Low-altitude unmanned aerial vehicle three-dimensional three-dimensional route network planning method

Abstract

This application discloses a three-dimensional route network planning method for low-altitude unmanned aerial vehicles (UAVs), relating to the field of UAV route planning technology. The method includes: acquiring initial data for UAV route planning; extracting the BeiDou grid code corresponding to the initial data for spatial mapping to obtain a gridded distribution of the initial data, wherein the initial data includes UAV flight requirements; dividing the UAV flight requirement grid into a preset number of clusters using a preset clustering algorithm, with the center point of each cluster as a core requirement point; generating direct flight routes connecting all core requirement points using a preset planning algorithm; optimizing the direct flight routes using a preset optimization algorithm to obtain optimized direct flight routes; converting the optimized direct flight routes into a maximum flow tree; obtaining a basic UAV route network based on the maximum flow tree; and expanding the grid channels to obtain a three-dimensional route network. This method enables efficient UAV route network planning.

Description

Technical Field

[0001] This application relates to the field of unmanned aerial vehicle (UAV) route planning technology, and in particular to a method for planning three-dimensional route networks for low-altitude UAVs. Background Technology

[0002] Unmanned Aerial Vehicles (UAVs) are characterized by high flexibility and maneuverability, and are widely used in many fields such as geological exploration, agricultural plant protection, logistics transportation, and disaster relief. With the rapid development of UAV technology and the continuous expansion of its application scale, urban low-altitude airspace resources are becoming increasingly strained. Faced with ever-increasing flight demand, the airspace resource allocation method for planning flight routes for each pair of flight demand points faces severe challenges.

[0003] Most existing technologies focus on optimizing individual UAV routes, attempting to establish a balance between the economy of route length, flight safety, and overall operational efficiency. However, with the rapid growth in flight demand, simply adopting point-to-point direct flight route planning will lead to conflicts between the flight paths of different UAVs in low-altitude airspace, increasing route intersections and potential conflict areas, seriously threatening flight safety. Secondly, traditional UAV route planning methods are mostly based on simplified airspace models for route solving, lacking adaptability to real-world scenarios such as dense building clusters and air traffic flows in urban low-altitude environments.

[0004] The invention patent application with publication number CN118938954A provides a method and system for UAV 3D path planning based on BeiDou grid codes. Based on the UAV's position data and environmental data, a 3D environment model is generated. The position data includes the position codes of the UAV's starting point and the target point. Through A... * The algorithm determines at least one initial path information between the starting point and the target point; the initial path information is optimized using a particle swarm optimization algorithm to obtain the planned path information. However, this method only considers the flight requirements of a single UAV and lacks a unified plan for the flight requirements of multiple UAVs. Summary of the Invention

[0005] Therefore, it is necessary to address the lack of multi-UAV flight demand planning with good scenario adaptability in existing technologies, and to provide a three-dimensional route network planning method for low-altitude UAVs. The specific technical solution is as follows: This application proposes a three-dimensional airway network planning method for low-altitude unmanned aerial vehicles (UAVs), including the following steps: S1: Obtain initial data for UAV route planning, extract the BeiDou grid code corresponding to the initial data for spatial mapping, and obtain the gridded distribution of the initial data. The initial data includes UAV flight requirements. S2: The drone flight demand grid is divided into a preset number of clusters using a pre-defined clustering algorithm, and the center point of each cluster is taken as the core demand point. S3: Use a preset planning algorithm to generate a direct flight route connecting all core demand points, and use a preset optimization algorithm to optimize the direct flight route to obtain the optimized direct flight route; S4: Convert the optimized direct flight routes into a maximum flow tree, and obtain the basic UAV route network based on the maximum flow tree; S5: Expand the grid channels to obtain a three-dimensional airway network.

[0006] Furthermore, the initial data in step S1 also includes at least one of no-fly zones, green areas, bodies of water, and obstacles.

[0007] Furthermore, the preset clustering algorithm in step S2 is the k-means clustering algorithm, and the specific clustering process is as follows: S201: Determine the number of clusters based on the number of UAV flight demand grids, and obtain the center point of each cluster; S202: Calculate the Euclidean distance from each flight demand grid (i.e., the flight demand point) to the center point of each cluster, and assign the flight demand point to the cluster corresponding to the nearest center point; S203: Calculate the average coordinates of all flight demand points in the cluster on each axis according to the coordinate values ​​of the X-axis, Y-axis and Z-axis in the three-dimensional coordinate system, and use this as the updated center point of the cluster; S204: Repeat steps S202 and S203 until the coordinates of the center point of each cluster no longer change, thus completing the clustering.

[0008] Furthermore, the preset planning algorithm mentioned in step S3 is A. * The algorithm for generating direct flight routes connecting all core requirements is as follows: The BeiDou grid code corresponding to the terrain obstacle is converted into a geometric body in three-dimensional space. The boundaries on the X, Y, and Z axes of the terrain obstacle area are set and the no-fly zone and restricted flight altitude layer are marked. A three-dimensional airspace environment model containing terrain obstacles, no-fly zone, restricted flight altitude layer constraints and core requirements is constructed. Design A based on route length, flight energy consumption, and safety distance * The heuristic cost function of the algorithm:

[0009]

[0010]

[0011]

[0012] in, For the total cost, For the actual cost, For heuristic cost estimation, From the starting point of flight to the core demand point The cumulative route distance, From the starting point of flight to the core demand point The cumulative flight energy consumption is related to changes in flight distance and altitude. , These are the energy consumption coefficient per unit distance and the energy consumption coefficient per unit height change, respectively. , The first and the The height of each core requirement point From the starting point of flight to the core demand point The security cost, For safe distance threshold, Core demand points Distance to the nearest obstacle, if but , , , Let be the weight coefficient, and satisfy... ; Taking the flight start point O as the initial node and the flight end point D as the target node, through A * The algorithm searches step by step and outputs a sequence containing multiple discrete path points. ; Based on the order of discrete path points in the sequence, adjacent points are connected by straight lines to form continuous multi-segment broken line trajectories. Starting from the flight origin, each point is sequentially connected to its first adjacent discrete path point with a straight line until it reaches the flight destination, generating direct flight routes between the core requirement point pairs. The coordinates of the points along the direct flight route are converted into corresponding BeiDou grid codes to obtain a BeiDou grid code sequence. Based on the BeiDou grid code sequence, it is checked whether all core demand points are interconnected through the direct flight route. If there is an isolated demand point that is not connected to other core demand points by a direct flight route, it is necessary to search for the nearest connected core demand point to the isolated demand point and generate a direct flight route between the two to obtain a direct flight route that includes all core demand points.

[0013] The preset optimization algorithm mentioned in step S3 is the small snap optimization algorithm, and the specific optimization process is as follows: Based on the physical performance parameters of the UAV, a flight speed threshold is set, and the minimum turning radius at discrete path points at the preset speed is calculated. Construct the minimum snap objective function:

[0014] in, for The fourth derivative, For position functions, Let be the three-dimensional position coordinates of the UAV at time t. The key indicator of whether a flight path is smooth is the snap. The time interval for adjacent discrete path points; Between two adjacent discrete path points, a high-order polynomial is used to fit the flight path trajectory, in order to For example, The higher-order polynomial trajectory model is as follows:

[0015]

[0016]

[0017]

[0018]

[0019] in, These are the polynomial coefficients; For location, For speed, For acceleration, The degree of jerk. For snap; Using the UAV's flight speed and corresponding turning radius as constraints, the objective function is solved to minimize the snap, resulting in a smooth flight path and thus optimizing the direct flight route.

[0020] Furthermore, the maximum flow tree described in step S4 is generated based on the Kruskal algorithm, and the generation process is as follows: Generate the starting point matrix between core requirements.

[0021]

[0022] This indicates that from the core demand point To the core demand point The demand for drone flights, i.e., the weight; and satisfying the condition that... hour, ,when hour, ; The connections between core requirement points are treated as edges. All edges connecting core requirement points are sorted in descending order of weight to form an ordered edge list. Each core requirement point is initialized as an independent set, and each independent set contains only its own core requirement points. At the same time, each independent set is assigned a unique number, and the relationship between the core requirement point and the independent set is recorded. Traverse each edge sequentially according to the ordered edge list. Based on the Beidou grid code, determine whether the core demand points at both ends of the edge belong to different sets. If they belong to different sets, include the edge in the maximum spanning tree and merge the two sets. Stop traversing when the total number of core demand points minus the number of edges in the maximum flow tree equals 1, thus completing the generation of the maximum flow tree.

[0023] The specific content of obtaining the basic UAV route network based on the maximum flow tree in step S4 is as follows: Extract the UAV routes corresponding to all edges in the maximum flow tree, calculate the traffic flow for each route, and expand the route width for each direction to m consecutive BeiDou grids to obtain the basic UAV route network. Furthermore, after step S4, it is necessary to identify the intersections of the UAV flight path and alleviate the dense intersection areas, and design cloverleaf intersections and intersection traffic rules.

[0024] The specific content of identifying the intersections of the UAV flight path and clearing up densely intersecting areas is as follows: Intersections are identified when two or more air routes have consecutive grid overlap areas on the horizontal projection that are larger than a preset area. If the number of intersections within a preset range of airspace exceeds a preset value, the airspace is marked as a dense intersection area and cleared. The specific clearing process is as follows: Based on the traffic flow and airway connectivity criticality of the intersections, some intersections were selected and retained, while the remaining intersections were marked as intersections to be adjusted. Centered on the intersection to be adjusted, search for alternative grids that meet preset conditions, update the position of the intersection to be adjusted to the alternative grids, and reconstruct the relevant routes.

[0025] The specific details of the design of the cloverleaf intersection and the intersection traffic rules are as follows: A cloverleaf intersection is designed with the geometric center of the intersection as the origin. The cloverleaf intersection allocates the flight paths of UAVs going straight, turning left, and turning right to independent flight levels or dedicated passage ramps. Flight paths of different directions do not intersect horizontally or vertically. The dedicated passage ramps include a ring ramp, and a transition area, an emergency level, and an emergency landing area are also set up. The passage rules of the intersection include: before entering the intersection, UAVs pass through the transition area and adjust their flight altitude to the corresponding flight level. UAVs going straight and turning right fly in the same flight level, UAVs turning left occupy a separate flight level, UAVs going straight maintain a straight flight along their original direction, and UAVs turning left and right turn through the ring ramps respectively. The emergency level and emergency landing area are used for emergency flight and emergency landing of UAVs.

[0026] Beneficial effects: This application is based on the BeiDou grid code, and uses the k-means clustering algorithm and A... * The algorithm and the Kruskal algorithm are used to obtain the basic UAV route network. At the same time, cloverleaf intersections and intersection traffic rules are designed to achieve efficient planning of the flight requirements of multiple UAVs and have good scene adaptability. Attached Figure Description

[0027] Figure 1 This is a flowchart of a three-dimensional airway network planning method for low-altitude unmanned aerial vehicles (UAVs) according to an embodiment of this application; Figure 2 This is a schematic diagram of the drone flight requirements in the embodiments of this application; Figure 3 This is a schematic diagram of the center point of the cluster generated in the embodiments of this application; Figure 4 This is a schematic diagram of the direct flight path of the UAV in the embodiments of this application; Figure 5 This is a comparison diagram of UAV flight paths in the embodiments of this application; Figure 6 This is a schematic diagram of a cloverleaf-shaped intersection in an embodiment of this application; Figure 7 This is a schematic diagram of the north-south UAV intersection layer in this application embodiment; Figure 8 This is a schematic diagram of the east-west UAV intersection layer in the embodiments of this application; Figure 9 The figure shows the experimental results in the embodiments of this application. Detailed Implementation

[0028] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0029] Example 1 This embodiment provides a three-dimensional flight path network planning method for low-altitude unmanned aerial vehicles (UAVs), the flowchart of which is shown below. Figure 1 As shown, it includes the following steps: S1: Obtain initial data for UAV route planning, extract the BeiDou grid code corresponding to the initial data for spatial mapping, and obtain the gridded distribution of the initial data. The initial data includes UAV flight requirements. It should be noted that the initial data mentioned in step S1 also includes at least one of no-fly zones, green areas, bodies of water, and obstacles.

[0030] In one specific embodiment, the BeiDou grid is a seven-level grid (7m*7m*7m). Based on the BeiDou grid code, a three-dimensional spatial coordinate system with a horizontal X-axis, a vertical Y-axis, and a vertical Z-axis is constructed. The spatial position information represented by the BeiDou grid code can be converted into coordinate parameters in the three-dimensional coordinate system.

[0031] S2: The drone flight demand grid is divided into a preset number of clusters using a pre-defined clustering algorithm, and the center point of each cluster is taken as the core demand point. It should be noted that the preset clustering algorithm in step S2 is the k-means clustering algorithm, and the specific clustering process is as follows: S201: Determine the number of clusters based on the number of UAV flight demand grids, and obtain the center point of each cluster; S202: Calculate the Euclidean distance from each flight demand grid (i.e., the flight demand point) to the center point of each cluster, and assign the flight demand point to the cluster corresponding to the nearest center point; The expression for the Euclidean distance from each flight demand point to the center point is:

[0032] in, For Euclidean distance , These are the coordinates of the drone's flight requirement point and center point on the horizontal X-axis, respectively. , These are the coordinates of the drone's flight requirement point and center point on the vertical Y-axis, respectively. , These are the coordinates of the drone's flight requirement point and center point on the vertical Z-axis, respectively. The Euclidean distance from the flight requirement point to the center point can be calculated using the expression.

[0033] S203: Calculate the average coordinates of all flight demand points in the cluster on each axis according to the coordinate values ​​of the X-axis, Y-axis and Z-axis in the three-dimensional coordinate system, and use this as the updated center point of the cluster; S204: Repeat steps S202 and S203 until the coordinates of the center point of each cluster no longer change, thus completing the clustering.

[0034] Figure 2 This is a diagram illustrating the flight requirements of drones. Figure 3 This is a schematic diagram of the center point of the generated cluster, with the pentagram shape representing the location of the center point.

[0035] S3: Use a preset planning algorithm to generate a direct flight route connecting all core demand points, and use a preset optimization algorithm to optimize the direct flight route to obtain the optimized direct flight route; The preset planning algorithm in step S3 is the A* algorithm. The process of generating direct flight routes connecting all core demand points is as follows: The Beidou grid code corresponding to terrain obstacles (water bodies, green spaces, high-rise building complexes, etc.) is converted into a geometric object in three-dimensional space. The boundaries on the X, Y, and Z axes of the terrain obstacle area are set and the no-fly zone and restricted flight altitude layer are marked. A three-dimensional airspace environment model containing terrain obstacles, no-fly zone, restricted flight altitude layer constraints and core requirements is constructed. The heuristic cost function for the A* algorithm is designed based on route length, flight energy consumption, and safety distance.

[0036]

[0037]

[0038]

[0039] in, For the total cost, For the actual cost, For heuristic cost estimation, From the starting point of flight to the core demand point The cumulative route distance, From the starting point of flight to the core demand point The cumulative flight energy consumption is related to changes in flight distance and altitude. , These are the energy consumption coefficient per unit distance and the energy consumption coefficient per unit height change, respectively. , The first and the The height of each core requirement point From the starting point of flight to the core demand point The security cost, For safe distance threshold, Core demand points Distance to the nearest obstacle, if but , , , Let be the weight coefficient, and satisfy... ; Taking the flight start point O as the initial node and the flight end point D as the target node, through A * The algorithm searches step by step and outputs a sequence containing multiple discrete path points. ; Based on the order of discrete path points in the sequence, adjacent points are connected by straight lines to form continuous multi-segment broken line trajectories. Starting from the flight origin, each point is sequentially connected to its first adjacent discrete path point with a straight line until it reaches the flight destination, generating direct flight routes between the core requirement point pairs. The coordinates of the points along the direct flight route are converted into corresponding BeiDou grid codes. These BeiDou grid codes are then arranged in chronological order according to the flight route to obtain a BeiDou grid code sequence. Based on this sequence, it is verified whether all core requirement points are interconnected via direct flight routes. If an isolated requirement point exists without a direct flight route connection to other core requirement points, the nearest connected core requirement point is searched, and a direct flight route between them is generated. This results in a direct flight route encompassing all core requirement points. Figure 4 As shown.

[0040] The preset optimization algorithm is the small snap optimization algorithm, and the specific optimization process is as follows: Based on the physical performance parameters of the UAV, a flight speed threshold is set, and the minimum turning radius at the preset speed at discrete path points is calculated. If the turning radius corresponding to the discrete path point is less than the minimum turning radius, the position of the discrete path point is corrected. Construct the minimum snap objective function:

[0041] in, for The fourth derivative, For position functions, Let be the three-dimensional position coordinates of the UAV at time t. The key indicator of whether a flight path is smooth is the snap. The time interval for adjacent discrete path points; Between two adjacent discrete path points, a high-order polynomial is used to fit the flight path trajectory, in order to For example, The higher-order polynomial trajectory model is as follows:

[0042]

[0043]

[0044]

[0045]

[0046] in, These are the polynomial coefficients; For location, For speed, For acceleration, The degree of jerk. For snap; Using the UAV's flight speed and corresponding turning radius as constraints, the optimized route is ensured to conform to the UAV's physical performance. The objective function is solved to minimize the snap, and the polynomial coefficients are obtained. The continuity of position, velocity, acceleration, and jerk at the path points of adjacent polynomials is ensured to avoid trajectory jumps or abrupt changes, thereby obtaining a smooth route trajectory and completing the optimization of the direct flight route.

[0047] For waypoints requiring correction within the flight path, their corresponding BeiDou grid codes are searched based on their 3D coordinate parameters. The positions of these waypoints are adjusted and converted to new BeiDou grid positions according to the flight path optimization requirements. During this adjustment process, it is necessary to ensure that the BeiDou grid containing the new waypoint is not located in no-fly zones, terrain obstacles, or other restrictive areas, and that all continuity constraints are met. The corresponding polynomial coefficients are then recalculated to generate the adjusted direct flight path. The BeiDou grid codes corresponding to the adjusted waypoints are arranged in the order of the flight path to form a smoothed direct flight path between the core requirement points. This path is then output as a BeiDou grid sequence, generating a direct flight path covering each pair of core requirement points.

[0048] S4: Convert the optimized direct flight routes into a maximum flow tree, and obtain the basic UAV route network based on the maximum flow tree; It should be noted that the maximum flow tree described in step S4 is generated based on the Kruskal algorithm, and the generation process is as follows: Generate the origin-destination (OD) matrix between the core requirements.

[0049]

[0050] This indicates that from the core demand point To the core demand point The demand for drone flights, i.e., the weight; and satisfying the condition that... hour, ,when hour, ; The connections between core requirement points are used as edges. All edges connecting core requirement points are sorted in descending order of weight to form an ordered edge list. Each core requirement point is initialized as an independent set. Each independent set contains only its own core requirement points. At the same time, each independent set is assigned a unique number and the relationship between the core requirement point and the independent set is recorded. The uniqueness of the Beidou grid code is used to avoid confusion of node identifiers. Each edge is traversed sequentially from the ordered edge list. Based on the BeiDou grid code, it is determined whether the core requirement points at both ends of the edge belong to different sets. If they belong to different sets, adding the edge will not cause a loop in the maximum flow tree, so the edge is added to the maximum flow tree, and the two sets are merged. The traversal stops when the total number of core requirement points minus the number of edges in the maximum flow tree equals 1, thus completing the generation of the maximum flow tree. If they belong to the same set, adding the edge will violate the acyclic property of the maximum flow tree, causing route redundancy conflicts.

[0051] The specific content of obtaining the basic UAV route network based on the maximum flow tree in step S4 is as follows: Extract all UAV routes corresponding to edges in the maximum flow tree, calculate the traffic flow for each route, ensure that the traffic flow matches the grid capacity, expand the route width for each direction to m consecutive BeiDou grids, construct standardized route channels with multi-UAV parallel passage capability, and obtain the basic UAV route network, such as... Figure 5 As shown, Figure 5 The left side shows the UAV route map covering all demand points, and the right side shows the UAV route network generated by the Kruskal algorithm.

[0052] It should be noted that after step S4, it is necessary to identify the intersections of the UAV flight path and clear the dense intersection areas, and design cloverleaf intersections and intersection traffic rules.

[0053] The specific content of identifying the intersections of the UAV flight path and clearing up densely intersecting areas is as follows: Intersections are identified when two or more air routes have consecutive grid overlap areas on the horizontal projection that are larger than a preset area. If the number of intersections within a preset range of airspace exceeds a preset value, the airspace is marked as a dense intersection area and cleared. The specific clearing process is as follows: Based on the traffic flow and airway connectivity criticality of the intersections, some intersections are selected and retained, while the remaining intersections are marked as intersections to be adjusted. Using the intersections to be adjusted as the center, alternative grids that meet the preset conditions are searched, the positions of the intersections to be adjusted are updated to alternative grids, and the relevant airways are reconstructed.

[0054] In one specific embodiment, the process of managing and alleviating dense intersection areas includes: (1) The Beidou grid code at the center of the intersection is used as the representative location of the intersection. The Beidou grid code that defines the boundary of the intersection is used for spatial analysis. All air routes passing through the intersection and their corresponding traffic information are recorded in detail. The distribution of all intersections is statistically analyzed. The number of intersections per unit space is calculated. When the number exceeds the preset value, the area is marked as a dense intersection area and needs to be adjusted.

[0055] (2) For densely intersecting areas, considering factors such as traffic flow through the intersections, airway connectivity criticality, and flight constraint adaptability, priority is given to retaining intersections with high traffic flow, high connectivity criticality, and high flight constraint adaptability in densely populated areas. The remaining intersections are marked as intersections to be adjusted, and their BeiDou grid attributes are recorded simultaneously to provide data support for subsequent intersection site selection adjustments.

[0056] (3) Update the location of the intersection to be adjusted: Using the original grid code of the intersection to be adjusted as the center, and relying on the coordinate system of the Beidou grid, search for candidate grids that meet the following conditions: ① The grid itself and its surrounding adjacent grids are not no-fly zones and are free of obstacles; ② The distance to the grids of surrounding intersections is greater than the preset grid; ③ The three-dimensional space corresponding to the grid meets the performance parameters such as the minimum turning radius and altitude constraints of the UAV. Through the standardized spatial division of the Beidou grid, the search for candidate locations does not require additional adaptation to terrain data, and the grid's built-in constraint information can be directly called. ④ Update the location of the intersection to be adjusted to the candidate grid. ⑤ Adjust the air routes connected to the new intersection, reconstruct the Beidou grid sequence of the relevant air routes, and re-verify the air route traffic to ensure that the grid's traffic capacity can meet the air route traffic requirements, avoiding additional airspace carrying capacity verification after adjustment.

[0057] When flight routes are adjusted, new intersection density issues may arise, requiring multiple iterations to achieve global convergence. For the adjusted new flight routes, the BeiDou grid sequence is re-traversed, and whether the number of intersections per unit space exceeds a threshold is used as the criterion for determining dense areas.

[0058] (2) If a new dense area is identified, repeat the de-icing process to adjust the intersection position. During the iteration process, all adjustments are based on the Beidou grid code, and the grid sequence and constraint association information of the airway are updated synchronously to ensure that the airway and airspace constraints always match after each round of adjustment, without the need for additional constraint compatibility verification.

[0059] (3) Perform multiple iterations. The iteration terminates when there are no densely intersecting areas in the global airway and the safety constraints are met, and the flow rate within the grid is lower than the preset passage threshold. When the global airway reaches the optimal state, output the adjusted UAV flight route network.

[0060] The specific details of the design of the cloverleaf intersection and the intersection traffic rules are as follows: (1) Design a cloverleaf intersection with the geometric center of the intersection as the origin. The cloverleaf intersection allocates the flight paths of UAVs going straight, turning left, and turning right to independent flight levels or dedicated ramps. Flight paths of different directions do not intersect horizontally or vertically, eliminating the potential for traffic conflicts between UAVs traveling in different directions. A schematic diagram of the intersection design is shown below. Figure 6 As shown.

[0061] (2) Determine the entry and exit nodes of the intersection based on the direction of flight. Select the intersection points located at the boundary of the buffer zone and about to enter the buffer zone along the direction of flight, and mark them as entry nodes. Mark the boundary intersection points about to leave the buffer zone as exit nodes. Each entry node and exit node corresponds to a unique BeiDou grid code.

[0062] (3) Eliminate lateral conflicts between UAVs by vertically layering the airspace and constructing aerial intersections. Based on the center of the intersection, divide the airspace into k main flight layers + 1 emergency layer, with the height difference between each layer being a preset safe height h, to ensure sufficient operational space for UAVs to avoid vertical collisions. The emergency layer is located at the bottom of the intersection and is only used for temporary avoidance by malfunctioning UAVs; regular passage is prohibited.

[0063] (4) A transition zone is set up in the area where the UAV enters and leaves the intersection. Before entering the intersection, the UAV needs to complete the altitude adjustment in the transition zone. Virtual guide points are set up in the transition zone to avoid sudden vertical ascent and descent. By setting up virtual main lines and circular turning ramps at each altitude level, the UAV can turn in all directions. All routes are calibrated by Beidou grid and have no physical structure.

[0064] (5) With the center of the intersection as the intersection point, the main line adopts a straight design and extends to 1 km outside the intersection to connect with the outer route. The width of the main line is a Beidou grid, which allows UAVs to pass in formation in one direction. A virtual boundary is set at the edge of the main line. The navigation system on the UAV can identify the boundary in real time to avoid deviation. At the same time, two warning nodes for entering the intersection are set at 200 meters and 100 meters away from the intersection on the main line to trigger the UAV turning preparation procedure.

[0065] (6) Set up ramps to accommodate the left-turn needs of UAVs, achieving turning through circular detours to avoid intersection conflicts. Each circular ramp forms a virtual circular path with a width of three Beidou grids, centered on the center of the intersection, connecting with the adjacent main line. Set the turning radius according to the UAV flight parameters to avoid excessive centrifugal force. The UAV's flight speed within the ramp needs to be reduced to 70% of the main line speed, and the speed limit is enforced through the navigation system; the altitude remains unchanged during the turning process, and only the horizontal attitude is adjusted.

[0066] (7) In addition to the circular left-turn ramp, straight and right-turn paths need to be laid out simultaneously to form a complete traffic network. By laying out the right-turn path outside the main line, the UAV can enter the right-turn path from the main line in advance and only needs to complete a 90° turn to access the target main line without having to detour through the circular ramp, thus improving the efficiency of right turns.

[0067] (8) Set up an emergency landing zone below the center of the intersection. When the UAV malfunctions, the navigation system will guide it to land nearby. At the same time, reserve a space in the emergency layer for the malfunctioning UAV to hover temporarily.

[0068] The rules for passing through intersections include: before entering an intersection, drones pass through a transition area and adjust their flight altitude to the corresponding flight level. In north-south direction flights, drones going straight and turning right fly in the same flight level, while drones turning left occupy a separate flight level. Drones going straight maintain a straight flight along their original heading, while drones turning left and right use a roundabout to turn. The emergency level and emergency landing area are used for emergency flights and emergency landings of drones. Figure 7 and Figure 8 The diagram shows the layered layout of drone intersections in the north-south and east-west directions.

[0069] S5: Expand the grid channels to obtain a three-dimensional airway network.

[0070] It should be noted that, based on flight requirements, routes and intersections represented by single-layer grid code sequences are expanded into three-dimensional channels with multi-layer grid codes in the longitudinal profile. This leads to the planning of a low-altitude three-dimensional route network, specifically including: Collect flight demand data across all core demand points, including flight counts per unit time, drone type, and payload. Based on this data, classify flight demands into different levels and, in conjunction with drone physical performance parameters, determine the grid layer division criteria for each demand level on the longitudinal profile. For each airway's single-layer grid code sequence, the grid is expanded longitudinally according to its corresponding flight demand level. Based on the original airway's single-grid code, adjacent BeiDou grids are added upwards and downwards, forming multiple continuous grid code sequences in the longitudinal profile. During the expansion process, it is necessary to ensure that the newly added grids meet safety constraints, and that the grids at each altitude level remain consistent with the original airway grid in the horizontal direction, together forming the three-dimensional grid channel of the airway. The expanded route grid sequence and intersection grid sequence are subjected to coherence verification to ensure that the grid layers of adjacent routes can be smoothly connected longitudinally, and that the grid layers of intersections in each direction match the grid layers of all connected routes. A global review is performed to ensure that the airspace corresponding to all grid codes meets safety constraints and to eliminate potential spatial conflicts. A low-altitude three-dimensional airway network covering all core requirements has been constructed to enable UAVs with different flight needs to pass through three-dimensional space in an orderly manner, and to achieve efficient planning of low-altitude UAV airways based on Beidou grid codes.

[0071] Example 2 This embodiment verifies the proposed three-dimensional flight path network planning method for low-altitude unmanned aerial vehicles (UAVs) through simulation, including the following: In the actual verification process, a specific region was selected as the experimental area. This region possesses diverse geographical features, including urban built-up areas, industrial parks, and waterfront areas, resulting in a complex and typical airspace environment that can effectively simulate various scenarios that UAVs may encounter in actual operation. Firstly, based on the actual airspace usage requirements of this region, multiple UAV operation demand points were pre-defined. Following the low-altitude UAV three-dimensional route network planning method based on the BeiDou grid airspace proposed in this application, the generation of UAV operation routes for the BeiDou grid airspace was completed through steps such as demand point clustering, route generation, and intersection design.

[0072] Figure 9 The experiment presented the optimal UAV route network between the core requirement points, generated based on the BeiDou grid airspace.

[0073] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A three-dimensional flight path network planning method for low-altitude unmanned aerial vehicles (UAVs), characterized in that, Includes the following steps: S1: Obtain initial data for UAV route planning, extract the BeiDou grid code corresponding to the initial data for spatial mapping, and obtain the gridded distribution of the initial data. The initial data includes UAV flight requirements. S2: The drone flight demand grid is divided into a preset number of clusters using a pre-defined clustering algorithm, and the center point of each cluster is taken as the core demand point. S3: Use a preset planning algorithm to generate a direct flight route connecting all core demand points, and use a preset optimization algorithm to optimize the direct flight route to obtain the optimized direct flight route; S4: Convert the optimized direct flight routes into a maximum flow tree, and obtain the basic UAV route network based on the maximum flow tree; S5: Expand the grid channels to obtain a three-dimensional airway network.

2. The method for planning a three-dimensional flight path network for low-altitude unmanned aerial vehicles according to claim 1, characterized in that, The initial data in step S1 also includes at least one of no-fly zones, green areas, bodies of water, and obstacles.

3. The three-dimensional flight path network planning method for low-altitude unmanned aerial vehicles according to claim 1, characterized in that, The preset clustering algorithm mentioned in step S2 is the k-means clustering algorithm, and the specific clustering process is as follows: S201: Determine the number of clusters based on the number of UAV flight demand grids, and obtain the center point of each cluster; S202: Calculate the Euclidean distance from each flight demand grid, i.e., the flight demand point, to the center point of each cluster, and assign the flight demand point to the cluster corresponding to the nearest center point; S203: Calculate the average coordinates of all flight demand points in the cluster on each axis according to the coordinate values ​​of the X-axis, Y-axis and Z-axis in the three-dimensional coordinate system, and use this as the updated center point of the cluster; S204: Repeat steps S202 and S203 until the coordinates of the center point of each cluster no longer change, thus completing the clustering.

4. The three-dimensional flight path network planning method for low-altitude unmanned aerial vehicles according to claim 1, characterized in that, The preset planning algorithm mentioned in step S3 is A. * The algorithm for generating direct flight routes connecting all core requirements is as follows: The BeiDou grid code corresponding to the terrain obstacle is converted into a geometric body in three-dimensional space. The boundaries on the X, Y, and Z axes of the terrain obstacle area are set and the no-fly zone and restricted flight altitude layer are marked. A three-dimensional airspace environment model containing terrain obstacles, no-fly zone, restricted flight altitude layer constraints and core requirements is constructed. Design A based on route length, flight energy consumption, and safety distance * The heuristic cost function of the algorithm: in, For the total cost, For the actual cost, For heuristic cost estimation, From the starting point of flight to the core demand point The cumulative route distance, From the starting point of flight to the core demand point The cumulative flight energy consumption is related to changes in flight distance and altitude. , These are the energy consumption coefficient per unit distance and the energy consumption coefficient per unit height change, respectively. , The first and the The height of each core requirement point From the starting point of flight to the core demand point The security cost, For safe distance threshold, Core demand points Distance to the nearest obstacle, if but , , , Let be the weight coefficient, and satisfy... ; Taking the flight start point O as the initial node and the flight end point D as the target node, through A * The algorithm searches step by step and outputs a sequence containing multiple discrete path points. ; According to the order of discrete path points in the sequence, connect adjacent points with straight lines to form a continuous multi-segment broken line trajectory. Starting from the flight start point, connect it with the first adjacent discrete path point with straight lines in turn until it is connected to the flight end point, generating a direct flight route between the core requirement point pairs. The coordinates of the points along the direct flight route are converted into corresponding BeiDou grid codes to obtain a BeiDou grid code sequence. Based on the BeiDou grid code sequence, it is checked whether all core demand points are interconnected through the direct flight route. If there are isolated demand points that are not connected to the core demand points by a direct flight route, it is necessary to search for the nearest connected core demand point to the isolated demand point and generate a direct flight route between the two to obtain a direct flight route that includes all core demand points.

5. The three-dimensional flight path network planning method for low-altitude unmanned aerial vehicles according to claim 4, characterized in that, The preset optimization algorithm mentioned in step S3 is the minimum snap optimization algorithm, and the specific optimization process is as follows: Based on the physical performance parameters of the UAV, a flight speed threshold is set, and the minimum turning radius at discrete path points at the preset speed is calculated. Construct the minimum snap objective function: in, for The fourth derivative, For position functions, Let be the three-dimensional position coordinates of the UAV at time t. The key indicator of whether a flight path is smooth is the snap. The time interval for adjacent discrete path points; Between two adjacent discrete path points, a high-order polynomial is used to fit the flight path trajectory, in order to For example, The higher-order polynomial trajectory model is as follows: in, These are the polynomial coefficients; For location, For speed, For acceleration, For rapidity, For jump; Using the UAV's flight speed and corresponding minimum turning radius as constraints, the objective function is solved to minimize the snap, resulting in a smooth flight path and thus optimizing the direct flight route.

6. The three-dimensional flight path network planning method for low-altitude unmanned aerial vehicles according to claim 1, characterized in that, The maximum flow tree described in step S4 is generated based on the Kruskal algorithm, and the generation process is as follows: Generate the starting point matrix between core requirements. This indicates that from the core demand point To the core demand point The demand for drone flights, i.e., the weight; and satisfying the condition that... hour, ,when hour, ; The connections between core requirement points are treated as edges. All edges connecting core requirement points are sorted in descending order of weight to form an ordered edge list. Each core requirement point is initialized as an independent set, and each independent set contains only its own core requirement points. At the same time, each independent set is assigned a unique number, and the relationship between the core requirement point and the independent set is recorded. Traverse each edge sequentially according to the ordered edge list. Based on the Beidou grid code, determine whether the core demand points at both ends of the edge belong to different sets. If they belong to different sets, include the edge in the maximum spanning tree and merge the two sets. Stop traversing when the total number of core demand points minus the number of edges in the maximum flow tree equals 1, thus completing the generation of the maximum flow tree.

7. A three-dimensional flight path network planning method for low-altitude unmanned aerial vehicles according to claim 6, characterized in that, The specific content of obtaining the basic UAV route network based on the maximum flow tree in step S4 is as follows: Extract all UAV routes corresponding to all edges in the maximum flow tree, calculate the traffic flow of each route, and expand the route width of each direction to m consecutive BeiDou grids to obtain the basic UAV route network.

8. A three-dimensional flight path network planning method for low-altitude unmanned aerial vehicles according to claim 1, characterized in that, After step S4, the process also includes identifying intersections along the UAV flight path, clearing up dense intersection areas, and designing cloverleaf intersections and intersection traffic rules.

9. A three-dimensional flight path network planning method for low-altitude unmanned aerial vehicles according to claim 8, characterized in that, The specific content of identifying the intersections of the UAV flight path and clearing up densely intersecting areas is as follows: Intersections are identified when two or more air routes have consecutive grid overlap areas on the horizontal projection that are larger than a preset area. If the number of intersections within a preset range of airspace exceeds a preset value, the airspace is marked as a dense intersection area and cleared. The specific clearing process is as follows: Based on the traffic flow and airway connectivity criticality of the intersections, some intersections were selected and retained, while the remaining intersections were marked as intersections to be adjusted. Centered on the intersection to be adjusted, search for alternative grids that meet preset conditions, update the position of the intersection to be adjusted to the alternative grids, and reconstruct the relevant routes.

10. A three-dimensional flight path network planning method for low-altitude unmanned aerial vehicles according to claim 9, characterized in that, The specific details of the design of the cloverleaf intersection and the intersection traffic rules are as follows: With the geometric center of the intersection as the origin, a cloverleaf intersection is designed. The cloverleaf intersection allocates the flight paths of UAVs going straight, turning left, and turning right to independent flight layers or dedicated passage ramps. There are no horizontal or vertical intersections between flight paths of different directions. The dedicated passage ramps include a ring ramp and also include a transition area, an emergency layer, and an emergency landing area. The rules for passing through intersections include: before entering an intersection, drones pass through a transition area to adjust their flight altitude to the corresponding flight level. Drones going straight and turning right fly in the same flight level, while drones turning left occupy a separate flight level. Drones going straight maintain a straight flight along their original heading, while drones turning left and right use a roundabout to turn. The emergency level and emergency landing area are used for emergency flights and emergency landings of drones.

Images