Low-altitude flight path planning method, system, equipment and medium

Abstract

The invention discloses a low-altitude flight path planning method, system and device and a medium. The method comprises the following steps: determining a landing point of an unmanned aerial vehicle according to environmental data of a landing area of the unmanned aerial vehicle; a plurality of passable flight areas distributed in the flight direction of the unmanned aerial vehicle are obtained according to the landing point and the take-off point of the unmanned aerial vehicle; acquiring an obstacle in each passable flight area, and acquiring the time cost, the energy cost and the flight risk when the unmanned aerial vehicle flies between any two adjacent passable flight areas according to the flight space coordinates of the candidate flight points and the obstacle coordinates of the obstacles, the flight point in each passable flight area is determined from the multiple candidate flight points, and the low-altitude flight path corresponding to the unmanned aerial vehicle is generated according to the flight space coordinates of the flight points, the take-off point and the landing point, so that the task execution efficiency of the unmanned aerial vehicle is improved.

Description

[0001] Technical field

[0002] The application relates to the technical field of unmanned aerial vehicle path planning, and particularly relates to a low-altitude flight path planning method, system, device and medium. BACKGROUND

[0003] With the development of the Internet of Things and artificial intelligence technology, the utilization of unmanned aerial vehicle technology in tactical task execution is rapidly increasing due to its advantages in speed, flexibility and cost. However, the path planning of unmanned aerial vehicles also faces many challenges in the face of complex flight environments.

[0004] Current mainstream low-altitude path planning algorithms, such as the traditional A* algorithm and its variants, usually limit the core optimization target to the geometric shortest path or static obstacle avoidance. Such methods simply divide the airspace into binary regions of "passable" and "unpassable", and only consider geometric collision detection with obstacles when planning the path. This simplified model fails to integrate time, energy and safety risk into a unified optimization framework, resulting in the inability to form an optimal flight path, which in turn affects the efficiency and reliability of the unmanned aerial vehicle task execution system. SUMMARY

[0005] To solve the above technical problems, the application discloses a low-altitude flight path planning method, system, device and medium, which is used to improve the efficiency of unmanned aerial vehicle task execution.

[0006] To achieve the above purpose, in a first aspect, the application discloses a low-altitude flight path planning method, comprising:

[0007] obtaining a landing area of an unmanned aerial vehicle and obtaining environmental data of each candidate landing point in the landing area, so as to determine a corresponding landing point of the unmanned aerial vehicle from a plurality of candidate landing points according to the environmental data;

[0008] obtaining a plurality of passable flight areas distributed along the flight direction of the unmanned aerial vehicle by a preset neighborhood search algorithm according to the landing point and the takeoff point of the unmanned aerial vehicle; wherein each passable flight area comprises a plurality of candidate flight points;

[0009] obtaining obstacles in each passable flight area, and obtaining time cost, energy cost and flight risk of the unmanned aerial vehicle when flying between any two adjacent passable flight areas according to the flight space coordinates of the candidate flight points and the obstacle coordinates of the obstacles;

[0010] determining a flight point in each passable flight area from a plurality of candidate flight points according to the flight task of the unmanned aerial vehicle and the time cost, the energy cost and the flight risk.

[0011] The low-altitude flight path of the UAV is generated based on the flight space coordinates of the flight point, the take-off point, and the landing point.

[0012] This invention discloses a low-altitude flight path planning method that achieves comprehensive planning of low-altitude flight paths for unmanned aerial vehicles (UAVs) through multi-dimensional cost integration and dynamic path optimization. First, landing points are screened using environmental data to ensure the safety and terrain adaptability of the landing area, avoiding landing risks caused by single geometric factors. Second, based on the division of traversable flight areas according to flight direction distribution, a neighborhood search algorithm is used to generate a candidate path space that conforms to the UAV's motion constraints, providing a foundation for subsequent multi-dimensional cost calculations. By introducing a joint calculation model of time cost, energy cost, and flight risk, the dynamic relationship between obstacle coordinates and flight space coordinates is quantified into optimizable target parameters, overcoming the limitations of traditional binary obstacle avoidance. Differentiated weights are assigned to different cost factors according to flight mission requirements, achieving dynamic optimization of candidate flight points and ensuring a strong correlation between path planning and mission objectives. Finally, a low-altitude flight path is generated through a sequence of coordinate points, forming a globally optimal solution that balances efficiency, energy consumption, and safety, ensuring the efficiency of UAV mission execution.

[0013] As a preferred example, the step of acquiring the landing area of ​​the UAV and acquiring environmental data of each candidate landing point in the landing area, so as to determine the landing point corresponding to the UAV from a plurality of candidate landing points based on the environmental data, includes:

[0014] The landing area of ​​the UAV is generated based on the target coordinates of the preset target point and the preset area radius. The landing area is then divided into grids to obtain multiple candidate landing nodes corresponding to the landing area and to acquire the digital elevation data of the landing area.

[0015] The elevation value of each candidate landing node and the landing obstacles in the landing area, as well as the landing obstacle coordinates of each obstacle, are extracted based on the digital elevation data.

[0016] For any of the candidate landing nodes:

[0017] Multiple surrounding landing nodes corresponding to the candidate landing node are extracted through a preset neighborhood window, and the slope value corresponding to the candidate landing node is obtained based on the elevation value of the surrounding landing nodes.

[0018] The straight-line distance between the candidate landing node and each landing obstacle is obtained based on the landing obstacle coordinates.

[0019] The clearance region and the clearance area of ​​the candidate landing node are obtained according to the preset multi-neighbor connected domain analysis method.

[0020] The landing point of the UAV is selected from a plurality of candidate landing nodes based on the slope value, the straight-line distance of the obstacle, and the clearance area.

[0021] The above scheme achieves refined selection of landing points through multi-dimensional terrain feature fusion. First, a landing area is generated based on the target coordinates and region radius to ensure coverage of the potentially feasible range. A high-precision terrain model is established through grid division and digital elevation data acquisition. For each candidate node, the slope value is calculated using a neighborhood window to eliminate the impact of steep terrain on landing stability. Straight-line distances are calculated using obstacle coordinates to avoid the risk of collisions with obstacles during landing. Multi-neighbor connected domain analysis is used to quantify the clearance area, ensuring sufficient landing buffer space for the UAV. Finally, a joint evaluation is performed based on the three dimensions of slope, obstacle distance, and clearance area to select landing points that simultaneously meet the requirements of gentle terrain, safe obstacle distance, and adequate clearance area, thus solving the problem of insufficient landing point safety caused by the single-dimensional evaluation of traditional methods.

[0022] As a preferred example, the step of obtaining multiple passable flight areas distributed along the flight direction of the UAV based on the landing point and the UAV's takeoff point using a preset neighborhood search algorithm includes:

[0023] The terrain area where the UAV is located is determined based on the landing point coordinates and the takeoff point coordinates, and the flight search step size of the UAV is determined based on the slope of the terrain area.

[0024] The three-dimensional voxel grid data of the UAV during flight is obtained based on the landing point coordinates, the takeoff point coordinates, and the preset BeiDou grid segmentation technology; wherein, the three-dimensional voxel grid data includes multiple grid nodes and the node spatial coordinates corresponding to each grid node;

[0025] The flight direction and flight constraints of the UAV are determined based on the landing point coordinates, the takeoff point coordinates, and the preset cone angle threshold.

[0026] For any one of the aforementioned grid nodes;

[0027] Based on the flight constraints and the flight search step size, a neighborhood search is performed on the three-dimensional voxel mesh data to obtain multiple surrounding mesh nodes of the mesh node in the flight direction.

[0028] Based on the node spatial coordinates corresponding to each of the multiple surrounding grid nodes, the passable flight area where the multiple surrounding grid nodes are located is obtained.

[0029] The above scheme addresses the insufficient path planning accuracy caused by fixed-step search in traditional methods by establishing a dynamic mapping relationship between terrain features and flight parameters. The flight search step size is dynamically adjusted based on the terrain slope, allowing for smaller step sizes on steep terrain to improve obstacle avoidance accuracy and larger step sizes on gentle terrain to improve search efficiency. Three-dimensional voxel grid data is generated using BeiDou grid partitioning technology to construct a three-dimensional grid model containing spatial coordinate information, providing a precise spatial reference for flight path search. The flight direction is determined based on takeoff and landing point coordinates and a cone angle threshold, forming a heading reference that conforms to actual flight physical constraints. Simultaneously, the cone angle threshold controls the adjustment range of the flight direction to avoid abrupt path changes. During the neighborhood search process, flight constraints and dynamic step size parameters are combined to achieve a progressive region expansion search in three-dimensional space, ensuring that the search range around each grid node not only meets the aircraft's maneuverability limitations but also effectively covers areas with changing terrain features. Finally, through correlation analysis of node spatial coordinates, a sequence of continuously distributed traversable areas along the flight direction is established, providing structured spatial support for subsequent multi-dimensional cost optimization.

[0030] As a preferred example, the step of obtaining the time cost, energy cost, and flight risk of the UAV when flying between any two adjacent passable flight areas based on the flight space coordinates of the candidate flight point and the obstacle coordinates of the obstacle includes:

[0031] Obtain the first candidate flight node of the first accessible flight area and the second candidate flight node of the second accessible flight area in two adjacent accessible flight areas;

[0032] The aircraft density between the first candidate flight node and the second candidate flight node is obtained based on the node spatial coordinates.

[0033] Obtain the wind speed and atmospheric turbulence levels between the first candidate flight node and the second candidate flight node;

[0034] The time cost of the UAV flying between the first candidate flight node and the second candidate flight node is obtained based on the aircraft density, the wind speed and the atmospheric turbulence.

[0035] Obtain the altitude difference between the first candidate flight node and the second candidate flight node;

[0036] Based on the climb drag coefficient, wind drag coefficient, and payload weight of the UAV and the altitude difference, the energy cost of the UAV when flying between the first candidate flight node and the second candidate flight node is obtained.

[0037] Based on the node spatial coordinates and the obstacle coordinates, candidate obstacles between the first candidate flight node and the second candidate flight node are obtained;

[0038] The flight risk of the UAV when flying between the first candidate flight node and the second candidate flight node is obtained based on the risk weight coefficient corresponding to the candidate obstacle, the distance between the UAV and the candidate obstacle, the aircraft density and the weight coefficient corresponding to the aircraft density.

[0039] The above scheme achieves comprehensive optimization of path planning by dynamically quantifying multi-dimensional cost factors between adjacent flight areas. First, by acquiring nodes in adjacent flight areas, local relationships between flight paths are established. Aircraft density is calculated based on node spatial coordinates, and combined with wind speed and atmospheric turbulence levels, the impact of path congestion and environmental disturbances on flight time is dynamically assessed, overcoming the limitations of traditional static time estimation. By introducing the physical relationship between altitude difference and the UAV's climb drag coefficient, wind resistance coefficient, and payload weight, aircraft dynamic parameters are combined with terrain features to accurately calculate energy consumption, avoiding errors from estimating energy consumption solely based on straight-line distance. Furthermore, candidate obstacles are identified using obstacle coordinates, and a dynamic risk assessment model is constructed by combining risk weight coefficients, obstacle distances, and aircraft density to quantify path safety risks. The weight coefficient corresponding to aircraft density reflects the impact of airspace traffic complexity on risk, while the risk weight coefficient distinguishes the threat levels of different obstacles, thus achieving a balance between safety and efficiency in path selection.

[0040] As a preferred example, determining each flight point in the passable flight area from a plurality of candidate flight points based on the UAV's flight mission, time cost, energy cost, and flight risk includes:

[0041] When the UAV flies from the first accessible flight area to the second accessible flight area, for any second candidate flight node in the second accessible flight area:

[0042] The time weight corresponding to the time cost, the cost weight corresponding to the energy cost, and the risk weight corresponding to the flight risk are determined based on the flight mission of the UAV.

[0043] Based on the time weight, cost weight, risk weight, time cost, energy cost, and risk weight, the flight cost between the second candidate flight node and the first passable flight area is obtained through a preset weight summation formula.

[0044] The second candidate flight node with the lowest flight cost is selected from multiple second candidate flight nodes as the flight point of the second passable flight area.

[0045] The above scheme addresses the shortcomings of traditional methods with their single optimization objective by establishing a multi-dimensional cost dynamic decision-making mechanism. First, it dynamically assigns weight coefficients to three dimensions—time, energy, and risk—based on flight mission type, reflecting the different priorities under different mission scenarios. For example, emergency missions are given a higher weight for time, while long-endurance missions prioritize energy. Second, a weighted summation formula is used to linearly weight the normalized time cost, energy cost, and flight risk, constructing a comprehensive flight cost assessment model. Finally, a minimum cost node selection mechanism is used to progressively construct the globally optimal path between adjacent flight areas. The dynamic correlation mechanism between flight missions and weight coefficients overcomes the limitations of traditional fixed-weight models, and the multi-objective linearly weighted decision model enables rapid solutions under complex constraints.

[0046] As a preferred example, generating the low-altitude flight path corresponding to the UAV based on the flight space coordinates of the flight point, the takeoff point, and the landing point includes:

[0047] The initial flight path of the UAV is generated based on the takeoff point coordinates of the takeoff point, the landing point coordinates of the landing point, and the flight space coordinates of each of the multiple flight points.

[0048] The initial flight path is optimized according to the preset curve fitting formula and curvature constraint algorithm to obtain the low-altitude flight path corresponding to the UAV.

[0049] The above scheme generates an initial path by integrating multi-dimensional spatial coordinates and improves the feasibility and safety of the path based on dynamic optimization algorithms. First, an initial flight path is generated based on the spatial coordinates of the takeoff point, landing point, and intermediate flight points, ensuring the basic connectivity of the path in three-dimensional space. Then, a curve fitting formula is used to smooth the initial path, eliminating broken lines or abrupt changes, thereby reducing the complexity of aircraft maneuver control. Simultaneously, a curvature constraint algorithm is used to limit the curvature of the fitted path, ensuring that the path conforms to physical maneuverability constraints such as the minimum turning radius of the UAV, avoiding the risk of loss of control due to path curvature exceeding the aircraft's performance limits. By combining initial path generation with dynamic optimization, the spatial rationality of the path planning is preserved while enhancing the smoothness and executability of the path, thus achieving comprehensive optimization of the flight path under multi-dimensional constraints.

[0050] As a preferred example, generating the low-altitude flight path corresponding to the UAV based on the flight space coordinates of the flight point, the takeoff point, and the landing point further includes:

[0051] When the UAV flies along the low-altitude flight path, it acquires new obstacle detection results in real time according to the preset detection radius;

[0052] When a new obstacle is detected, a new path is selected from the low-altitude flight paths based on the real-time positioning of the UAV and the preset path planning radius.

[0053] Multiple passable flight areas are obtained from the path to be replanned, and a new planned path is generated based on the time cost, energy cost and flight risk between the passable flight areas;

[0054] The path to be replanned is replaced with the newly planned path to generate the low-altitude flight path corresponding to the UAV.

[0055] The above scheme addresses the problem of sudden changes in the flight environment through dynamic monitoring and local path reconstruction mechanisms. New obstacle detection results are acquired in real time based on a preset detection radius, balancing the real-time nature of environmental perception with computational resource consumption by limiting the detection range. When a new obstacle is detected, a path to be replanned is selected based on real-time positioning and the path planning radius, avoiding the computational burden of global path reconstruction through local path adjustment. After obtaining the passable flight area of ​​the affected region, a new planned path is generated by combining time cost, energy cost, and flight risk, ensuring the continuity of multi-dimensional path optimization objectives. Finally, dynamic updates of the flight path are achieved through path replacement operations, maintaining the stability of the original path planning framework while enhancing the system's responsiveness to sudden obstacles. The synergistic effect of each step forms a closed-loop dynamic adjustment mechanism. The preset parameter settings for the detection radius and path planning radius, along with the design of retaining the original cost calculation model during local path optimization, collectively achieve compatibility between dynamic path adjustment and global optimization objectives.

[0056] Secondly, the present invention discloses a low-altitude flight path planning system, including a landing point optimization module, a passability analysis module, a flight cost module, a flight point screening module, and a path planning module.

[0057] The landing point optimization module is used to obtain the landing area of ​​the UAV and obtain environmental data of each candidate landing point in the landing area, so as to determine the landing point corresponding to the UAV from multiple candidate landing points based on the environmental data.

[0058] The accessibility analysis module is used to obtain multiple accessible flight areas distributed along the flight direction of the UAV based on the landing point and the UAV's takeoff point using a preset neighborhood search algorithm; wherein each accessible flight area includes multiple candidate flight points;

[0059] The flight cost module is used to obtain obstacles in each of the passable flight areas, and to obtain the time cost, energy cost and flight risk of the UAV when flying between any two adjacent passable flight areas based on the flight space coordinates of the candidate flight point and the obstacle coordinates of the obstacles;

[0060] The flight point screening module is used to determine each flight point in the passable flight area from multiple candidate flight points based on the UAV's flight mission, time cost, energy cost, and flight risk.

[0061] The path planning module is used to generate a low-altitude flight path for the UAV based on the flight space coordinates of the flight point, the take-off point, and the landing point.

[0062] This invention discloses a low-altitude flight path planning system that achieves comprehensive planning of low-altitude flight paths for unmanned aerial vehicles (UAVs) through multi-dimensional cost integration and dynamic path optimization. First, landing points are screened using environmental data to ensure the safety and terrain adaptability of the landing area, avoiding landing risks caused by single geometric factors. Second, based on the division of traversable flight areas according to flight direction distribution, a neighborhood search algorithm is used to generate a candidate path space that conforms to the UAV's motion constraints, providing a foundation for subsequent multi-dimensional cost calculations. By introducing a joint calculation model of time cost, energy cost, and flight risk, the dynamic relationship between obstacle coordinates and flight space coordinates is quantified into optimizable target parameters, overcoming the limitations of traditional binary obstacle avoidance. Differentiated weights are assigned to different cost factors according to flight mission requirements, achieving dynamic optimization of candidate flight points and ensuring a strong correlation between path planning and mission objectives. Finally, a low-altitude flight path is generated through a sequence of coordinate points, forming a globally optimal solution that balances efficiency, energy consumption, and safety, ensuring the efficiency of UAV mission execution.

[0063] Thirdly, the present invention discloses a terminal device, including a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements a low-altitude flight path planning method as described in the first aspect.

[0064] Fourthly, the present invention discloses a computer-readable storage medium comprising: a stored computer program, wherein, when the computer program is executed, the device on which the computer-readable storage medium is located is controlled to perform a low-altitude flight path planning method as described in the first aspect. Attached Figure Description

[0065] To more clearly illustrate the technical solution of this application, the drawings used in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0066] Figure 1 This is a flowchart illustrating a low-altitude flight path planning method disclosed in an embodiment of the present invention;

[0067] Figure 2 This is a schematic diagram of the structure of a low-altitude flight path planning system disclosed in an embodiment of the present invention. Detailed Implementation

[0068] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0069] Reference Figure 1 To improve the efficiency and reliability of UAVs during mission execution, this embodiment discloses a low-altitude flight path planning method, which mainly includes:

[0070] Step 101: Obtain the landing area of ​​the UAV and obtain environmental data for each candidate landing point in the landing area, so as to determine the landing point corresponding to the UAV from multiple candidate landing points based on the environmental data.

[0071] Step 102: Based on the landing point and the takeoff point of the UAV, obtain multiple passable flight areas distributed along the flight direction of the UAV using a preset neighborhood search algorithm; wherein each passable flight area includes multiple candidate flight points.

[0072] Step 103: Obtain obstacles in each of the passable flight areas, and obtain the time cost, energy cost and flight risk of the UAV when flying between any two adjacent passable flight areas based on the flight space coordinates of the candidate flight point and the obstacle coordinates of the obstacles.

[0073] Step 104: Determine the flight point in each of the accessible flight areas from the multiple candidate flight points based on the UAV's flight mission, time cost, energy cost, and flight risk.

[0074] Step 105: Generate the low-altitude flight path corresponding to the UAV based on the flight space coordinates of the flight point, the take-off point, and the landing point.

[0075] In this first embodiment, to avoid interference with the drone's mission execution during landing, and to ensure the reliability of the mission execution by achieving a precise and reliable landing of the drone, this embodiment implements step 101 through the following steps. Specifically, step 101 includes:

[0076] Step 1011: Generate the landing area of ​​the UAV based on the target coordinates of the preset target point and the preset area radius, divide the landing area into a grid, obtain multiple candidate landing nodes corresponding to the landing area, and obtain the digital elevation data of the landing area.

[0077] Step 1012: Extract the elevation value of each candidate landing node and the landing obstacles in the landing area and the landing obstacle coordinates of each obstacle based on the digital elevation data;

[0078] Step 1013: For any candidate landing node: extract multiple surrounding landing nodes corresponding to the candidate landing node through a preset neighborhood window, and obtain the slope value corresponding to the candidate landing node according to the elevation value of the surrounding landing nodes; obtain the straight-line distance between the candidate landing node and each landing obstacle according to the landing obstacle coordinates;

[0079] Step 1014: Obtain the clearance region corresponding to the candidate landing node and the clearance area of ​​the clearance region according to the preset multi-neighbor connected component analysis method;

[0080] Step 1015: Select the landing point of the UAV from the multiple candidate landing nodes based on the slope value, the straight-line distance of the obstacle, and the clearance area.

[0081] Specifically, in this embodiment, firstly, a preset target point for the drone's landing is determined from the task data corresponding to the drone. Then, the landing area of ​​the drone is generated based on the coordinates of the target point and a preset area radius. For example, a three-dimensional spatial area with a radius of 500 meters is selected as the landing area of ​​the drone, centered on the target point.

[0082] Next, a grid model corresponding to the landing area can be constructed using BeiDou grid partitioning technology and digital elevation data (such as a Digital Elevation Model, DEM) of the landing area. Preferably, the grid model can also be optimized based on the lidar point cloud data and obstacle database corresponding to the landing area, such as obstacle data including trees, cables, and buildings, to obtain a refined grid model. It should be noted that the grid resolution of the refined grid model can be dynamically adjusted according to the scene complexity, such as 5m × 5m × 0.5m.

[0083] After constructing the grid model, the slope corresponding to each candidate landing node in the landing area can be calculated using a third-order inverse distance weighted method based on the digital elevation data of the landing area. Specifically, for each candidate landing node in the grid model, multiple surrounding landing nodes corresponding to each candidate landing node are extracted using a preset neighborhood window. For example, a 3×3 neighborhood window can be used to extract eight surrounding landing nodes corresponding to each candidate landing node. Next, the elevation value of each surrounding landing node is obtained, and the slope value corresponding to each candidate landing node is obtained according to a preset slope calculation formula, ultimately forming a slope matrix covering the entire candidate area. The slope value calculation formula is as follows: Among them, the and The elevation value represents the elevation value corresponding to each of the candidate landing nodes; where the slope represents the gradient value.

[0084] Next, the minimum straight-line distance between each candidate landing node and the obstacle is calculated, thus obtaining the obstacle straight-line distance for the non-UAV's real-time flight distance. This involves identifying obstacle coordinates such as trees, cables, and buildings using LiDAR point cloud data. For each candidate landing node (grid center), the Euclidean distance to the nearest obstacle is calculated, and the minimum value is taken as the obstacle straight-line distance for that point. Simultaneously, a multi-neighborhood connected domain analysis method is employed, such as using an 8-neighborhood connected domain analysis algorithm, to mark the connectivity of the clearance grid including each candidate landing node. This involves dividing adjacent clearance grids (e.g., vertically, horizontally, or diagonally adjacent) into the same connected region. By counting the number of grids contained in each connected region and combining this with the grid area (e.g., a 5m × 5m grid corresponds to 25㎡), the clearance area corresponding to each candidate landing node is calculated.

[0085] Next, the landing point for the UAV is selected from multiple candidate landing nodes based on preset slope thresholds, obstacle straight-line distance thresholds, and clearance area thresholds. For example, candidate landing nodes with a slope less than 8° are selected to ensure a gentle landing terrain and avoid the risk of tipping over. Areas with obstacle straight-line distances greater than three times the UAV's altitude are selected to ensure a safe buffer with obstacles during landing, and areas with a continuous clearance area greater than 2πR are also selected. 2 (R is the radius of the required airspace, i.e. the area of ​​the circular airspace that needs to be guaranteed around the landing point) connecting areas to ensure sufficient space for landing and emergency go-around.

[0086] In the preferred embodiment of this example, when multiple valid candidate landing points are selected as the final landing point after environmental analysis, a multi-objective evaluation model can be used to generate the globally optimal landing point. Specifically, a quantitative safety score model (safety_score) is constructed, which is calculated using a weighted average of three levels of indicators. For example, the adaptability weight corresponding to the slope value (weight 0.4): with an 8° threshold, it is calculated as (1-dem_slope / 8), and the gentler the slope, the higher the score; the safety margin corresponding to the straight-line distance from the obstacle (weight 0.4): with 3H as the benchmark, it is calculated as (dist_min / 3H), and the farther the distance from the obstacle, the higher the score; the space adequacy corresponding to the clearance area (weight 0.2): with 2πR... 2 As a standard, calculate (area / 2πR) 2 The larger the continuous clearance area, the higher the score. A weighted sum of the three factors yields a safety coefficient, ranging from 0 to 1; a higher score indicates better landing safety. Then, based on the evaluation values ​​from the multi-objective evaluation model, the effective candidate landing point corresponding to the evaluation values ​​is selected as the landing point.

[0087] In this embodiment, the above steps achieve refined selection of landing points through multi-dimensional terrain feature fusion. First, a landing area is generated based on the target coordinates and region radius to ensure coverage of the potentially feasible range. A high-precision terrain model is established through grid division and digital elevation data acquisition. For each candidate node, the slope value is calculated using a neighborhood window to eliminate the impact of steep terrain on landing stability. Straight-line distances are calculated using obstacle coordinates to avoid the risk of collisions with obstacles during landing. Multi-neighbor connected domain analysis is used to quantify the clearance area, ensuring the UAV has sufficient landing buffer space. Finally, a joint evaluation is performed based on the three dimensions of slope, obstacle distance, and clearance area to select landing points that simultaneously meet the requirements of gentle terrain, safe obstacle distance, and adequate clearance area, thus solving the problem of insufficient landing point safety caused by the single-dimensional evaluation of traditional methods.

[0088] In this first embodiment, in order to improve the efficiency of path rules by reducing the amount of data processing, thereby improving the efficiency of UAV mission execution, step 102 can be performed to screen for passable flight areas. The steps mainly include:

[0089] Step 1021: Determine the terrain area where the UAV is located based on the landing point coordinates and the takeoff point coordinates, and determine the flight search step length of the UAV based on the slope of the terrain area;

[0090] Step 1022: Obtain three-dimensional voxel grid data of the UAV during flight based on the landing point coordinates, the takeoff point coordinates, and the preset BeiDou grid partitioning technology; wherein, the three-dimensional voxel grid data includes multiple grid nodes and the node spatial coordinates corresponding to each grid node;

[0091] Step 1023: Determine the flight direction and flight constraints of the UAV based on the landing point coordinates, the takeoff point coordinates, and the preset cone angle threshold.

[0092] Step 1024: For any one of the grid nodes; perform a neighborhood search on the three-dimensional voxel grid data according to the flight constraints and the flight search step size to obtain multiple surrounding grid nodes of the grid node in the flight direction;

[0093] Step 1025: Based on the node spatial coordinates corresponding to each of the multiple surrounding grid nodes, obtain the passable flight area where the multiple surrounding grid nodes are located.

[0094] Specifically, in this embodiment, the terrain region where the UAV is located is first determined based on the landing point coordinates and the takeoff point coordinates, and the flight search step size of the UAV is determined according to the slope of the terrain region. The slope of the terrain region is used as the basis for determining the flight search step size. The expression for calculating the real-time optimization step size is: Wherein, base_step is the base step size; The absolute value of the slope of the terrain region, the The larger the value, the steeper the slope of the terrain, and the smaller the step size, ensuring search accuracy in complex terrain.

[0095] While determining the flight search step size, the three-dimensional space containing the landing point coordinates and the takeoff point coordinates is discretized into a voxel grid based on the Beidou grid partitioning technology to obtain the three-dimensional voxel grid data of the UAV during flight; wherein, the three-dimensional voxel grid data includes a dynamic attribute set: basic attributes: spatial coordinates (x,y,z), accessibility (pre-determined by lidar point cloud and terrain data, marking obstacle areas).

[0096] Then, a gradient constraint expansion method is used to replace the traditional 26-neighborhood search, reducing invalid computation and improving adaptability to complex terrain. For each grid node in the 3D voxel mesh data, the unit direction vector from the current grid node to the target grid node is calculated. Then, flight constraints are generated based on a preset cone angle threshold, i.e., based on the unit direction vector and the cone angle threshold. For example, when the cone angle threshold is 30°, candidate neighbor nodes are generated within a cone angle range of ±30° of the unit direction vector, focusing on the target direction and reducing redundant searches. After the candidate nodes are generated, their drivability is immediately verified, such as determining whether the node is an obstacle to filter out invalid nodes, obtaining multiple surrounding grid nodes for each grid node in the flight direction. Then, based on the node spatial coordinates corresponding to each of the multiple surrounding grid nodes, the drivable flight area where the multiple surrounding grid nodes are located is obtained.

[0097] In this embodiment, the above steps address the insufficient path planning accuracy caused by fixed-step-size search in traditional methods by establishing a dynamic mapping relationship between terrain features and flight parameters. The flight search step size is dynamically adjusted based on the terrain slope, allowing for smaller step sizes to improve obstacle avoidance accuracy on steep terrain and larger step sizes to improve search efficiency on gentle terrain. Three-dimensional voxel grid data is generated using BeiDou grid partitioning technology to construct a three-dimensional grid model containing spatial coordinate information, providing a precise spatial reference for flight path search. The flight direction is determined based on takeoff and landing point coordinates and a cone angle threshold, forming a heading reference that conforms to actual flight physical constraints. Simultaneously, the cone angle threshold controls the adjustment range of the flight direction to avoid abrupt path changes. During the neighborhood search process, flight constraints and dynamic step-size parameters are combined to achieve a progressive region expansion search in three-dimensional space, ensuring that the surrounding search range of each grid node not only meets the aircraft's maneuverability limitations but also effectively covers areas with changing terrain features. Finally, through correlation analysis of node spatial coordinates, a sequence of continuously distributed traversable areas along the flight direction is established, providing structured spatial support for subsequent multi-dimensional cost optimization.

[0098] In this first embodiment, to improve the efficiency and reliability of the UAV during mission execution, step 103 can calculate the time cost, energy cost, and flight risk of the UAV during flight through the following steps, thereby improving the efficiency and reliability of path planning. Step 103 includes:

[0099] Step 1031: Obtain the first candidate flight node of the first accessible flight area and the second candidate flight node of the second accessible flight area in two adjacent accessible flight areas;

[0100] Step 1032: Obtain the aircraft density between the first candidate flight node and the second candidate flight node based on the node spatial coordinates;

[0101] Step 1033: Obtain the wind speed and atmospheric turbulence levels between the first candidate flight node and the second candidate flight node;

[0102] Step 1034: Calculate the time cost of the UAV flying between the first candidate flight node and the second candidate flight node based on the aircraft density, the wind speed, and the atmospheric turbulence level.

[0103] Step 1035: Obtain the altitude difference between the first candidate flight node and the second candidate flight node; based on the climb drag coefficient, wind drag coefficient, and payload weight of the UAV and the altitude difference, obtain the energy cost of the UAV flying between the first candidate flight node and the second candidate flight node; obtain candidate obstacles between the first candidate flight node and the second candidate flight node based on the node spatial coordinates and the obstacle coordinates; obtain the flight risk of the UAV flying between the first candidate flight node and the second candidate flight node based on the risk weight coefficient corresponding to the candidate obstacle, the distance between the UAV and the candidate obstacle, the aircraft density, and the weight coefficient corresponding to the aircraft density.

[0104] Specifically, in this embodiment, after screening the passable flight areas, the flight cost and flight risk corresponding to each candidate flight node in the passable flight areas are calculated according to a preset improved A* 3D path algorithm. Specifically, the improved A* 3D path algorithm includes a dynamic cost layer integrating a 3D cost model of time, energy, and risk; an environmental perception layer that fuses BeiDou grid spatial data and airborne sensor data in real time; and an adaptive search layer based on a heuristic function that dynamically adjusts the search strategy according to the task type.

[0105] This refers to any two adjacent passable flight areas distributed along the flight direction; and multiple second candidate flight nodes in the first passable flight area where the UAV is already located and the second passable flight area it is about to fly to. Next, two core lists are initialized, including an open list and a closed list. Multiple second candidate flight nodes are added to the open list, and initial cost parameters are set: actual cost g = 0, and the combined cost of flying from the first candidate flight node to the second candidate flight node.

[0106] In this embodiment, for any second candidate flight node, the time cost, energy cost, and flight risk of the UAV flying between the first candidate flight node and the second candidate flight node are calculated using the time cost model, energy cost model, and real-time risk quantification model in the preset three-dimensional cost modeling unit.

[0107] In this embodiment, when accumulating the time cost gt(n), the time cost represents the actual time consumed from the starting point to the current node (unit: seconds), which is affected by wind speed, airspace congestion, and airflow disturbance. The calculation formula is gt(n) = g t (parent) + Δt, where Δt is the time increment of the current segment (parent→n), calculated by integration using the following formula: Δt=∫[1 / (v0+v_wind·cosθ)]·ds (v0 is the drone's cruising speed, v_wind is the real-time wind speed vector, θ is the angle between the wind direction and the flight direction, and ds is the path micro-segment length). Simultaneously, correction terms are introduced: airspace congestion impact: Δt=Δt·(1+ρ)ρ is the airspace congestion index, 0≤ρ≤1, calculated based on ADS-B data: ρ=number of aircraft in the current area / (airspace volume × safe distance threshold)); airflow disturbance impact: Δt=Δt·(1+ω) (ω is the airflow disturbance coefficient, ω=0.05×|surface temperature difference|, updated in real-time by terrain thermal effects), etc., thus obtaining the time cost model; where the time cost model is: Δt=∫(v,ρ,ω)ds; where the time cost model calculates the time increment along path s. The integral represents the cumulative time along the path. Wind speed *v* affects the drone's flight speed; a tailwind increases speed, while a headwind decreases it. Airspace congestion *ρ* represents the density of other aircraft within a given airspace, affecting the degree to which the drone needs to avoid them, thus increasing the time. *ω* represents airflow disturbance, indicating the degree of atmospheric turbulence, which requires the drone to adjust its flight attitude and speed, increasing flight time. These three parameters collectively affect the drone's actual flight speed, thereby influencing the time required to traverse path segment *ds*.

[0108] And cumulative energy consumption g e When (n), the cumulative energy consumption represents the actual energy consumption (unit: joules) from the starting point to the current node, which is related to changes in altitude, horizontal distance, and load. The calculation formula is: g e (n)=g e (parent) + E, where g e(parent) represents the energy consumption of the preceding node. This leads to the energy cost model, expressed as: E = Σ[K_m·|Δh| + k_f(v)]·e^β. This model calculates the energy consumption during the drone's flight. The summation symbol Σ signifies dividing the path into segments, calculating the energy consumption of each segment (especially the climb segment), and then summing them. K_m represents the climb drag coefficient, indicating the energy required to climb a unit height, which is related to the drone's weight and aerodynamic characteristics. Δh represents the altitude change, i.e., the altitude at which it climbs or descends. Δh is negative during descent, but energy consumption is usually lower during descent. k_f(v) represents the drag function, indicating the additional energy consumption caused by the relationship between wind speed and drone speed during level flight. This function is usually related to speed v, as air resistance is proportional to the square of speed, making it a quadratic function. β: payload coefficient, related to the weight of the payload carried by the drone. The greater the payload, the higher the energy consumption. Here, we use the exponential form e^β to illustrate that the impact of payload on energy consumption is non-linear and grows rapidly. The entire energy cost model is divided into two main parts: one is the energy change due to climb (or descent) (K_m·|Δh|), and the other is the energy consumption to overcome wind resistance (k_f(v)). This is then multiplied by an exponential term (e^β) related to payload, because payload affects not only climb energy consumption but also level flight energy consumption.

[0109] When calculating the real-time risk R(n) of the combined obstacle threat and traffic conflict risk, the formula for calculating the real-time risk is: R(n) = λ1·(1 / d_obs) 2 The real-time risk quantification model is R = λ_obs·d^{-2} + λ_tra·δ (where d_obs is the distance from the node to the nearest obstacle, δ is the regional traffic density, and λ1 and λ2 are risk weights). This model quantifies risk as a weighted sum of three parts: λ_obs: obstacle risk weight coefficient, used to adjust the contribution of the obstacle risk term; d: distance from the UAV to the nearest obstacle. Note the exponent of -2, meaning the risk increases sharply as the distance decreases (similar to an inverse square relationship); λ_tra: traffic risk weight coefficient, used to adjust the risk from air traffic density. δ is the traffic density, representing the number or density of other aircraft within a certain airspace around the UAV. Higher density means higher risk. The risk model consists of two parts: obstacle risk (proportional to the negative square of the obstacle distance) and air traffic risk (proportional to traffic density). The weight coefficients λ_obs and λ_tra are then used to adjust the proportion of these two parts in the total risk.

[0110] In this embodiment, the above steps achieve comprehensive optimization of path planning by dynamically quantifying multi-dimensional cost factors between adjacent flight areas. First, by acquiring nodes in adjacent flight areas, local correlations of flight paths are established. Aircraft density is calculated based on node spatial coordinates, and combined with wind speed and atmospheric turbulence levels, the impact of path congestion and environmental disturbances on flight time is dynamically assessed, overcoming the limitations of traditional static time estimation. By introducing the physical relationship between altitude difference and the UAV's climb drag coefficient, wind resistance coefficient, and payload weight, aircraft dynamic parameters are combined with terrain features to accurately calculate energy consumption, avoiding errors from estimating energy consumption solely based on straight-line distance. Furthermore, candidate obstacles are identified using obstacle coordinates, and a dynamic risk assessment model is constructed by combining risk weight coefficients, obstacle distances, and aircraft density to quantify path safety risks. The weight coefficient corresponding to aircraft density reflects the impact of airspace traffic complexity on risk, while the risk weight coefficient distinguishes the threat levels of different obstacles, thereby achieving a balance between safety and efficiency in path selection.

[0111] In this first embodiment, to improve the efficiency and accuracy of path planning, thereby enhancing the efficiency and reliability of the UAV in performing its missions, step 104 involves selecting flight points based on a comprehensive cost and risk assessment. Specifically, step 104 includes:

[0112] Step 1041: When the UAV flies from the first accessible flight area to the second accessible flight area, for any second candidate flight node in the second accessible flight area:

[0113] Step 1042: Determine the time weight corresponding to the time cost, the cost weight corresponding to the energy cost, and the risk weight corresponding to the flight risk based on the flight mission of the UAV;

[0114] Step 1043: Based on the time weight, cost weight, risk weight, time cost, energy cost, and risk weight, obtain the flight cost between the second candidate flight node and the first passable flight area using a preset weight summation formula;

[0115] Step 1044: Select the second candidate flight node with the lowest flight cost from multiple second candidate flight nodes as the flight point of the second passable flight area.

[0116] Specifically, in this embodiment, a time weight corresponding to the time cost, a cost weight corresponding to the energy cost, and a risk weight corresponding to the flight risk are adaptively allocated according to the task type. For example, a medical emergency mission: ω t ≥0.7, ω e ≤0.2, ωr ≤0.1 (priority timeliness); Logistics and delivery task: ω e ≥0.6, ω t ≤0.3, ω r ≤0.1 (prioritizing low energy consumption); Disaster reconnaissance mission: ω r ≥0.5, ω t ≤0.3, ω e ≤0.2 (prioritizing low risk). Where, ω t Represents time weight, ω e Represents cost weight, ω r Represents risk weight.

[0117] Next, the flight cost between the second candidate flight node and the first passable flight area is obtained according to a preset weighted summation formula. The formula for calculating the flight cost is: f(n) = ω t ·g t (n)+ω e ·g e (n)+ω r ·R(n), where f(n) represents the flight cost.

[0118] In this embodiment, the above steps address the shortcomings of traditional methods with their single optimization objective by establishing a multi-dimensional cost dynamic decision-making mechanism. First, weight coefficients for time, energy, and risk are dynamically assigned based on flight mission type, reflecting the priority differences under different mission scenarios. For example, emergency missions are given a higher weight for time, while long-endurance missions emphasize energy. Second, a weighted summation formula is used to linearly weight the normalized time cost, energy cost, and flight risk, constructing a comprehensive flight cost evaluation model. Finally, a minimum cost node selection mechanism is used to progressively construct the globally optimal path between adjacent flight areas. The dynamic correlation mechanism between flight missions and weight coefficients overcomes the limitations of the traditional fixed-weight model, and the multi-objective linear weighted decision-making model enables rapid solutions under complex constraints.

[0119] In this first embodiment, to ensure the accuracy of the final generated low-altitude flight path, step 105 can optimize the path through the following steps. Step 105 includes:

[0120] Step 1051: Generate the initial flight path corresponding to the UAV based on the takeoff point coordinates, the landing point coordinates, and the flight space coordinates of each of the multiple flight points;

[0121] Step 1052: Optimize the initial flight path according to the preset curve fitting formula and curvature constraint algorithm to obtain the low-altitude flight path corresponding to the UAV.

[0122] Step 1053: When the UAV flies according to the low-altitude flight path, it acquires new obstacle detection results in real time according to the preset detection radius; when a new obstacle is detected, it selects a path to be replanned from the low-altitude flight path according to the real-time positioning of the UAV and the preset path planning radius; it acquires multiple passable flight areas in the path to be replanned, and generates a new planned path according to the time cost, energy cost and flight risk between the passable flight areas; it replaces the path to be replanned with the new planned path to generate the low-altitude flight path corresponding to the UAV.

[0123] Specifically, in this embodiment, an initial flight path is generated based on the takeoff point coordinates, the landing point coordinates, and the flight space coordinates of each of the multiple flight points. A curve fitting formula, such as cubic B-spline curve fitting, is used to eliminate sawtooth fluctuations and ensure the trajectory is continuous and differentiable. Simultaneously, a curvature constraint algorithm is used to ensure that the radius of curvature at each point on the path is greater than the minimum turning radius of the UAV, satisfying the physical motion limits. Finally, a low-altitude flight path corresponding to the UAV is generated, covering the core mission requirements. This low-altitude flight path is the time-optimal path, the energy-optimal path, and the risk-optimal path. The time-optimal path minimizes total flight time by avoiding highly congested airspace and optimizing flight speed parameters (such as acceleration during tailwinds), making it suitable for time-sensitive tasks such as medical emergency rescue. The energy-optimal path minimizes energy consumption by smoothing altitude changes (reducing |Δh|) and avoiding headwinds and strong airflow areas, making it suitable for cost-sensitive tasks such as logistics and delivery. The risk-optimal path minimizes the cumulative R(n) value by staying away from areas with dense obstacles, low-altitude airspace conflict zones, and disaster-affected areas, making it suitable for high-risk scenarios such as disaster reconnaissance. All generated low-altitude flight paths come with quantitative evaluation indicators (total time, total energy consumption, and risk value), allowing users to select as needed or for the system to automatically match the task type.

[0124] To address sudden environmental changes during flight (such as temporary obstacles or sudden airflow changes), a low-latency response mechanism is established: when the airborne LiDAR detects a new obstacle (such as birds or a temporary no-fly zone) within 50 meters of the current path, local replanning is immediately initiated. A spherical region with a radius of 50 meters centered on the current UAV position is defined as the replanning boundary. Only the path segment within this region is adjusted, preserving the unaffected initial section of the original path to reduce computational load. Within this local area, based on the cost function from step 2 and the expansion strategy from step 3, the optimal sub-path is re-searched, ensuring that the new path satisfies a curvature radius ≥ the UAV's minimum turning radius and a replanning response time ≤ 100ms.

[0125] In this embodiment, the above steps generate an initial path by integrating multi-dimensional spatial coordinates and improve the feasibility and safety of the path based on a dynamic optimization algorithm. First, an initial flight path is generated based on the spatial coordinates of the takeoff point, landing point, and intermediate flight points, ensuring the basic connectivity of the path in three-dimensional space. Then, the initial path is smoothed using a curve fitting formula to eliminate broken line segments or abrupt changes, thereby reducing the complexity of aircraft maneuver control. Simultaneously, a curvature constraint algorithm is used to limit the curvature of the fitted path, ensuring that the path conforms to physical maneuverability constraints such as the minimum turning radius of the UAV, avoiding the risk of loss of control due to path curvature exceeding the aircraft's performance limits. By combining initial path generation with dynamic optimization, the spatial rationality of the path planning is preserved while enhancing the smoothness and executability of the path, thus achieving comprehensive optimization of the flight path under multi-dimensional constraints.

[0126] In this embodiment one implementation, refer to Figure 1 The low-altitude flight path planning method shown employs a cloud-based collaborative architecture when controlling the aircraft to perform flight missions. Based on the computational load and real-time requirements of the functional modules, the modules are distributed across a backend server (including cloud and edge computing nodes) and the UAV's onboard system. The specific division of labor and collaborative process are as follows: The backend server runs core modules including global path planning (relying on GPU parallel computation of the Pareto optimal path in a million-level point cloud scenario, and using LRU caching to pre-compute the static cost map) and a landing point optimization engine (requiring large-scale analysis of slope, obstacles, and airspace area to generate Pareto solution sets). These modules rely on powerful computing capabilities (such as GPU parallelism and large-scale spatial data analysis) and global information (overall airspace congestion and multi-source meteorological data), and have less stringent real-time requirements (second-level to minute-level response is sufficient). The UAV's onboard system runs lightweight modules covering dynamic environmental response mechanisms (real-time fusion of onboard sensor data to update local risk values), local path replanning (rapid adjustment within ±30° cone angle constraints based on the global path in the event of sudden obstacles), and emergency obstacle avoidance (millimeter-level response for hovering, emergency climb, and other actions). These modules require ultra-low latency (obstacle avoidance response <100ms), rely on local sensor data (not suitable for backhaul), and can ensure basic flight safety in weak network / network outage scenarios.

[0127] The collaborative operation process for controlling drones to perform flight missions, taking medical supply delivery as an example, is as follows: Mission Issuance Phase: The backend inputs the target point and airspace radius and initiates global planning. The onboard system receives the mission instructions. Communication requirements at this stage are low bandwidth (only transmitting instructions). Global Planning Phase: The backend runs an improved A* algorithm to generate 3-5 Pareto paths. The onboard system waits for the results. Communication requirements are medium bandwidth (transmitting path points). Flight Execution Phase: The backend monitors disaster prediction and pushes environmental updates (such as no-fly zone expansion). The onboard system flies along the global path and updates local risk values ​​in real time. Communication requirements are high bandwidth, supporting streaming updates. Sudden Obstacle Response Phase: No backend operation is required. The onboard system autonomously completes local replanning based on cone angle constraints, requiring no communication. Landing Phase: The backend pushes optimized landing point coordinates. The onboard system performs a precise landing and visually confirms airspace conditions. Communication requirements are low bandwidth, transmitting coordinates. Through this architecture, the system can utilize the powerful computing capabilities of the backend to handle complex global tasks and rely on the onboard system to meet real-time response requirements, achieving efficient collaborative operation.

[0128] On the other hand, refer to Figure 2 This embodiment also discloses a low-altitude flight path planning system, which mainly includes a landing point optimization module 201, a passability analysis module 202, a flight cost module 203, a flight point screening module 204, and a path planning module 205.

[0129] The landing point optimization module 201 is used to acquire the landing area of ​​the UAV and acquire environmental data of each candidate landing point in the landing area, so as to determine the landing point corresponding to the UAV from multiple candidate landing points based on the environmental data.

[0130] The accessibility analysis module 202 is used to obtain multiple accessible flight areas distributed along the flight direction of the UAV based on the landing point and the take-off point of the UAV using a preset neighborhood search algorithm; wherein each accessible flight area includes multiple candidate flight points.

[0131] The flight cost module 203 is used to acquire obstacles in each of the passable flight areas, and to acquire the time cost, energy cost and flight risk of the UAV when flying between any two adjacent passable flight areas based on the flight space coordinates of the candidate flight point and the obstacle coordinates of the obstacles.

[0132] The flight point screening module 204 is used to determine each flight point in the passable flight area from multiple candidate flight points based on the UAV's flight mission, time cost, energy cost, and flight risk.

[0133] The path planning module 205 is used to generate a low-altitude flight path for the UAV based on the flight space coordinates of the flight point, the take-off point, and the landing point.

[0134] Based on the above embodiment of a low-altitude flight path planning method, this embodiment also provides a terminal device, which includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor. When the processor executes the computer program, it implements the low-altitude flight path planning method described in this embodiment.

[0135] For example, in this embodiment, the computer program can be divided into one or more modules, which are stored in the memory and executed by the processor to complete the present invention. The one or more modules may be a series of computer program instruction segments capable of performing specific functions, which describe the execution process of the computer program in the terminal device. The terminal device may be a desktop computer, laptop, handheld computer, cloud server, or other computing device. The terminal device may include, but is not limited to, a processor and memory. The processor may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. The general-purpose processor may be a microprocessor or any conventional processor, etc. The processor is the control center of the terminal device, connecting various parts of the terminal device through various interfaces and lines.

[0136] Based on the above-described method embodiments, this embodiment also provides a computer-readable storage medium, including a stored computer program, wherein the computer program, when running, controls the device where the computer-readable storage medium is located to execute the low-altitude flight path planning method described in this embodiment. The modules / units integrated into the device / terminal device, if implemented as software functional units and sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the method described in this embodiment can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, a recording medium, a USB flash drive, a portable hard drive, a magnetic disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), an electrical carrier signal, a telecommunication signal, and a software distribution medium, etc.

[0137] This embodiment discloses a low-altitude flight path planning method, system, device, and medium. Through multi-dimensional cost integration and dynamic path optimization, it achieves comprehensive planning of low-altitude flight paths for UAVs. First, landing points are screened using environmental data to ensure the safety and terrain adaptability of the landing area, avoiding landing risks caused by single geometric factors. Second, based on the division of traversable flight areas according to flight direction distribution, a neighborhood search algorithm is used to generate a candidate path space that conforms to the UAV's motion constraints, providing a foundation for subsequent multi-dimensional cost calculations. By introducing a joint calculation model of time cost, energy cost, and flight risk, the dynamic relationship between obstacle coordinates and flight space coordinates is quantified into optimizable target parameters, overcoming the limitations of traditional binary obstacle avoidance. Differentiated weights are assigned to different cost factors according to flight mission requirements, achieving dynamic optimization of candidate flight points and ensuring a strong correlation between path planning and mission objectives. Finally, a low-altitude flight path is generated through a sequence of coordinate points, forming a globally optimal solution that balances efficiency, energy consumption, and safety, ensuring the efficiency of UAV mission execution.

[0138] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above descriptions are merely specific embodiments of the present invention and are not intended to limit the scope of protection of the present invention. In particular, it should be noted that any modifications, equivalent substitutions, improvements, etc., made by those skilled in the art within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A low-altitude flight path planning method, characterized in that, include: The landing area of ​​the drone is obtained and environmental data of each candidate landing point in the landing area is obtained, so as to determine the landing point corresponding to the drone from multiple candidate landing points based on the environmental data; Based on the landing point and the takeoff point of the UAV, a preset neighborhood search algorithm is used to obtain multiple passable flight areas distributed along the flight direction of the UAV; wherein each passable flight area includes multiple candidate flight points; Obtain obstacles in each of the accessible flight areas, and obtain the time cost, energy cost and flight risk of the UAV when flying between any two adjacent accessible flight areas based on the flight space coordinates of the candidate flight point and the obstacle coordinates of the obstacles; Based on the UAV's flight mission, time cost, energy cost, and flight risk, each flight point in the passable flight area is determined from a plurality of candidate flight points; The low-altitude flight path of the UAV is generated based on the flight space coordinates of the flight point, the take-off point, and the landing point.

2. The low-altitude flight path planning method according to claim 1, characterized in that, The step of acquiring the landing area of ​​the UAV and acquiring environmental data for each candidate landing point within the landing area, so as to determine the landing point corresponding to the UAV from a plurality of candidate landing points based on the environmental data, includes: The landing area of ​​the UAV is generated based on the target coordinates of the preset target point and the preset area radius. The landing area is then divided into grids to obtain multiple candidate landing nodes corresponding to the landing area and to acquire the digital elevation data of the landing area. The elevation value of each candidate landing node and the landing obstacles in the landing area, as well as the landing obstacle coordinates of each obstacle, are extracted based on the digital elevation data. For any of the candidate landing nodes: Multiple surrounding landing nodes corresponding to the candidate landing node are extracted through a preset neighborhood window, and the slope value corresponding to the candidate landing node is obtained based on the elevation value of the surrounding landing nodes. The straight-line distance between the candidate landing node and each landing obstacle is obtained based on the landing obstacle coordinates. The clearance area and the clearance region of the candidate landing node are obtained according to the preset multi-neighbor connected domain analysis method. The landing point of the UAV is selected from a plurality of candidate landing nodes based on the slope value, the straight-line distance of the obstacle, and the clearance area.

3. The low-altitude flight path planning method according to claim 2, characterized in that, The method involves obtaining multiple passable flight areas distributed along the flight direction of the UAV based on the landing point and the UAV's takeoff point using a preset neighborhood search algorithm, including: The terrain area where the UAV is located is determined based on the landing point coordinates and the takeoff point coordinates, and the flight search step size of the UAV is determined based on the slope of the terrain area. The three-dimensional voxel grid data of the UAV during flight is obtained based on the landing point coordinates, the takeoff point coordinates, and the preset BeiDou grid segmentation technology; wherein, the three-dimensional voxel grid data includes multiple grid nodes and the node spatial coordinates corresponding to each grid node; The flight direction and flight constraints of the UAV are determined based on the landing point coordinates, the takeoff point coordinates, and the preset cone angle threshold. For any one of the aforementioned grid nodes; Based on the flight constraints and the flight search step size, a neighborhood search is performed on the three-dimensional voxel mesh data to obtain multiple surrounding mesh nodes of the mesh node in the flight direction. Based on the node spatial coordinates corresponding to each of the multiple surrounding grid nodes, the passable flight area where the multiple surrounding grid nodes are located is obtained.

4. The low-altitude flight path planning method according to claim 3, characterized in that, The step of obtaining the time cost, energy cost, and flight risk of the UAV when flying between any two adjacent passable flight areas based on the flight space coordinates of the candidate flight points and the obstacle coordinates of the obstacles includes: Obtain the first candidate flight node of the first accessible flight area and the second candidate flight node of the second accessible flight area in two adjacent accessible flight areas; The aircraft density between the first candidate flight node and the second candidate flight node is obtained based on the node spatial coordinates. Obtain the wind speed and atmospheric turbulence levels between the first candidate flight node and the second candidate flight node; The time cost of the UAV flying between the first candidate flight node and the second candidate flight node is obtained based on the aircraft density, the wind speed and the atmospheric turbulence. Obtain the altitude difference between the first candidate flight node and the second candidate flight node; Based on the climb drag coefficient, wind drag coefficient, and payload weight of the UAV and the altitude difference, the energy cost of the UAV when flying between the first candidate flight node and the second candidate flight node is obtained. Based on the node spatial coordinates and the obstacle coordinates, candidate obstacles between the first candidate flight node and the second candidate flight node are obtained; The flight risk of the UAV when flying between the first candidate flight node and the second candidate flight node is obtained based on the risk weight coefficient corresponding to the candidate obstacle, the distance between the UAV and the candidate obstacle, the aircraft density and the weight coefficient corresponding to the aircraft density.

5. The low-altitude flight path planning method according to claim 4, characterized in that, The step of determining each flight point in the passable flight area from multiple candidate flight points based on the UAV's flight mission, time cost, energy cost, and flight risk includes: When the UAV flies from the first accessible flight area to the second accessible flight area, for any second candidate flight node in the second accessible flight area: The time weight corresponding to the time cost, the cost weight corresponding to the energy cost, and the risk weight corresponding to the flight risk are determined based on the flight mission of the UAV. Based on the time weight, cost weight, risk weight, time cost, energy cost, and risk weight, the flight cost between the second candidate flight node and the first passable flight area is obtained through a preset weight summation formula. The second candidate flight node with the lowest flight cost is selected from multiple second candidate flight nodes as the flight point of the second passable flight area.

6. The low-altitude flight path planning method according to claim 1, characterized in that, The step of generating the low-altitude flight path corresponding to the UAV based on the flight space coordinates of the flight point, the takeoff point, and the landing point includes: The initial flight path of the UAV is generated based on the takeoff point coordinates of the takeoff point, the landing point coordinates of the landing point, and the flight space coordinates of each of the multiple flight points. The initial flight path is optimized according to the preset curve fitting formula and curvature constraint algorithm to obtain the low-altitude flight path corresponding to the UAV.

7. The low-altitude flight path planning method according to claim 1, characterized in that, The step of generating the low-altitude flight path corresponding to the UAV based on the flight space coordinates of the flight point, the takeoff point, and the landing point further includes: When the UAV flies along the low-altitude flight path, it acquires new obstacle detection results in real time according to the preset detection radius; When a new obstacle is detected, a new path is selected from the low-altitude flight paths based on the real-time positioning of the UAV and the preset path planning radius. Multiple passable flight areas are obtained from the route to be replanned, and a new planned route is generated based on the time cost, energy cost and flight risk between the passable flight areas; The path to be replanned is replaced with the newly planned path to generate the low-altitude flight path corresponding to the UAV.

8. A low-altitude flight path planning system, characterized in that, It includes a landing point optimization module, a accessibility analysis module, a flight cost module, a flight point selection module, and a path planning module; The landing point optimization module is used to obtain the landing area of ​​the UAV and obtain environmental data of each candidate landing point in the landing area, so as to determine the landing point corresponding to the UAV from multiple candidate landing points based on the environmental data. The accessibility analysis module is used to obtain multiple accessible flight areas distributed along the flight direction of the UAV based on the landing point and the UAV's takeoff point using a preset neighborhood search algorithm; wherein each accessible flight area includes multiple candidate flight points; The flight cost module is used to obtain obstacles in each of the passable flight areas, and to obtain the time cost, energy cost and flight risk of the UAV when flying between any two adjacent passable flight areas based on the flight space coordinates of the candidate flight point and the obstacle coordinates of the obstacles; The flight point screening module is used to determine each flight point in the passable flight area from multiple candidate flight points based on the UAV's flight mission, time cost, energy cost, and flight risk. The path planning module is used to generate a low-altitude flight path for the UAV based on the flight space coordinates of the flight point, the take-off point, and the landing point.

9. A terminal device, characterized in that, The system includes a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor, wherein when the processor executes the computer program, it implements a low-altitude flight path planning method as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, include: A stored computer program, wherein, when the computer program is executed, it controls the device containing the computer-readable storage medium to perform a low-altitude flight path planning method as described in any one of claims 1-7.

Images