory method: the flight path sequence is represented as a directed graph, where nodes represent path points, edges represent flight paths, and edge weights contain flight time and priority information. A topological sort is used to determine the execution order that meets the temporal constraints. The generated flight path sequence contains precise position coordinates, timestamps, and speed commands, and is distributed to the flight control systems of each UAV in a standard data format. The distribution process is carried out through a secure communication channel to ensure the reliability of data transmission. After receiving the flight path sequence, each UAV executes the planned path through the flight path tracking controller.

[0106] Following step S215, the following steps are also included: The system acquires dynamic obstacle feature information from environmental data, updates the obstacle avoidance strategy of the trajectory sequence based on the feature information, and obtains the adjusted trajectory sequence.

[0107] In this step, the dynamic obstacle feature information represents attributes such as the position, speed, and motion trend of the moving obstacle; the obstacle avoidance strategy refers to the specific method and path adjustment rule for the UAV to avoid the obstacle; and the adjusted flight path sequence refers to the flight path optimized by avoiding obstacles.

[0108] Feature extraction and motion prediction are performed on the acquired dynamic obstacle data. Feature extraction includes calculating the current position and velocity vector of the obstacle and analyzing historical trajectory data to predict future motion trends. The dynamic window method is used to update the obstacle avoidance strategy: first, the dangerous area of the obstacle is calculated, which takes into account the size and motion speed of the obstacle; then, it is detected whether the existing flight path intersects with the dangerous area; for the flight path segment with collision risk, an obstacle avoidance path is generated based on the artificial potential field method. The potential field construction sets the obstacle as a repulsive source and the target point as an attractive source, and the safe flight path is calculated by superimposing the force field. Path adjustment needs to ensure that the motion constraints of the UAV are met, while maintaining the time sequence relationship of the original flight path.

[0109] Based on the obstacle avoidance strategy, an optimized flight path sequence is generated, the execution state of the optimized flight path sequence is judged, and task control is realized in response to the execution state. The adjusted flight path sequence is distributed to the corresponding UAV for execution.

[0110] In this step, the optimized flight path sequence refers to the complete flight path after obstacle avoidance processing; the execution state represents the progress and abnormal conditions of the flight path execution; and the task control refers to the real-time management and adjustment of the flight task.

[0111] According to the updated obstacle avoidance strategy, a new flight path sequence is generated using a trajectory optimization algorithm. Multiple evaluation indicators are set in the optimization process: path safety, flight time, and energy consumption. The flight path parameters are iteratively optimized by gradient descent method until the evaluation requirements are met. The execution state of the optimized flight path is monitored: the actual position of the UAV is compared with the planned path in real time, and the tracking error is calculated; it is detected whether there are abnormal conditions such as timeout and deviation. Based on the state monitoring result, task control is performed: path re-planning is triggered when the tracking error exceeds the threshold; and emergency handling process is started when abnormal state occurs. Task control instructions are issued to each UAV in real time through the communication network to ensure reliable execution of the collaborative task.

[0112] The remaining power state parameters of each UAV are acquired, and the available flight time corresponding to the remaining power state parameters is calculated.

[0113] In this step, the remaining power state parameter refers to the numerical characteristics of the remaining capacity of the battery; the available flight time represents the flight time that can be maintained under the current load; and the state parameters include real-time data such as voltage, current, and temperature.

[0114] The battery management system collects the power state data of the UAV, including real-time voltage, discharge current and battery temperature. Based on these parameters, a power prediction model is established: first, estimate the remaining power percentage according to the current voltage and discharge curve; then calculate the actual available capacity combined with the load power and environmental temperature; finally, through the energy consumption model, the available flight time is obtained. The calculation process takes into account multiple factors: the influence of flight speed and maneuver on power consumption, the influence of temperature on battery performance, and the reservation of safety margin. Through these calculations, an accurate estimate of the remaining flight time is obtained.

[0115] According to the available flight time, the task allocation weight is determined, and the UAV flight parameters are adjusted according to the task allocation weight to generate a return path planning that meets the power constraint.

[0116] In this step, the task allocation weight refers to the task priority coefficient determined according to the remaining power; the flight parameters include speed, height and other control quantities; the return path refers to the flight path of the UAV safely returning to the starting point.

[0117] Based on the calculated available flight time, a priority evaluation system for task allocation is established. The weight calculation takes into account the following factors: safety margin of remaining power, completion progress of current task, energy demand of return distance. The flight parameters are dynamically adjusted: adjust the cruising speed and climb rate according to the remaining power to optimize energy utilization efficiency. The return path planning uses the A* algorithm with the minimum energy consumption as the optimization goal. The wind field influence is considered in the path generation process, and the flight altitude that is beneficial to energy saving is selected. Through these measures, the UAV can safely return when the power is insufficient, and accidental landing is avoided.

[0118] Next, the UAV group cooperative path planning system in the embodiment of the present application will be described from the perspective of hardware processing. Please refer to Figure 3 , which is a schematic diagram of an entity device structure of the UAV group cooperative path planning system in the embodiment of the present application.

[0119] It should be noted that, Figure 3 The structure of the UAV group cooperative path planning system shown is only an example, and should not impose any limitation on the function and use range of the embodiment of the present application.

[0120] As Figure 3As shown, the UAV swarm cooperative path planning system includes a central processing unit (CPU) 301, which can perform various appropriate actions and processes, such as executing the methods described in the above embodiments, according to programs stored in a read-only memory (ROM) 302 or loaded from a storage section 308 into a random access memory (RAM) 303. Various programs and data required for system operation are also stored in the RAM 303. The CPU 301, the ROM 302, and the RAM 303 are connected to each other through a bus 304. An input / output (I / O) interface 305 is also connected to the bus 304.

[0121] Connected to the I / O interface 305 are an input section 306 including an audio input device, a push button switch, and the like; an output section 307 including a liquid crystal display (LCD), an audio output device, an indicator, and the like; a storage section 308 including a hard disk and the like; and a communication section 309 including a network interface card such as a LAN (Local Area Network) card, a modem, and the like. The communication section 309 performs communication processing via a network such as the Internet. A drive 310 is also connected to the I / O interface 305 as necessary. A removable medium 311 such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, and the like is attached to the drive 310 as necessary, so that a computer program read therefrom is installed into the storage section 308 as necessary.

[0122] In particular, according to embodiments of the present application, the processes described above with reference to the flowcharts can be implemented as a computer software program. For example, embodiments of the present application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing a computer program for executing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network by the communication section 309, and / or installed from the removable medium 311. When the computer program is executed by the central processing unit (CPU) 301, various functions defined in the present application are performed.

[0123] Note that specific examples of computer-readable storage media can include but are not limited to an electrical connection having one or more wires, a portable computer diskette, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or Flash memory), an optical fiber, a portable compact disc read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing. In the present disclosure, computer-readable storage media can be any tangible medium that can contain, or store a program for use by or in connection with an instruction execution system, apparatus, or device.

[0124] The flow charts and block diagrams in the drawings are illustrations of possible architectures, functions, and operations of systems, methods, and computer program products according to various embodiments of the present disclosure. In some embodiments, each block in the flow charts or block diagrams can represent a module, a segment, or a portion of code, which includes one or more executable instructions for implementing the specified logical functions. It should also be noted that in some alternative implementations, the functions noted in the blocks can occur out of the order noted in the figures.

[0125] Specifically, the UAV group cooperative path planning system in the embodiment includes a processor and a memory, and the memory stores a computer program. When the computer program is executed by the processor, the UAV group cooperative path planning method provided in the above embodiment is implemented.

[0126] As another aspect, the present disclosure also provides a computer-readable storage medium. The storage medium can be included in the UAV group cooperative path planning system described in the above embodiments, or can exist independently without being assembled into the UAV group cooperative path planning system. The storage medium carries one or more computer programs. When the one or more computer programs are executed by a processor of the UAV group cooperative path planning system, the UAV group cooperative path planning system implements the UAV group cooperative path planning method provided in the above embodiments.

[0127] The above embodiments are only used to illustrate the technical solutions of the present disclosure, rather than limiting them. Although the present disclosure has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that they can still make modifications to the technical solutions recorded in the foregoing embodiments, or make equivalent replacements to some technical features thereof; and these modifications or replacements do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present disclosure.

[0128] In the above embodiments, the term "when" can be interpreted to mean "if" or "after" or "in response to determining" or "in response to detecting" depending on the context. Similarly, the phrase "on determining" or "if detecting" can be interpreted to mean "if determining" or "in response to determining" or "on detecting" or "in response to detecting" depending on the context.

[0129] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by a computer program instructing the relevant hardware to complete, and the program can be stored in a computer readable storage medium. When the program is executed, it can include the processes of the above-mentioned embodiments. The aforementioned storage medium includes ROM or random storage memory RAM, magnetic disk or optical disk and various storage program codes.

Claims

1. A method for cooperative path planning of a UAV swarm, characterized in that, The method is applied to a UAV group cooperative path planning system, and comprises the following steps: Collecting environmental data within the detection range of each UAV, and determining the spatial occupancy state of a plurality of grid cells in a probability map according to the environmental data, wherein the environmental data comprises terrain height data, obstacle distribution data and dynamic target data; Dividing a search space into three-dimensional grid cells, and calculating an unexplored probability value of each grid cell according to the spatial occupancy state, wherein the unexplored probability value represents the possibility of the existence of an unsearched area in each grid cell; Extracting the change data of the unexplored probability value of the grid cell in the time dimension, calculating the probability value difference of the unexplored probability value of the grid cell at adjacent time instants, and identifying a dynamic change region of the environment according to the change data; Extracting the unexplored probability value gradient information between the grid cells, determining the boundary features of the environmental change region according to the probability value difference and the gradient information, and generating path node data based on the boundary features; Compressing the path node data into lightweight path information, wherein the lightweight path information comprises node coordinates, node priority calculated based on probability distribution, and timing constraint relationship determined based on the environmental change trend; Generating a flight path sequence satisfying the timing constraint based on the lightweight path information and the current position information of each UAV, and distributing the flight path sequence to the corresponding UAV.

2. The method of claim 1, wherein, The step of collecting environmental data within the detection range of each UAV, and determining the spatial occupancy state of a plurality of grid cells in a probability map according to the environmental data, specifically comprises the following steps: Receiving ranging data and image data collected by a laser radar and a camera carried by each UAV; performing noise filtering on the ranging data to generate three-dimensional point cloud data; Performing coordinate alignment and splicing on the three-dimensional point cloud data, extracting ground points and obstacle points in the three-dimensional point cloud data, and obtaining terrain height data and obstacle distribution data; Comparing the position differences of targets in images at different time instants, marking moving targets in the images, and obtaining dynamic target data; and mapping the terrain height data, the obstacle distribution data and the dynamic target data to a three-dimensional grid map to obtain the spatial occupancy state of each grid cell; Generating a probability map according to the spatial occupancy state of each grid cell.

3. The method of claim 1, wherein, The step of extracting the change data of the unexplored probability value of the grid cell in the time dimension, calculating the probability value difference of the unexplored probability value of the grid cell at adjacent time instants, and identifying a dynamic change region of the environment according to the change data, specifically comprises the following steps: Collecting and recording the unexplored probability value sequence of each grid cell within a set time window, and calculating the unexplored probability value difference of the grid cell at adjacent sampling time instants within the time window; Screening out grid cells with a probability value change exceeding a preset change threshold, and counting the change frequency of each significant change grid cell within the time window; Clustering the grid cells based on spatial proximity and change pattern similarity, extracting the spatial range and change features of each cluster region, and outputting an environmental dynamic change region marked with a change frequency.

4. The method of claim 1, wherein, The step of extracting the unexplored probability value gradient information between the grid cells, determining the boundary features of the environmental change region according to the probability value difference and the gradient information, and generating path node data based on the boundary features specifically includes: calculating the unexplored probability value gradient between adjacent grid cells, extracting the grid cells with probability value mutation according to the gradient size; connecting the spatially adjacent high-gradient grid cells to construct a boundary contour, performing feature analysis on the boundary contour and extracting key nodes, and calculating the environmental change intensity at each key node; generating priorities according to the node positions and change intensities, and outputting path node data containing the node positions and priorities.

5. The method of claim 1, wherein, The step of generating a flight path sequence meeting the timing constraints based on the lightweight path information and the current position information of each unmanned aerial vehicle, and distributing the flight path sequence to the corresponding unmanned aerial vehicle specifically includes: obtaining the real-time position and motion state parameters of each unmanned aerial vehicle, and calculating the shortest flight time of each unmanned aerial vehicle to each path node; allocating exploration tasks to the unmanned aerial vehicles based on the flight time and node priority, generating an initial flight path meeting the task allocation, and verifying the feasibility of the flight path according to the performance parameters of the unmanned aerial vehicles; optimizing the flight path segments in the initial flight path that do not meet the constraints, establishing the timing relationship of the flight paths of multiple unmanned aerial vehicles, and generating a flight path sequence, and distributing the flight path sequence to the corresponding unmanned aerial vehicles.

6. The method of claim 1, wherein, After the step of generating a flight path sequence meeting the timing constraints based on the lightweight path information and the current position information of each unmanned aerial vehicle, and distributing the flight path sequence to the corresponding unmanned aerial vehicle, the method further includes: obtaining dynamic obstacle feature information in the environmental data, updating the obstacle avoidance strategy of the flight path sequence according to the feature information, and obtaining an adjusted flight path sequence; generating an optimized flight path sequence based on the obstacle avoidance strategy, judging the execution state of the optimized flight path sequence, and realizing task control in response to the execution state, and distributing the adjusted flight path sequence to the corresponding unmanned aerial vehicles for execution.

7. The method of claim 6, wherein, After the step of generating a flight path sequence meeting the timing constraints based on the lightweight path information and the current position information of each unmanned aerial vehicle, and distributing the flight path sequence to the corresponding unmanned aerial vehicle, the method further includes: obtaining the remaining power state parameters of each unmanned aerial vehicle, and calculating the available flight time corresponding to the remaining power state parameters; determining the task allocation weight according to the available flight time, adjusting the flight parameters of the unmanned aerial vehicles according to the task allocation weight, and generating a return path plan meeting the power constraint.

8. A UAV swarm cooperative path planning system, characterized in that, The unmanned aerial vehicle group cooperative path planning system includes one or more processors and a memory; the memory is coupled to the one or more processors, the memory is used to store computer program code, the computer program code includes computer instructions, and the one or more processors invoke the computer instructions to make the unmanned aerial vehicle group cooperative path planning system execute the method as claimed in any one of claims 1-7.

9. A computer-readable storage medium comprising instructions, characterized in that, When the instructions run on the unmanned aerial vehicle group cooperative path planning system, the unmanned aerial vehicle group cooperative path planning system executes the method as claimed in any one of claims 1-7.

10. A computer program product, characterised in that, When the computer program product runs on the UAV swarm cooperative path planning system, the UAV swarm cooperative path planning system is caused to perform the method according to any one of claims 1-7.

Images