Three-dimensional airspace rasterization management method and system based on multistage dynamic division

Abstract

The invention belongs to the technical field of three-dimensional airspace management, and provides a three-dimensional airspace rasterizing management method and system based on multistage dynamic division, and the method comprises the steps: dividing height layers for a three-dimensional airspace, carrying out the rasterizing division of the three-dimensional airspace in each height layer, and carrying out the rasterizing division of the three-dimensional airspace; a live-action three-dimensional model of a three-dimensional airspace of a target area is constructed through multi-source data fusion, passable grids are marked, selection priority values of the passable grids are calculated through multi-source data fusion, when an aircraft is about to enter the three-dimensional airspace, route dynamic priority values of the grids are initialized, and a basic planning route of the aircraft is generated. Detecting the collision risk of the aircraft on the grids in the basic planning route, dynamically adjusting the route dynamic priority value, updating to obtain the actual planning route of the aircraft, monitoring the space between the aircrafts in the three-dimensional airspace in real time, and judging whether the collision risk exists between the aircrafts; and if there is a collision risk signal, triggering an aircraft obstacle avoidance and path re-planning strategy, and dynamically adjusting and updating the grid side length.

Description

Technical Field

[0001] The present invention belongs to the technical field of three-dimensional airspace management, and specifically relates to a three-dimensional airspace grid management method and system based on multi-level dynamic division. Background Art

[0002] With the rapid development of aviation technology and the gradual advancement of the low-altitude opening policy, flight activities in the three-dimensional airspace are becoming increasingly frequent and diverse. Various flight tasks such as unmanned aerial vehicle (UAV) logistics distribution, low-altitude tourism sightseeing, agricultural and forestry plant protection operations, and emergency rescue flights are constantly emerging, and the three-dimensional airspace has gradually become a new growth point in the aviation field.

[0003] In the prior art, the three-dimensional airspace is mostly divided by fixed altitude layers and simple grid divisions, without considering the performance of aircraft and mission requirements, resulting in difficulty in balancing airspace utilization efficiency and safety. However, the existing methods lack refined evaluation of the spatial distribution of obstacles, and cannot accurately identify high-risk areas. Traditional route planning algorithms only consider path distance, do not incorporate real-time airspace status, and lack a multi-aircraft cooperative collision avoidance mechanism. As airspace traffic increases, the risk of time and space overlap of aircraft in shared grids intensifies, and it is difficult for the prior art to detect and eliminate collision hazards in real time.

[0004] In addition, the three-dimensional airspace environment changes dynamically, but the grid accuracy of the existing grid planning is fixed and cannot be adjusted dynamically according to the collision risk. The existing technology's fusion of data such as obstacle distribution, airspace density, and aircraft trajectories remains at a simple superposition level, without forming a scientific quantitative evaluation system, ignoring the actual impact of obstacle positions on flight paths, resulting in a disconnection between the planning results and actual safety requirements. In the prior art, the grid side length is fixed, cannot be adjusted dynamically according to the collision risk, lacks the collection and analysis of collision risk signals during flight, and cannot optimize grid division and planning strategies through historical data.

[0005] In view of the above problems, the present invention proposes a three-dimensional airspace grid management method and system based on multi-level dynamic division. Summary of the Invention

[0006] In order to make up for the deficiencies of the prior art and solve at least one of the technical problems proposed in the background art.

[0007] The technical solution adopted by the present invention to solve the technical problems is: A three-dimensional airspace grid management method based on multi-level dynamic division, including:

[0008] Dividing altitude layers for the three-dimensional airspace, and making grid planning for the three-dimensional airspace according to the altitude layers to divide grids;

[0009] Obtain the grid obstacle index for each grid and mark the passable grids. For the passable grids, calculate the grid selection priority value through multi-source data fusion. When the aircraft is about to enter the three-dimensional airspace, initialize the route dynamic priority value of the grid according to the selection priority value, and generate the actual planned route for the aircraft;

[0010] Specifically, based on any divided grid, determine whether there is an overlapping part within the grid with the static no-fly zone, temporary control zone, and flight restriction zone. If there is no overlapping part, calculate the obstacle index of the grid through the voxel position weighting method;

[0011] Compare the obstacle index of the grid with the obstacle ratio threshold. If the obstacle index of a single grid is less than the obstacle ratio threshold, mark the grid as a passable grid;

[0012] The selection priority data includes the obstacle index of the grid, the real-time aircraft density in the altitude layer where the grid is located, and the spatial distribution dispersion of the aircraft;

[0013] Count the number of aircraft in the upper altitude layer at the current moment, and combine with the altitude layer volume to calculate the real-time aircraft density in the altitude layer where the grid is located. Obtain the position coordinates of all aircraft in the upper altitude layer at the current moment for data processing to get the centroid coordinates of all aircraft in the altitude layer. Use the Euclidean distance formula to process the position coordinates of each aircraft and the centroid coordinates to obtain the spatial distribution dispersion of the aircraft in the altitude layer where the grid is located at the current moment;

[0014] Obtain the selection priority data of the grid at the current moment. After integrating the selection priority data, perform reverse normalization processing and normalization processing respectively. Subjectively assign weights to the processed selection priority data based on the analytic hierarchy process, calculate the subjective weight of the selection priority data, objectively assign weights to the selection priority data through the entropy weight method, and calculate the objective weight. Fuse and calculate the subjective weight and the objective weight to obtain the final fusion weight;

[0015] Perform weighted fusion calculation on the processed selection priority data at the current moment according to the corresponding final fusion weight to obtain the selection priority value of the passable grid;

[0016] When the aircraft is about to enter the three-dimensional airspace during flight, mark the aircraft as a to-be-flown aircraft, and select the starting flight altitude layer, starting flight node, and target flight node for the to-be-flown aircraft;

[0017] Define a route analysis period, and set the route dynamic priority value for the passable grid within the route analysis period, with the initial value being the selection priority value of the corresponding passable grid at the starting point of the route analysis period;

[0018] Within the selected altitude layer type, combined with the dynamic priority value of the route in the grid, use the improved A* algorithm to generate a basic planned route for the aircraft to be flown;

[0019] Obtain the actual planned routes of all aircraft in actual flight, and conduct a coincidence analysis with the basic planned route of the aircraft to be flown to determine whether there are overlapping grids. If there are, calculate the time periods for the aircraft to be flown to pass through each overlapping grid, marked as the passing time periods for the aircraft to be flown, and calculate the time periods for the aircraft in actual flight with overlapping grids to pass through the overlapping grids, marked as the passing time periods for actual flight;

[0020] Compare the passing time periods for the aircraft to be flown and the passing time periods for actual flight within the grid to obtain the time interval between the passing time periods for the aircraft to be flown and the passing time periods for actual flight in the grid;

[0021] Obtain the time interval between the passing time periods for the aircraft to be flown and the passing time periods for actual flight in the grid. If it is less than or equal to the safety time interval threshold, mark the grid as a collision risk grid, process the data of the time intervals between the passing time periods for the aircraft to be flown and the passing time periods for actual flight in all collision risk grids, and adjust the dynamic priority value of the route for all grids with collision risks;

[0022] According to the adjusted dynamic priority value of the route, use the improved A* algorithm to re-plan the route for the aircraft to be flown, update the basic planned route until there are no grids that can be marked as collision risk grids, clear the mark of the aircraft to be flown for the aircraft, and mark the basic planned route as the actual planned route of the aircraft;

[0023] Continuously monitor the distance between aircraft in the three-dimensional airspace during flight, judge the collision risk in real time. If there is a collision risk, generate a collision risk signal and trigger the aircraft obstacle avoidance and path re-planning strategy, and optimize the management of the grid division parameters based on the number of generated collision risk signals

[0024] Specifically, define a flight detection period, and in the flight detection period, obtain the distances of all aircraft in each altitude layer on the plane in real time. If the distance between two aircraft is less than or equal to the distance safety threshold, generate a collision risk signal;

[0025] Calculate the expected position of the aircraft based on the actual planned path, compare and calculate the position deviation value between the actual position and the expected position of the aircraft, and mark the aircraft with the larger position deviation value among the two aircraft as the obstacle avoidance aircraft;

[0026] Taking the actual position of the obstacle avoidance aircraft at the current moment as the starting flight node and the target flight node remaining unchanged, use the improved A* algorithm to re-search for a path within a range of 5 path nodes in radius to conduct local path re-planning for the obstacle avoidance aircraft. If there is no feasible path in the local path re-planning, then conduct global re-planning and update the actual planned path of the obstacle avoidance aircraft;

[0027] Obtain the number of times of generating collision risk signals during the flight detection period, and optimize the management of grid division parameters;

[0028] Adopt a proportional-integral adjustment model to dynamically adjust the grid side length of the three-dimensional airspace grid planning division in the target area, calculate the adjustment value of the grid side length of the three-dimensional airspace grid planning division, adjust the grid side length, update and optimize the grid division parameters in the three-dimensional airspace grid planning according to the adjusted grid side length, and apply it to the physical airspace.

[0029] A three-dimensional airspace grid management system based on multi-level dynamic division includes the following modules:

[0030] Initial planning module: Divide height layers for the three-dimensional airspace, and perform grid planning for the three-dimensional airspace according to the height layers to divide grids;

[0031] Priority analysis module: Obtain the obstacle index of each grid and mark the passable grids. For the passable grids, calculate the grid selection priority value through multi-source data fusion;

[0032] Route planning module: When the aircraft is about to enter the three-dimensional airspace, initialize the route dynamic priority value of the grid according to the selection priority value, and generate an actual planned route for the aircraft;

[0033] Parameter optimization module: Continuously monitor the distance between aircraft in the three-dimensional airspace during flight, judge the collision risk in real time. If there is a collision risk, generate a collision risk signal, and trigger the aircraft obstacle avoidance and path replanning strategy, and optimize the management of grid division parameters based on the number of times of generating collision risk signals.

[0034] The beneficial effects of the present invention are as follows:

[0035] 1. The present invention significantly improves the accuracy and safety of route planning. By dividing height layers and rasterizing the three-dimensional airspace, accurately calculating the grid obstacle index by combining multi-source data, accurately marking the passable grids, planning a path for the aircraft that better fits the actual airspace environment, calculating the selection priority value based on the selection priority data fusion and initializing the route dynamic priority value, enabling the route planning to fully consider various key factors. Before flight, identify the collision risk grids in advance through the coincidence period analysis and adjust the route dynamic priority value to achieve route replanning, effectively avoiding potential collision risks. This dynamic and accurate planning method allows the aircraft to avoid dangerous areas and choose the optimal route to fly, greatly reducing the probability of flight accidents and providing a solid guarantee for the flight safety in the three-dimensional airspace.

[0036] 2. The present invention enhances the adaptability and planning self-optimization ability during flight. During flight, it continuously monitors the distance between aircraft and judges the collision risk in real time. Once a risk exists, it can quickly generate a collision risk signal and trigger obstacle avoidance and path replanning strategies, perform local or global replanning for the obstacle-avoiding aircraft, and timely adjust the flight path to ensure the safe flight of the aircraft. At the same time, according to the number of collision risk signals generated during the flight detection period, a proportional-integral adjustment model is used to dynamically optimize the grid division parameters and achieve adaptive adjustment of the grid side length. This self-optimization mechanism based on actual flight data enables the planning method to continuously adapt to the complex and changeable three-dimensional airspace environment, continuously improve the planning quality, and provide strong support for the efficient and orderly development of three-dimensional airspace flight activities. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] The present invention will be further described below with reference to the accompanying drawings.

[0038] Figure 1 is a flowchart of the steps of the three-dimensional airspace grid management method based on multi-level dynamic division according to an embodiment of the present invention;

[0039] Figure 2 is a system module architecture diagram of the three-dimensional airspace grid management system based on multi-level dynamic division according to an embodiment of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0040] In order to make the technical means, creative features, achieved purposes and effects of the present invention easy to understand, the present invention will be further described below in conjunction with specific embodiments.

[0041] Embodiment 1

[0042] Please refer to Figure 1 as shown. The three-dimensional airspace grid management method based on multi-level dynamic division according to an embodiment of the present invention includes the following steps:

[0043] S1: Divide height layers for the three-dimensional airspace, perform grid division on the three-dimensional airspace within each height layer to obtain a number of grids, and construct a real three-dimensional model of the target area three-dimensional airspace through multi-source data fusion;

[0044] According to the low-altitude classification standard of the International Civil Aviation Organization (ICAO) and the aircraft performance parameters, divide the three-dimensional airspace of the target area into several horizontally parallel height layers. The height layer types include the micro-unmanned aircraft layer, the medium-sized unmanned aircraft layer, and the manned aircraft layer. Each height layer type is divided into several height layers according to different stratification step lengths, and each height layer corresponds to a specific aircraft type and flight mission;

[0045] Specifically, the height range of the micro - UAV layer is (0m, 50m], and the layer - division step size of the micro - UAV layer is set to 10 meters. The height range of the medium - sized UAV layer is (50m, 300m], and the layer - division step size of the medium - sized UAV layer is set to 50 meters. The height range of the manned aircraft layer is (300m, 1000m], and the layer - division step size of the manned aircraft layer is set to 100 meters;

[0046] It should be noted that the micro - UAV layer is applicable to consumer - grade UAVs with a wheelbase less than or equal to 25 cm, which can perform short - range aerial photography and agricultural plant protection tasks. In the micro - UAV layer, due to the low flight height and strong mobility of the aircraft, fine layer - division is required to avoid low - altitude obstacles such as hitting trees. The medium - sized UAV layer is applicable to logistics UAVs and mapping UAVs. When setting the layer - division step size, the endurance efficiency and obstacle - avoidance space need to be taken into account. The manned aircraft layer is applicable to electric vertical take - off and landing aircraft and helicopters. When setting the layer - division step size, sufficient safety intervals and emergency heights need to be reserved;

[0047] The calculation formula for the number k of any height h (true height, unit: meter) in the corresponding height - layer type is:

[0048]

[0049] where h step represents the layer - division step size of the height - layer type where the height h is located;

[0050] If the height h is 0 meters, it is defaultly classified into layer 1. If it exceeds 1000 meters, it will be incorporated into the medium - and - high - altitude control airspace and does not participate in the three - dimensional airspace grid planning;

[0051] For each height - layer, the height - layer is evenly divided into several square grids in the horizontal direction. The side lengths of the grids divided in the same height - layer are the same in the horizontal direction, obtaining the grid planning of the three - dimensional airspace of the target area;

[0052] Combining the height - layer division and the horizontal grid coordinates, a three - level coding system is designed, adopting a three - level nested structure of height - layer type - height - layer number - horizontal grid coordinates to encode the latitude, longitude, and height information step by step;

[0053] Specifically, in the three - level coding system, the first level is the height - layer type, where the micro - UAV layer is numbered 1, the medium - sized UAV layer is numbered 2, and the manned aircraft layer is numbered 3. The second level is the height - layer number, which retains three - digit characters of the number k of the height in the corresponding height - layer type. The third level is the horizontal grid coordinates, which are composed of the row number and column number each retaining four - digit characters;

[0054] Among them, the row number and column number are mapped from latitude and longitude. The southwest corner vertex of the target area in the horizontal direction at the same altitude layer is selected as the coordinate origin, and the relative distances between the latitude and longitude of the center point of the square grid in the horizontal direction and the coordinate origin are calculated. Based on the relative distances, data processing is performed with the side length of the square grid in the horizontal direction to obtain the corresponding row number and column number;

[0055] Use high-resolution commercial satellites to cover the target area to obtain satellite images, and acquire basic geographic information such as terrain and building outlines to provide a macroscopic base for the 3D model. Deploy multi-rotor drones equipped with five-lens tilt cameras and lidar to collect centimeter-level accurate point cloud data, and capture details such as building facades and high-voltage towers. For complex areas, manually supplement texture data through drones to obtain texture images, reducing model voids and streaking phenomena. Use ground lidar to supplement scanning in key areas to obtain ground lidar data, enhancing the local accuracy and anti-interference ability of the model. Georeference the images taken by the multi-rotor drones using ground control points (GCPs) to eliminate distortion errors, and achieve tone uniformity and seamless stitching of multi-angle aerial images through image processing software to generate a global orthophoto image;

[0056] Integrate satellite images, point cloud data, and ground lidar data, combine the Kalman filtering algorithm to eliminate multi-source data conflicts, construct a 3D data cube under a unified spatio-temporal coordinate system, perform deep classification on the point cloud data in the 3D data cube through deep learning algorithms to distinguish ground object types such as vegetation, buildings, and roads, combine the multi-view stereo vision (MVS) algorithm to generate a dense point cloud, and generate a preliminary 3D model through Poisson surface reconstruction. Use an AI model to automatically match texture images with the preliminary 3D model, optimize texture details, and repair streaked areas to achieve high-fidelity model rendering and obtain a real-scene 3D model;

[0057] It should be noted that the role of this step is to establish a hierarchical and grid-based framework for the 3D airspace, provide a spatial benchmark for subsequent obstacle detection and path planning, and through a high-precision 3D model, achieve refined modeling of obstacles and support grid passability assessment;

[0058] S2: Based on the real-scene 3D model, calculate the obstacle index within the grid, mark the passable grids, and for the passable grids, calculate the selection priority value of the passable grids through multi-source data fusion;

[0059] Based on any grid divided, obtain the ranges of static no-fly zones, temporary control zones, and restricted flight zones within the 3D airspace of the target area, and judge whether there are overlapping parts within the grid with the ranges of static no-fly zones, temporary control zones, and restricted flight zones;

[0060] If there are overlapping parts, mark the grid as an impassable grid;

[0061] If there is no overlapping part, obtain the real-scene three-dimensional model, extract the obstacles within the grid from the real-scene three-dimensional model, and further analyze the impact of the spatial distribution of the obstacles within the grid on flight safety. Through the three-dimensional model voxelization technology, the grid is evenly divided into m×m×m voxels, and the obstacle index O of the grid is calculated by the voxel position weighting method. f ;

[0062] Specifically, mark the center coordinates of each voxel in the grid as (x a , y b , z c ), where a, b, and c are the index numbers of the voxel in the x-axis, y-axis, and z-axis directions in the grid, respectively, and the counting starts from the lower left corner of the grid and increases sequentially. Mark the coordinates of the center point of the grid as (x v , y v , z v ), and define the voxel position weight w (a,b,c) as:

[0063]

[0064] where d (a,b,c) represents the distance between the center point of the voxel and the center point of the grid, and the calculation formula is:

[0065]

[0066] It should be noted that the closer the voxel is to the center point of the grid, the greater the impact of the obstacles in the voxel on the flight path, the higher the voxel position weight w v , the closer it is to 1, and the weight of the obstacles located at the edge approaches 0 and can be ignored;

[0067] Count the voxels occupied by obstacles, and perform weighted summation calculation in combination with the voxel volume and the voxel position weight w v to obtain the weighted obstacle volume of the grid, and perform a ratio process on the weighted obstacle volume and the grid volume to calculate the obstacle index O of the grid f ;

[0068] Compare the obstacle index O f of the grid with the obstacle occupancy threshold;

[0069] If the obstacle index of a single grid is greater than or equal to the obstacle occupancy threshold, mark the grid as an impassable grid;

[0070] If the obstacle index of a single grid is less than the obstacle occupancy threshold, mark the grid as a passable grid;

[0071] For passable grids, combining the obstacle index of the upper grid at the current moment, the real-time aircraft density of the altitude layer where the grid is located, and the spatial distribution dispersion degree of the aircraft, a quantitative analysis is carried out on the selection priority of the grid when performing aircraft path planning;

[0072] Specifically, define a time sliding window with the current moment as the end point. As time progresses, the time sliding window slides according to the change of the current moment, and the length of the time sliding window is Δt;

[0073] Deploy a low-altitude surveillance radar array and an ADS-B receiving station network. At the current moment, count the aircraft falling into the altitude layer where the grid is located within the time sliding window, and statistically mark it as the number of aircraft N in the altitude layer at the current moment k , where k represents the number of the altitude layer within the corresponding altitude layer type;

[0074] Combined with the statistically obtained number of aircraft N k and the altitude layer volume V, through the formula:

[0075]

[0076] Calculate the real-time aircraft density ρ of the altitude layer where the grid is located k , where the altitude layer volume V is calculated from the total number of grids in the altitude layer and the grid volume;

[0077] By deploying a low-altitude surveillance radar array and an ADS-B receiving station network, collect and record the flight information of the aircraft entering the altitude layer. Among them, the flight information includes the position information of all aircraft in the altitude layer at the current moment, and the time when the aircraft enters the altitude layer. Mark the serial number i for the aircraft according to the time when the aircraft enters the altitude layer, define the position coordinates of the aircraft as the plane coordinates of the aircraft center point in the altitude layer, and mark it as

[0078] Perform a summation and averaging process on the position coordinates of all aircraft in the altitude layer at the current moment to obtain the centroid coordinates of all aircraft in the altitude layer. Use the Euclidean distance formula to calculate the distance between each aircraft position coordinate and the centroid coordinate, and perform a summation and averaging process on the obtained distances to obtain the spatial distribution dispersion degree D of the aircraft in the altitude layer where the grid is located at the current moment k ;

[0079] It should be noted that the spatial distribution dispersion degree D k reflects the degree of dispersion of the aircraft's spatial distribution in the altitude layer. The larger the spatial distribution dispersion degree, the more dispersed the aircraft distribution, and the higher the risk of collision when a new aircraft enters the altitude layer. The smaller the spatial distribution dispersion degree, the more concentrated the aircraft distribution, and the lower the risk of collision when a new aircraft enters the altitude layer;

[0080] Obtain the selection priority data of the grid at the current moment, where the selection priority data includes the obstacle index O of the grid f , the real-time aircraft density ρ at the height layer where the grid is located k and the spatial distribution dispersion D of the aircraft k , and perform data fusion processing on the selection priority data;

[0081] Specifically, integrate the selection priority data at the current moment, integrate the obstacle indexes of all passable grids into an obstacle index group, integrate the real-time aircraft densities of all height layers in any height layer type into an aircraft density group, and integrate the spatial distribution dispersions of the aircraft in all height layers in any height layer type into a distribution dispersion group;

[0082] Perform inverse normalization processing on the obstacle index group and the aircraft density group respectively, perform Z-Score normalization processing on the distribution dispersion group, map the selection priority data at the current moment to the value range of [0,1], and mark the obstacle index and the real-time aircraft density after inverse normalization processing as O′ f and ρ′ k , and mark the spatial distribution dispersion of the aircraft after Z-Score normalization processing as D′ k ;

[0083] Subjectively weight the selection priority data based on the Analytic Hierarchy Process (AHP), construct a pairwise comparison judgment matrix for the selection priority data, calculate the subjective weight of the selection priority data through matrix operations on the judgment matrix, objectively weight the selection priority data through the entropy weight method, calculate the information entropy of the selection priority data, perform data processing on the information entropy, calculate the objective weight of the selection priority data, and use the weighted average method to fuse the subjective weight and the objective weight of the selection priority data to obtain the final fusion weight of the selection priority data;

[0084] Based on any passable grid, perform weighted fusion calculation on the selection priority data O′ f , ρ′ k and D′ k at the current moment after Z-Score normalization or inverse normalization processing according to the corresponding final fusion weight to obtain the selection priority value of the passable grid;

[0085] It should be noted that the selection priority value is in the value range of [0,1]. The closer the selection priority value is to 0, the higher the comprehensive risk of the grid is indicated, and the lower the selection priority is. The closer the selection priority value is to 1, the lower the comprehensive risk of the grid is indicated, and the higher the selection priority is;

[0086] It should be noted that the function of this step is to quantify the grid risk, distinguish between passable and impassable areas, and provide a basis for path planning priority. Through data fusion technology, multi-dimensional risks such as obstacles and airspace density are comprehensively evaluated to improve the scientific nature of grid evaluation;

[0087] S3: When the aircraft is about to enter and fly in the three-dimensional airspace, initialize the route dynamic priority value of the grid based on the selection priority value of the grid. Use the improved A* algorithm to generate the basic planned route of the aircraft according to the aircraft type and the route dynamic priority value. Detect the collision risk of the aircraft on the grid within the basic planned route, dynamically adjust the route dynamic priority value based on the collision risk, and update to obtain the actual planned route of the aircraft based on the dynamic adjustment;

[0088] When the aircraft is about to enter and fly in the three-dimensional airspace, mark the aircraft as a to-be-flown aircraft, and select the altitude layer type for the to-be-flown aircraft based on the to-be-flown aircraft type;

[0089] For any altitude layer within the altitude layer type, obtain the selection priority values of all passable grids within the altitude layer and perform a summation process to obtain the selection priority sum of the altitude layer. Sort the altitude layers from largest to smallest according to the selection priority sum, and use the altitude layer with the largest selection priority sum as the starting flight altitude layer of the to-be-flown aircraft;

[0090] Select the starting flight node for the to-be-flown aircraft within the starting flight altitude layer. Obtain the coordinates of the center point of the to-be-flown aircraft at the current moment, and use the Euclidean distance formula to calculate the distances between the coordinates of the center point of the to-be-flown aircraft and the coordinates of the center points of each passable grid within the starting flight altitude layer. Use the grid center point with the shortest distance from the coordinates of the center point of the to-be-flown aircraft as the starting flight node of the to-be-flown aircraft;

[0091] Similarly, obtain the target position coordinates of the to-be-flown aircraft, use the Euclidean distance formula to calculate the distances between the target position coordinates of the to-be-flown aircraft and the coordinates of the center points of each passable grid within the starting flight altitude layer, and use the grid center point with the shortest distance from the target position coordinates of the to-be-flown aircraft as the target flight node of the to-be-flown aircraft;

[0092] Taking the time when the starting flight node and the target flight node of the to-be-flown aircraft are determined as the starting point, define a route analysis period. Set the route dynamic priority value for all passable grids within the altitude layer type during the route analysis period, and assign the initial value of the route dynamic priority value as the selection priority value of the corresponding passable grid at the starting point of the route analysis period;

[0093] Within the selected altitude layer type, use the improved A* algorithm to generate the basic planned route for the to-be-flown aircraft;

[0094] Specifically, the center point of each passable grid within the altitude layer type is defined as a path node, and the path node is assigned with three-dimensional coordinate information, a parent node pointer, an actual cost g(n), a heuristic estimated cost h(n), a comprehensive cost f(n), and a status flag;

[0095] Among them, the role of the status flag is to mark whether the path node has been explored. The initial status flag is unexplored. The parent node pointer points to the previous path node of the path node during the path search process for final path backtracking. The actual cost g(n) represents the cumulative cost from the starting flight node to the path node n. The initial value is set to infinity indicating that it has not been explored and is gradually updated during the search process. The heuristic estimated cost h(n) represents the estimated cost from the path node n to the target flight node and is calculated using the Manhattan distance formula. The formula for the comprehensive cost f(n) is:

[0096]

[0097] Among them, S p (n) represents the dynamic priority value of the grid route corresponding to the path node n. α represents the weight coefficient emphasizing the influence of the dynamic priority value of the route. β represents the weight coefficient controlling the smoothness of the heading. θ(n) represents the recommended heading angle of the aircraft to be flown at the path node. θ pre (n) represents the heading angle of the parent node of the path node n;

[0098] Set up a data list and initialize the data list. The data list includes an open list and a closed list. The open list is used to store nodes to be explored. Initialize the open list. Initially, only the starting flight node is added to this list and g(starting node) = 0 is set. The closed list is used to store nodes that have been explored. Initialize the closed list. Initially, it is empty;

[0099] Perform a search loop for path nodes within the altitude layer type. When the open list is not empty, continuously select the optimal path node with the smallest f(n) value from the open list as the current node to be expanded, and move the current node to be expanded from the open list to the closed list;

[0100] It should be noted that selecting the optimal path node with the smallest f(n) value from the open list ensures that the path branch with the lowest cost is explored first. Among them, the dynamic priority value S of the route in the f(n) formula p (n) makes the algorithm tend to select passable grids with low risk. At the same time, the smoothness of the planned path is constrained by β·|θ(n) - θ pre (n)| to avoid frequent turning;

[0101] If the current node to be expanded is not the target node, mark the status of the current node to be expanded as explored, and analyze all adjacent path nodes of the current node to be expanded;

[0102] Specifically, based on any adjacent path node, if the adjacent path node is already in the closed list, it means that the adjacent path node and its subsequent paths have been fully explored, and skip the adjacent path node without analysis;

[0103] If the adjacent path node is not in the closed list and not in the open list, add the adjacent path node to the open list, calculate and update the actual cost from the starting flight node through the current path node to the adjacent path node. Based on the updated actual cost, calculate and update the comprehensive cost to ensure that the path cost is always optimal;

[0104] If the current node to be expanded is the target flight node, it indicates that a complete path from the starting node to the target node has been successfully found. The algorithm terminates the current search process and enters the subsequent path generation stage. The algorithm generates the final path by backtracking the parent node pointer. Starting from the target flight node, backtrack along the parent node pointer step by step to the starting node, record the nodes passed through in sequence to form a node sequence, and reverse the node sequence to obtain the complete path from the starting flight node to the target flight node, which is the basic planned route generated by the improved A* algorithm for the aircraft to be flown;

[0105] Obtain the aircraft that are in flight among all altitude levels within the altitude level type, mark them as actually flying aircraft, obtain the actual planned routes of all actually flying aircraft, perform a coincidence analysis on the passable grids within all actual planned routes and the passable grids within the basic planned route of the aircraft to be flown, and determine whether there are overlapping grids. If not, mark the basic planned route of the aircraft to be flown as the actual planned route of the aircraft to be flown;

[0106] If there are overlapping grids, calculate the time period for the aircraft to be flown to pass through each overlapping grid, mark it as the to-be-flown passing time period of the overlapping grid, calculate the time period for the actually flying aircraft with overlapping grids to pass through the overlapping grid, and mark it as the actually-flying passing time period of the overlapping grid;

[0107] Based on any overlapping grid, compare the to-be-flown passing time period and the actually-flying passing time period within the overlapping grid, and obtain the time interval between the to-be-flown passing time period and the actually-flying passing time period. Among them, the time interval represents the difference between the starting point of the later time sequence period and the key point of the earlier time sequence period. If the to-be-flown passing time period and the actually-flying passing time period overlap, the time interval is negative;

[0108] Compare the time interval between the to-be-flown passing time period and the actually-flying passing time period with the safety time interval threshold;

[0109] If the time interval is greater than the safety time interval threshold, it is determined that there is no collision risk between the to-be-flown aircraft and the corresponding actually-flown aircraft within the overlapping grid;

[0110] If the time interval is less than or equal to the safety time interval threshold, it is determined that there is a collision risk between the to-be-flown aircraft and the corresponding actually-flown aircraft within the overlapping grid, and the overlapping grid is marked as a collision risk grid;

[0111] If the overlapping grid is marked as a collision risk grid, adjust the dynamic priority value of the route of the collision risk grid;

[0112] Specifically, based on the safety time interval threshold, obtain the time intervals between the to-be-flown passing time period and the actually-flown passing time period within all collision risk grids for normalization processing to obtain the time collision risk value R of the collision risk grid col , through the formula:

[0113] S′ phx =S phx ·(1 - δ·R col )

[0114] Adjust the dynamic priority value of the route of the collision risk grid, where S′ phx represents the dynamically adjusted dynamic priority value of the route, S phx represents the dynamic priority value of the route before dynamic adjustment, δ represents the penalty coefficient, and update the dynamic priority value of the collision risk grid route:

[0115] S phx =S′ phx

[0116] After completing the update of the dynamic priority value of the route, clear the overlapping grid mark and the collision risk grid mark of the grid;

[0117] Adjust the dynamic priority value of the route for all grids with collision risks. According to the adjusted dynamic priority value of the route, use the improved A* algorithm to re-plan the route for the to-be-flown aircraft and update the basic planned route;

[0118] Continue to analyze the collision risk between the to-be-flown aircraft and the actually-flown aircraft based on the updated basic planned route until there are no grids that can be marked as collision risk grids, stop the route analysis period, mark the basic planned route as the actual planned route of the to-be-flown aircraft, control the to-be-flown aircraft to fly in the three-dimensional airspace according to the actual planned route, and clear the to-be-flown aircraft mark of the aircraft and mark the aircraft as an actually-flown aircraft;

[0119] It should be noted that the function of this step is to generate a safe and efficient initial flight route based on grid priorities, ensure that the aircraft selects a path with low risk and high suitability, avoid the risk of multi-aircraft collision through dynamic adjustment of priority values and replanning mechanisms, and ensure the safety of airspace passage;

[0120] S4: When the aircraft is flying in the three-dimensional airspace according to the actual planned flight route, the distance between aircraft in the three-dimensional airspace is monitored in real time to determine whether there is a collision risk between aircraft. If there is a collision risk, a collision risk signal is generated and the aircraft obstacle avoidance and path replanning strategy is triggered. Based on the number of times the collision risk signal is generated, the grid side length in the grid-based planning is dynamically adjusted and updated;

[0121] After the grid-based planning of the three-dimensional airspace in the target area is put into use, a flight detection period is defined. The end point of the flight detection period is also the starting point of the next flight detection period, and the duration of all flight detection periods is the same;

[0122] During the flight detection period, using the Euclidean distance formula, the distance between the center points of all aircraft in each altitude layer on the plane is obtained in real time. If the distance between the center points of two aircraft is less than or equal to the distance safety threshold, a collision risk signal is generated, and the position deviation value between the actual positions of the two aircraft at the current moment and the expected positions of the aircraft flying based on the actual planned path is calculated and compared. The aircraft with the larger position deviation value among the two aircraft is marked as the obstacle avoidance aircraft;

[0123] Plan an obstacle avoidance path for the obstacle avoidance aircraft. Taking the actual position of the obstacle avoidance aircraft at the current moment as the starting flight node and the target flight node remaining unchanged, the improved A* algorithm is used to re-search for a path within a radius of 5 path nodes to perform local path replanning for the obstacle avoidance aircraft. If there is no feasible path in the local path replanning, the improved A* algorithm is used to perform global replanning for the obstacle avoidance aircraft, update the dynamic priority value of the grid route, and update the actual planned path of the obstacle avoidance aircraft;

[0124] Based on any flight detection period, obtain the number of times the collision risk signal is generated during the flight detection period. If the number of times the collision risk signal is greater than the number threshold, it indicates that there is a collision risk problem in the grid-based planning of the three-dimensional airspace in the target area. Based on the number of times the collision risk signal is generated during the flight detection period, the grid side length divided by the grid-based planning of the three-dimensional airspace in the target area is dynamically adjusted;

[0125] Specifically, using the proportional-integral adjustment model, through the formula:

[0126]

[0127] Calculate the adjustment value ΔL of the grid side length divided by the grid-based planning of the three-dimensional airspace, where K prepresents the ratio adjustment coefficient, τ represents the non-linear adjustment factor, which makes the adjustment amplitude approach saturation as the error increases, avoiding violent oscillations, EMA t represents the error smoothing value of the difference between the number of generated collision risk signals and the number threshold within the detection period, calculated by the exponential moving average method, and η represents the adaptive adjustment factor that balances safety requirements and computational efficiency;

[0128] Adjust the grid side length according to the calculated adjustment value of the grid side length divided by the three-dimensional airspace gridization plan, update the three-dimensional airspace gridization plan according to the adjusted grid side length, and apply it to the physical airspace;

[0129] It should be noted that the function of this step is to monitor the operation status of the airspace in real time, quickly respond to sudden collision risks, ensure flight safety, dynamically optimize the airspace division accuracy by adaptively adjusting the grid side length, balance safety requirements and computational efficiency, and achieve system self-optimization;

[0130] The technical solution of the embodiment of the present invention is as follows: Divide height layers for the three-dimensional airspace, perform gridization on the three-dimensional airspace within each height layer to obtain a number of grids, construct a real three-dimensional model of the target area three-dimensional airspace through multi-source data fusion, calculate the obstacle index within the grid based on the real three-dimensional model, mark the passable grids, for the passable grids, calculate the selection priority value of the passable grids through multi-source data fusion, when the aircraft is about to enter the three-dimensional airspace during flight, initialize the route dynamic priority value of the grid based on the selection priority value of the grid, use the improved A* algorithm to generate the basic planned route of the aircraft according to the aircraft type and the route dynamic priority value, detect the collision risk of the aircraft on the grid within the basic planned route, dynamically adjust the route dynamic priority value based on the collision risk, and update the actual planned route of the aircraft based on the dynamic adjustment. When the aircraft flies in the three-dimensional airspace according to the actual planned route, monitor the distance between aircraft in the three-dimensional airspace in real time, judge whether there is a collision risk between aircraft, if there is a collision risk, generate a collision risk signal and trigger the aircraft obstacle avoidance and path replanning strategy, and dynamically adjust and update the grid side length in the gridization plan based on the number of generated collision risk signals.

[0131] Embodiment 2

[0132] As Figure 2 shown, the three-dimensional airspace gridization management system based on multi-level dynamic division described in the embodiment of the present invention includes the following modules:

[0133] Initial planning module: Divide height layers for the three-dimensional airspace and perform gridization planning for the three-dimensional airspace according to the height layers to divide grids;

[0134] Priority analysis module: Obtain the grid obstacle index of each grid and mark the passable grids. For the passable grids, calculate the grid selection priority value through multi-source data fusion;

[0135] Route planning module: When the aircraft is about to enter the three-dimensional airspace, initialize the route dynamic priority value of the grid according to the selection priority value, and generate an actual planned route for the aircraft;

[0136] Parameter optimization module: Continuously monitor the distance between aircraft in the three-dimensional airspace during flight, judge the collision risk in real time. If there is a collision risk, generate a collision risk signal, and trigger the aircraft obstacle avoidance and path replanning strategy, and optimize and manage the grid division parameters based on the number of times the collision risk signal is generated.

[0137] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited by the above embodiments. The above embodiments and the descriptions in the specification only illustrate the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A three-dimensional airspace rasterization management method based on multi-level dynamic partitioning, characterized in that: Including: Dividing altitude layers for the three-dimensional airspace, and making a rasterized plan for the three-dimensional airspace according to the altitude layers to divide the grids; Obtaining the obstacle index of each grid and marking the passable grids. For the passable grids, calculating the grid selection priority value through multi-source data fusion. When the aircraft is about to enter the three-dimensional airspace, initializing the route dynamic priority value of the grid according to the selection priority value to generate the actual planned route for the aircraft; Continuously monitoring the distance between aircraft in the three-dimensional airspace during flight, judging the collision risk in real time. If there is a collision risk, generating a collision risk signal, and triggering the aircraft obstacle avoidance and path replanning strategy, and optimizing and managing the grid division parameters based on the number of generated collision risk signals.

2. The three-dimensional airspace rasterization management method based on multi-level dynamic partitioning according to claim 1, wherein: The method of marking the passable grids is as follows: Based on any of the divided grids, judging whether there is an overlapping part with the static no-fly zone, temporary control zone and restricted flight zone within the grid. If there is no overlapping part, calculating the obstacle index of the grid by the voxel position weighting method; Comparing the obstacle index of the grid with the obstacle ratio threshold. If the obstacle index of a single grid is less than the obstacle ratio threshold, marking the grid as a passable grid.

3. The three-dimensional airspace rasterization management method based on multi-level dynamic partitioning according to claim 1, wherein: The calculation method of the selection priority value is as follows: Obtaining the selection priority data of the upper grid at the current moment. After integrating the selection priority data, performing reverse normalization processing and normalization processing respectively. Subjectively weighting the processed selection priority data based on the analytic hierarchy process to calculate the subjective weight of the selection priority data. Objectively weighting the selection priority data by the entropy weight method to calculate the objective weight of the selection priority data, and fusing and calculating the subjective weight and the objective weight to obtain the final fusion weight; Performing weighted fusion calculation on the processed selection priority data at the current moment according to the corresponding final fusion weight to obtain the selection priority value of the passable grid.

4. The three-dimensional airspace rasterization management method based on multi-level dynamic partitioning according to claim 3, wherein: The acquisition method of the selection priority data is as follows: The selection priority data includes the obstacle index of the grid, the real-time aircraft density in the altitude layer where the grid is located, and the spatial distribution dispersion degree of the aircraft; Counting the number of aircraft in the altitude layer at the current moment, and combining with the altitude layer volume to calculate the real-time aircraft density in the altitude layer where the grid is located. Obtaining the position coordinates of all aircraft in the altitude layer at the current moment for data processing to obtain the centroid coordinates of all aircraft in the altitude layer. Using the Euclidean distance formula to process the position coordinates of each aircraft and the centroid coordinates to obtain the spatial distribution dispersion degree of the aircraft in the altitude layer where the grid is located at the current moment.

5. The three-dimensional airspace rasterization management method based on multi-level dynamic partitioning according to claim 1, wherein: The method of generating the actual planned route for the aircraft is as follows: Obtaining the time interval between the to-be-flown through period and the actually-flown through period of the grid. If it is less than or equal to the safety time interval threshold, marking the grid as a collision risk grid, processing the time intervals between the to-be-flown through periods and the actually-flown through periods in all collision risk grids, and adjusting the route dynamic priority value of all grids with collision risks. According to the adjusted dynamic priority value of the route, use the improved A* algorithm to re-plan the route for the aircraft to be flown, update the basic planned route until there are no grids that can be marked as collision risk grids, clear the mark of the aircraft to be flown, and mark the basic planned route as the actual planned route of the aircraft.

6. The three-dimensional airspace rasterization management method based on multi-level dynamic partitioning according to claim 5, characterized in that: The way to obtain the time interval between the to-be-flown passing period and the actually-flown passing period is as follows: Obtain the actual planned routes of all actually-flown aircraft, perform coincidence analysis with the basic planned route of the aircraft to be flown, judge whether there are coincident grids. If there are, calculate the time period for the aircraft to be flown to pass through each coincident grid, mark it as the to-be-flown passing period, calculate the time period for the actually-flown aircraft with coincident grids to pass through the coincident grid, and mark it as the actually-flown passing period; Compare the to-be-flown passing period and the actually-flown passing period in the grid to obtain the time interval between the to-be-flown passing period and the actually-flown passing period of the grid.

7. The three-dimensional airspace rasterization management method based on multi-level dynamic partitioning according to claim 6, wherein: The way to obtain the basic planned route is as follows: When the aircraft is about to enter the three-dimensional airspace for flight, mark the aircraft as an aircraft to be flown, and select the starting flight altitude layer, starting flight node and target flight node for the aircraft to be flown; Define a route analysis period, set the dynamic priority value of the route for the passable grids within the route analysis period, and assign the initial value as the selection priority value of the corresponding passable grid at the starting point of the route analysis period; Within the selected altitude layer type, combine the dynamic priority value of the route of the grid, and use the improved A* algorithm to generate the basic planned route for the aircraft to be flown.

8. The three-dimensional airspace rasterization management method based on multi-level dynamic partitioning according to claim 1, characterized in that: The aircraft obstacle avoidance and path re-planning strategy is as follows: Define a flight detection period, and in the flight detection period, obtain the distances of all aircraft in each altitude layer on the plane in real time. If the distance between two aircraft is less than or equal to the distance safety threshold, generate a collision risk signal; Calculate the expected position of the aircraft based on the actual planned path, compare and calculate the position deviation value between the actual position and the expected position of the aircraft, and mark the aircraft with the larger position deviation value among the two aircraft as the obstacle avoidance aircraft; Take the actual position of the obstacle avoidance aircraft at the current moment as the starting flight node, keep the target flight node unchanged, and use the improved A* algorithm to re-search for a path within a range of 5 path nodes in radius to perform local path re-planning for the obstacle avoidance aircraft. If there is no feasible path in the local path re-planning, perform global re-planning and update the actual planned path of the obstacle avoidance aircraft.

9. The three-dimensional airspace rasterization management method based on multi-level dynamic partitioning according to claim 1, wherein: The way to optimize and manage the grid division parameters is as follows: Obtain the number of times of generating collision risk signals during the flight detection period, and optimize and manage the grid division parameters; Adopt a proportional-integral adjustment model to dynamically adjust the grid side length of the grid division of the three-dimensional airspace grid of the target area, calculate the adjustment value of the grid side length of the three-dimensional airspace grid division, adjust the grid side length, update and optimize the grid division parameters in the three-dimensional airspace grid division according to the adjusted grid side length, and apply it to the physical airspace.

10. A three-dimensional airspace raster management system based on multi-level dynamic partitioning, characterized in that: It includes the following modules: Initial planning module: Divide altitude layers for the three-dimensional airspace, and perform grid division for the three-dimensional airspace according to the altitude layers; Priority analysis module: Obtain the grid obstacle index for each grid and mark the passable grids. For the passable grids, calculate the grid selection priority value through multi-source data fusion; Route planning module: When the aircraft is about to enter the three-dimensional airspace, initialize the route dynamic priority value of the grid according to the selection priority value, and generate the actual planned route for the aircraft; Parameter optimization module: Continuously monitor the distance between aircraft in the three-dimensional airspace during flight, judge the collision risk in real time. If there is a collision risk, generate a collision risk signal, and trigger the aircraft obstacle avoidance and path replanning strategy, and optimize and manage the grid division parameters based on the number of times the collision risk signal is generated.

Images