A system for urban airspace mapping and UAV path planning

Abstract

This invention relates to the field of UAV airspace planning technology and discloses an urban airspace map construction and UAV path planning system. The system collects building outline point clouds, real-time wind direction, and controlled area boundary data through an airspace feature fusion module, correlates spatial locations, and calculates obstacle distribution density to generate a three-dimensional dynamic feature set of the airspace. A constraint modeling module divides the airspace into grid units, marks building height differences and wind direction abrupt change zones, forming a constraint label set. A trajectory segmentation processing module analyzes the turning angles of trajectory points in non-abrupt change zones to evaluate path continuity. A conflict node screening module identifies potential conflict nodes with turning deviations exceeding thresholds and located in wind direction abrupt change zones. A path optimization output module calculates the shortest obstacle avoidance path between nodes, ultimately generating an airspace map and globally optimized path planning results, effectively improving the safety and efficiency of UAV flights in urban airspace.

Description

Technical Field

[0001] This invention relates to the field of UAV airspace planning technology, specifically to a system for urban airspace map construction and UAV path planning. Background Technology

[0002] With the acceleration of urbanization and the rapid development of drone technology, drones are increasingly widely used in urban environments, covering multiple fields such as logistics delivery, power line inspection, emergency rescue, and urban surveying. Urban airspace, as the core carrier of drone activities, is far more complex than open areas. It not only includes static obstacles such as dense high-rise buildings, bridges, and communication towers, but is also affected by dynamic meteorological factors such as real-time wind direction and speed. Furthermore, it contains airspace control zones such as airport clear zones and military control zones. These factors together constitute a complex and ever-changing urban airspace environment.

[0003] Unmanned aerial vehicle (UAV) path planning technology still faces numerous limitations when dealing with complex urban airspace environments. At the airspace information acquisition level, existing technologies primarily focus on acquiring static geographic information, such as building outline data, while often neglecting the impact of dynamic changes in meteorological factors on UAV flight. This results in planned paths that are difficult to adapt to real-time wind conditions, increasing flight energy consumption and safety risks. While some systems incorporate meteorological data, they fail to achieve deep correlation between building features, meteorological parameters, and airspace control information, making it difficult to form a comprehensive description of airspace characteristics.

[0004] In terms of constraint handling, traditional methods tend to have a coarse airspace division, often using large-scale grid cells, which fails to accurately capture the differences in obstacle distribution and weather abrupt changes in different areas. For areas with significant variations in building height, they fail to effectively label the passage restrictions caused by these height differences; and in areas with sudden wind direction changes, the lack of targeted constraint markers makes drones prone to attitude instability when traversing such areas.

[0005] The trajectory analysis and conflict identification processes also have shortcomings. Existing technologies often rely on static trajectory data when assessing trajectory continuity, failing to fully consider UAV flight altitude limitations and dynamic environmental parameters, leading to biased judgments on path adaptability. In conflict node identification, most focus only on obstacle collision risks, neglecting the potential dangers created by sudden wind changes and deviations in steering angle. This results in some high-risk nodes not being detected in time, thus affecting the effectiveness of path optimization.

[0006] Existing airspace maps are mostly static models, which cannot reflect the dynamic changes of the airspace environment in real time. This often leads to a mismatch between the path planning results based on static maps and the real-time environment in practical applications. This not only reduces the operating efficiency of UAVs, but may also cause air traffic conflicts, thus restricting the safe and efficient application of UAVs in complex urban environments. Summary of the Invention

[0007] The purpose of this invention is to provide an urban airspace map construction and UAV path planning system to solve the problems mentioned in the background art.

[0008] To achieve the above objectives, the present invention provides an urban airspace map construction and UAV path planning system, the system comprising:

[0009] The airspace feature fusion module collects point cloud data of building outlines, real-time wind direction values ​​from meteorological sensors, and boundary coordinates of airspace control areas within the urban airspace. It then associates spatial location information and calculates the distribution density of airspace obstacles to generate a three-dimensional dynamic feature set of the airspace.

[0010] The constraint modeling module extracts the obstacle density and wind direction change values ​​from the three-dimensional dynamic feature set of the airspace, divides the airspace into units according to the spatial grid, marks the building height range and wind direction change segments in each unit, and generates an airspace constraint label set.

[0011] The trajectory segmentation processing module obtains the airspace units of the non-abrupt segments in the airspace constraint label set, extracts the historical flight trajectory point sequence, analyzes the turning angle between trajectory points in combination with the UAV flight altitude limit, evaluates the path continuity strength in the dynamic environment, and generates trajectory continuity analysis results.

[0012] The conflict node screening module identifies trajectory points in the trajectory continuity analysis results whose turning angle deviation is greater than the safety threshold and are located in the wind direction change section, and marks them as potential conflict nodes to form a trajectory conflict node set.

[0013] The path optimization output module obtains the coordinates of all nodes and corresponding environmental parameters in the set of conflicting flight paths, calculates the shortest obstacle avoidance path between nodes, and generates an airspace map and the UAV's global optimized path planning results.

[0014] Preferably, the three-dimensional dynamic feature set of the airspace includes obstacle density hierarchy, spatial grid coding, and normalized meteorological factors;

[0015] The specific set of spatial constraint conditions includes building height range labels, wind direction change zone labels, and density difference rates of adjacent spatial units.

[0016] The results of the track continuity analysis include the cumulative deviation of the turning angle, the influence coefficient of the altitude change rate, and the comparison of the path breakage response under dynamic interference conditions.

[0017] The set of flight path conflict nodes includes the spatial coordinates of the conflict points, the characteristics of sudden wind speed changes at the conflict points, and the ratio of the turning angle of the conflict points to the safety radius offset.

[0018] The airspace map and the UAV global optimized path planning results include a three-dimensional airspace grid model, an optimized path node sequence, and path conflict avoidance labels.

[0019] Preferably, the spatial feature fusion module includes:

[0020] The multi-source data acquisition submodule acquires urban building laser point cloud data and real-time wind direction values ​​from meteorological monitoring stations, extracts the coordinates of building vertices and the locations of meteorological data acquisition points from the point cloud data, maps the coordinates to a unified spatial reference system, and generates a spatial basic data set.

[0021] The dynamic factor weighting submodule dynamically assigns weights based on the building vertex density and wind direction change rate in the aforementioned airspace basic data set. It then spatially superimposes the weighted obstacle distribution density with the real-time wind direction intensity value to generate a three-dimensional dynamic feature set for the airspace.

[0022] Preferably, the constraint modeling module includes:

[0023] The mesh division calculation submodule extracts the spatial location code from the three-dimensional dynamic feature set of the spatial domain, divides the spatial domain into cubic units with a preset mesh size, counts the maximum and minimum number of building vertices in each unit, and calculates the height range within the unit.

[0024] The abrupt change section detection submodule extracts the gradient of wind direction sensor data change between adjacent units based on the height range of the airspace units, marks the unit boundaries where the gradient value exceeds the abrupt change threshold, and integrates the building height range label and the wind direction abrupt change section label to generate an airspace constraint condition label set.

[0025] Preferably, the track segmentation processing module includes:

[0026] The flight track data extraction submodule filters airspace units labeled as non-wind direction change zones based on the airspace constraint label set, calls the timestamps and three-dimensional coordinate data of historical UAV flight track points, and constructs a spatial sequence of track points in chronological order.

[0027] The continuity assessment submodule calculates the three-dimensional turning angle and altitude change rate between adjacent trajectory points based on the spatial sequence of trajectory points, and identifies trajectory segments whose cumulative turning angle exceeds the continuity interruption threshold by combining the flight altitude limit threshold within the airspace unit, thereby generating track continuity analysis results.

[0028] Preferably, the conflict node filtering module includes:

[0029] The node positioning submodule extracts the coordinates of trajectory points with turning angle deviations greater than the safety threshold from the trajectory continuity analysis results, overlays the airspace constraint conditions to mark the spatial range of concentrated wind direction change sections, and filters the deviation trajectory points located within the change sections.

[0030] The conflict feature analysis submodule calls the real-time wind speed value and UAV safe hovering radius data of the filtered trajectory points, calculates the offset ratio between the wind speed change intensity and the safe radius, marks the trajectory points whose offset ratio exceeds the conflict judgment standard as potential conflict nodes, and generates a set of flight path conflict nodes.

[0031] Preferably, the path optimization output module includes:

[0032] The obstacle avoidance path calculation submodule obtains the three-dimensional coordinates and corresponding wind speed offset ratios of all nodes in the set of conflicting flight paths. Using the building obstacle density level as a constraint, it uses a spatial topology algorithm to calculate the shortest obstacle avoidance path between nodes.

[0033] The global path integration submodule, based on the shortest obstacle avoidance path sequence, integrates historical waypoint data from non-conflict zones, reconstructs the complete flight path according to temporal continuity, and generates an airspace map. Figure 3 3D grid model and UAV global optimization path planning results.

[0034] Preferably, the system further includes:

[0035] The real-time dynamic correction module accesses the real-time position coordinates of the UAV and the updated airspace weather data, extracts the next flight point from the global optimized path planning results, compares the real-time wind speed value with the wind speed change characteristics in the planned path, and dynamically adjusts the turning angle threshold of the path nodes.

[0036] Preferably, the real-time dynamic correction module includes:

[0037] The environmental data matching submodule obtains the real-time location coordinates and wind speed sensor values ​​transmitted by the UAV, calls the conflict feature data in the set of flight path conflict nodes, and calculates the environmental similarity between the current location and the nearest conflict node.

[0038] Based on the environmental similarity results, when the real-time wind speed offset ratio approaches the conflict threshold, the path replanning submodule recalculates the obstacle avoidance path sequence starting from the current node and updates the track point data in the global optimized path planning results.

[0039] Preferably, the system further includes:

[0040] The airspace map update module periodically collects newly added building point cloud data and airspace control rule change information, extracts the historical spatial grid codes from the airspace three-dimensional dynamic feature set, recalculates the obstacle distribution density after the addition of new data, and iteratively generates a new version of the airspace three-dimensional dynamic feature set.

[0041] Compared with the prior art, the beneficial effects of the present invention are:

[0042] This urban airspace mapping and UAV path planning system significantly enhances the path planning capabilities of UAVs in complex urban airspace environments through multi-module collaborative operation. The airspace feature fusion module breaks through the limitations of traditional airspace information collection, correlating building outline point cloud data, real-time meteorological parameters, and airspace control boundary coordinates. By calculating obstacle distribution density, the generated three-dimensional dynamic feature set of the airspace comprehensively reflects the static structure and dynamic changes of the airspace, making the system's perception of the airspace environment more three-dimensional and accurate, and avoiding planning deviations caused by missing information.

[0043] The constraint modeling module divides the airspace into units using a spatial grid, precisely marking the differences in building height and abrupt changes in wind direction within each unit. This resulting set of airspace constraint labels provides detailed constraint data for subsequent route planning. This refined constraint modeling approach allows the system to accurately identify high-risk areas and traffic restrictions within the airspace, ensuring that planned routes meet safety requirements while better reflecting the actual environmental characteristics of different airspace units, thus reducing the problem of unreasonable routes caused by ambiguity in constraints.

[0044] The track segmentation module combines historical flight trajectory point sequences with UAV flight altitude limitations to analyze the turning angles between trajectory points and evaluate the strength of path continuity. The generated track continuity analysis results can effectively reflect the stability of the path in dynamic environments. This process fully considers the flight characteristics of the UAV and dynamic environmental changes, enabling the planned track to maintain good continuity when traversing different airspace units, reducing the increased energy consumption and attitude adjustment pressure on UAV flight caused by sudden path changes.

[0045] The conflict node screening module identifies and marks trajectory points whose turning angle deviation exceeds the safety threshold and are located in areas of sudden wind changes as potential conflict nodes. The resulting set of flight path conflict nodes accurately pinpoints high-risk locations in the airspace. This targeted conflict identification method can detect potential flight risks in a dynamic environment in advance, providing clear optimization targets for subsequent path optimization and reducing the possibility of UAVs encountering sudden dangers during flight.

[0046] The path optimization output module calculates the shortest obstacle avoidance path based on the coordinates of conflict nodes and corresponding environmental parameters. The generated airspace map and the global optimized path planning results organically combine airspace environment visualization and path optimization. This result can clearly present the overall environmental characteristics of urban airspace and provide UAVs with an efficient flight path that adapts to the dynamic environment and avoids potential conflicts, making UAV operations in urban airspace more adaptable and safer. Attached Figure Description

[0047] Figure 1 This is a schematic diagram illustrating the working principle of the urban airspace map construction and UAV path planning system described in this invention.

[0048] Figure 2 A structural diagram of the system's data set;

[0049] Figure 3 Workflow diagram for the constraint modeling module;

[0050] Figure 4 Workflow diagram for the conflict node filtering module. Detailed Implementation

[0051] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0052] Please see Figure 1 This invention provides a system for constructing urban airspace maps and planning UAV paths. The system includes the collaborative operation of an airspace feature fusion module, a constraint condition modeling module, a trajectory segmentation processing module, a conflict node screening module, and a path optimization output module.

[0053] In the implementation process, after the airspace feature fusion module is activated, it collects building outline point cloud data, real-time wind direction values ​​from meteorological sensors, and airspace control zone boundary coordinates from the urban airspace. The building outline point cloud data is acquired through LiDAR scanning and contains a set of three-dimensional coordinate points, while the meteorological sensor data comes from real-time monitoring stations distributed throughout the city. The airspace control zone boundary coordinates are extracted from the airspace management database, covering the boundary point sequence of the geographic information system. All data are associated with spatial location information and mapped to a unified three-dimensional reference system through coordinate system transformation. The point cloud density algorithm is used to calculate the distribution density of building obstacles. This density is based on the number of point clouds per unit volume. Combined with wind direction values, time series analysis is performed to generate a set of three-dimensional dynamic features of the airspace.

[0054] The constraint modeling module then processes the feature set: it extracts obstacle density and wind direction change values ​​from the feature set, with the wind direction change values ​​including direction and intensity parameters; it divides the airspace into cubic units using a preset grid size, with the side length of each unit set to a fixed number of meters at the city scale; it calculates the building height range within the unit by identifying the Z-axis difference between the highest and lowest points within the unit; simultaneously, it identifies abrupt wind direction change zones by detecting gradient changes in adjacent units, with the gradient threshold preset according to the meteorological model; after labeling the height range and abrupt change zone, it integrates the density difference rate calculation results of adjacent units to generate an airspace constraint label set.

[0055] The track segmentation module operates as follows: it filters airspace units labeled as non-wind direction change zones from the annotation set; non-change zones are determined based on units with wind gradients below the change threshold; historical flight trajectory point sequences are extracted from the flight log database, including timestamps and three-dimensional coordinates; the sequences are reassembled in chronological order, and combined with the UAV's pre-configured flight altitude limit threshold, the turning angle between trajectory points is calculated through vector analysis; the cumulative value of the angle deviation is used to evaluate the path continuity strength in dynamic environments, such as the impact of instantaneous changes in urban wind fields, and generates track continuity analysis results.

[0056] The conflict node screening module then analyzes the results: it identifies trajectory points with turning angle deviations greater than a safety threshold, which is determined based on the UAV dynamics model; it overlays the wind direction change zone range in the annotation set and uses a spatial overlap algorithm to filter points located in the change zone; these points are labeled as potential conflict nodes, forming a set of flight path conflict nodes, which includes location and environmental feature data.

[0057] The path optimization output module processes this set: obtains the coordinates of all nodes and real-time wind speed values; calculates obstacle avoidance paths between nodes using the shortest path algorithm, with building density as the obstacle weight; integrates trajectory point data from non-conflict sections and reassembles the complete flight path sequence; outputs a 3D grid model of the airspace map and the global optimized path planning results, with the map constructed based on grid cell encoding.

[0058] Example 1: See Figure 2The construction of the three-dimensional dynamic feature set in the spatial domain relies on a multi-level data structure. The obstacle density level is determined through analysis within spatial grid cells. This level reflects the intensity of point cloud distribution, and the level division is based on the statistical distribution range of point cloud quantity per unit volume. Density values ​​are categorized using a clustering algorithm to form discrete level indexes, with different indices corresponding to different levels of building obstacle concentration. The spatial grid coding system is implemented using a three-dimensional spatial partitioned tree structure. Each cube cell generates a unique identifier based on its geometric center coordinates. The identifier contains level location information and uses hash mapping to establish a fast retrieval mechanism. Normalized meteorological factors are processed by integrating wind direction data from meteorological monitoring stations. The initial wind direction vector and intensity scalar are standardized to remove dimensional differences. The normalization process uses a linear scaling algorithm to map the original values ​​to the [0,1] interval. The feature set is stored as a tree-structured data table, with the main index being the spatial grid code, and child nodes carrying density levels, meteorological factor values, and timestamp attributes respectively.

[0059] The generation of the spatial constraint label set relies on multi-dimensional spatial attribute labeling. The calculation of building height range labels is performed within the grid cell, obtaining the maximum and minimum values ​​through fast sorting of the point cloud Z-coordinates. The range value is appended to the cell data structure as a floating-point label. Wind direction change zone labels are established using spatial neighborhood analysis, starting with the calculation of wind direction change gradients between adjacent cells: a selected cell and its east, south, and upward-facing adjacent cells form an analysis cell group, and the absolute value of the angular difference of the wind direction vector within the group is calculated. When the difference in any direction exceeds a preset gradient threshold, the cell boundary labeling mechanism is triggered. Boundary labeling results are stored as spatial polygon objects, with polygon vertices composed of cell boundary coordinates. The density difference rate between adjacent spatial cells is calculated using a double-loop algorithm: while traversing all cells, its density level value is simultaneously compared with that of its 26 neighboring cells, and the difference rate is taken as the percentage of the level difference to the maximum level. The final label set is stored in the form of a three-dimensional matrix, with matrix coordinates corresponding to spatial grid positions. Each matrix element contains a numerical label for the height range, a Boolean label for the change zone, and a floating-point value for the density difference rate.

[0060] The results of the track continuity analysis are generated based on time-series track data processing. The calculation of the cumulative deviation value of the turning angle adopts an iterative accumulation mechanism: starting from the first item of the historical track point sequence, the three-dimensional vector turning angle between adjacent track points is calculated one by one, and the angle value is converted to a radian value using the unit vector dot product formula. The system sets a continuity check window; when the arithmetic sum of the radian changes within the window exceeds a preset angle threshold, it is recorded as a deviation event. The altitude change rate influence coefficient is calculated as the ratio of the altitude difference of the track points to the time difference; the time difference is smoothed and filtered to eliminate noise interference. This coefficient is coupled with the turning angle to establish a weighted relationship, used to quantify the degree of influence of altitude changes on track abrupt changes. The path breakage response comparison under dynamic disturbance conditions adopts a two-parameter analysis method: a mapping model between wind speed disturbance intensity and path breakage frequency is established, and the probability distribution of track breakage under different wind levels is statistically analyzed based on the historical database. The analysis results include a three-dimensional angle deviation heatmap, an altitude-wind speed coupling coefficient table, and a path breakage frequency distribution curve; the data is stored in a structured query database.

[0061] The formation of the conflict node set utilizes spatial cross-validation technology. The spatial coordinates of the conflict points are extracted from a flight track continuity analysis database, including latitude, longitude, and altitude information. The wind speed abrupt change characteristics at the conflict points are extracted through temporal correlation analysis: based on the conflict point timestamp, a meteorological database is traced back to retrieve the maximum wind rate of change within five minutes before and after that moment. The wind rate of change is calculated based on second-level sampling data, taking the maximum value of the vector modulus difference of consecutive sampling points. The establishment of the conflict point turning angle and safe radius offset ratio relies on the UAV physical parameter database: the safe hovering radius is preset to a fixed value based on the UAV's dynamic parameters, and the offset ratio is defined as the product of the actual turning radian value at that point and the reciprocal of the safe radius. The offset ratio threshold is set as a configurable parameter; a conflict marker is triggered when the calculation result exceeds the threshold. The final set is stored in a spatial index database, with each record containing four-dimensional coordinates (spatial + temporal), wind characteristic values, offset ratio values, and a conflict level label.

[0062] The output of airspace maps and UAV global optimized path planning results adopts a hybrid data encapsulation format. The construction of the 3D airspace mesh model relies on a spatial mesh coding system to map unit attribute data to corresponding geometries. The model data includes a point cloud distribution heatmap layer, a meteorological change gradient layer, and a controlled area boundary layer, which are loaded and displayed in a layered rendering mode. The optimized path node sequence is output by the obstacle avoidance calculation submodule, and the sequence consists of connected 3D coordinate points. Each path node carries a set of environmental parameter labels, including the density level, wind intensity, and steering constraint values ​​of the passed units. Path conflict avoidance labels are generated based on the set of flight path conflict nodes: a region with a preset buffer zone from the conflict point is marked on the optimized path trajectory line, and the buffer radius is adaptively scaled according to the conflict level. The labels use binary flags to indicate the safety status of each segment of the path trajectory, and simultaneously generate a set of alternative path suggestion points. The planning results are encapsulated in the form of a time series package, including the basic path coordinate sequence, conflict label array, and environmental parameter mapping table, supporting visualization front-end calls and flight control protocol transmission.

[0063] Example 2: See Figure 3 After the multi-source data acquisition submodule starts, it connects to the urban infrastructure database system. The point cloud datasets generated by the laser scanning equipment are input in binary format. Each dataset contains the coordinates of hundreds of thousands of building outline vertices, and the coordinates use the local coordinate system as the reference system. The meteorological monitoring station sends real-time wind direction records through the IoT transmission layer. Each record contains the direction angle, wind speed scalar, and collection timestamp. The station's geographical coordinates are stored independently in the location registry. The coordinate mapping operation calls the 3D transformation matrix library: the point cloud vertices in the local coordinate system are processed by a seven-parameter transformation formula to output the 3D coordinates in the WGS84 ellipsoidal coordinate system; the meteorological station coordinates are directly associated with the same reference system through GPS positioning information. The unification process of spatial location information adopts a dynamic projection mechanism: each data point is attached with a coordinate system identifier, and the coordinate transformation service is called in real time to complete the mapping. After processing, the airspace basic data group is organized in a spatiotemporal joint index structure. The time dimension index uses time-division storage blocks, and the spatial dimension index establishes an R-tree structure. Each data node contains point cloud density statistics, wind direction sequence values, and control boundary topology relationships.

[0064] The dynamic factor weighting submodule initiates a dual weight allocation process after receiving the basic data set. Building vertex density is calculated using a kernel density estimation algorithm: the number of points in each grid cell is used as the initial density value, a bandwidth parameter is introduced to control the spatial smoothness of the density, and the density output value is normalized to a standard distribution range. The wind direction change rate is calculated using a sliding window analysis method: the standard deviation of the wind direction angle within the time window is used as the base value for the change rate, and the window size is dynamically configured based on the characteristics of the urban wind field. The weight allocation model employs an adaptive adjustment mechanism: the building density weight coefficient increases with the increase in regional building complexity, and the coefficient adjustment is based on the comparison value of the density distribution of adjacent cells; the wind direction change rate weight maintains a proportional relationship with its actual change rate value, and the weight ratio is automatically increased during periods of high change rate. The weighted obstacle distribution density and real-time wind direction intensity value are dynamically superimposed in three-dimensional space: the superposition process uses a unit attribute fusion algorithm, independently calculating the linear weighted sum of density level and wind direction factor for each grid cell, where the wind direction intensity value is converted into a scalar wind pressure coefficient for calculation. The output constructs a three-dimensional dynamic feature space object. Each spatial grid cell object stores the density level, wind direction normalized intensity, combined weight value, and time validity marker. The overall data structure supports streaming data transmission interface.

[0065] The grid partitioning calculation submodule extracts spatial location encoding sequences from the feature set, with the encoding symbol consisting of a three-digit integer partition index. During system initialization, spatial grid size parameters are preset: the cell side length is set to an integer value representing the typical resolution size of urban airspace, and the segmentation process employs a three-dimensional uniform partitioning algorithm. Building vertex counts are performed within cells using point cloud clustering: the spatial encoding index is used to quickly retrieve point sets within the cell, the Z-axis elevation distribution characteristics of the point sets are calculated, the maximum value is detected using a sorting algorithm to obtain the 99th percentile, and the minimum value is determined using the 5th percentile to reduce outlier interference. The height range is calculated using statistical methods: a kernel density function is used to fit the point cloud height distribution curve, and the range is the difference between the highest and lowest points in the curve's peak interval. The processed height range is used as an inherent attribute of the cell and bound to the grid object; the data record includes the cell code, range value, and confidence parameter.

[0066] The abrupt change zone detection submodule performs spatial gradient analysis based on height range data. Wind direction sensor data from adjacent units are organized using a doubly linked list structure, with each node containing a wind direction time-domain sequence. The gradient calculation process consists of three layers: first, comparing the average wind direction angles on both sides of the unit interface; second, calculating the spatial projection angle between the wind direction vectors of the two units; and finally, analyzing the synchronization of wind direction variability based on a time window. The abrupt change threshold is set as a configurable variable group, including angle abrupt change threshold, vector variability threshold, and time synchronization threshold. When an adjacent unit combination meets any of the threshold conditions, a polygon annotation program is triggered: the spatial position of the interface is determined by connecting the geometric centers of the units; the starting coordinates of the annotated zone are taken from the intersection point of the unit boundaries, and the ending coordinates are taken from the next intersecting boundary point; the generated abrupt change zone polygons are stored in order of spatial topological relationships. The final constraint annotation set integrates three data types: a numerical table of building height ranges at the unit level, a wind direction abrupt change zone map described by spatial polygons, and a symmetric matrix of density difference rates between units. The data storage adopts a hybrid index structure: the key-value pairs of grid cells store height and density data, the spatial relation database stores polygon topology, and the difference rate matrix is ​​stored in a distributed memory array to support high-speed queries.

[0067] Example 3: See Figure 4 The system begins operation with the flight track data extraction submodule. This module first accesses the airspace constraint label set database, retrieving all airspace unit records labeled as non-wind direction change zones. These units are identified using a spatial grid coding system, with each code corresponding to a specific three-dimensional spatial region. The system establishes a query statement to filter units that meet the conditions, returning a result set containing the unit's unique identifier, geometric boundary parameters, and environmental attribute labels. Historical flight track data is loaded from a distributed storage system. Each track record contains a timestamp field, three-dimensional coordinate points, aircraft attitude parameters, and equipment status codes. During the data preprocessing stage, a time series integrity check is performed: when the time interval between adjacent records exceeds the maximum allowable interval set by the configuration parameters, the system automatically triggers a linear interpolation algorithm to generate supplementary track points. The interpolation process considers aircraft dynamic constraints to ensure that the generated points conform to kinematic laws. The processed track data is reassembled into a continuous spatial sequence in chronological order, and the storage structure adopts a four-dimensional array format, supporting fast range queries and sliding window analysis.

[0068] After receiving the preprocessed trajectory sequence, the continuity assessment submodule initiates a multi-dimensional feature extraction process. The core of this module is calculating the steering angle deviation, used to quantify the degree of trajectory abrupt changes. For two vector segments formed by three consecutive points in the trajectory sequence, the system calculates their spatial steering angle values. The calculation process uses vector geometry methods, first converting the coordinates of adjacent trajectory points into displacement vectors, and then deriving the included angle through dot product operations. The angle value calculation is standardized, and the results are uniformly converted to radians between 0 and π. Altitude change feature analysis is performed simultaneously; the system calculates the vertical velocity components between adjacent trajectory points and normalizes them using preset flight altitude limit parameters within the airspace unit. Dynamic environmental interference analysis establishes a trajectory stability evaluation model, which comprehensively considers the coupling effect of steering angle fluctuations and altitude change rates. The evaluation process uses a sliding window technique, with the window size dynamically adjusted according to the aircraft type. The cumulative steering angle deviation and altitude change characteristics of the data points within the window are weighted and fused to generate a path continuity strength index. The system assigns a continuity score to each trajectory point, reflecting the point's path maintenance capability under current environmental conditions. The score results are compared with preset safety thresholds to identify potential discontinuous trajectory segments. The analysis results are output as a structured dataset, which includes the original trajectory point sequence, enhanced feature vectors, and continuity markers.

[0069] The node localization submodule's workflow is based on spatial topology analysis. This module takes into account a set of anomalous trajectory points from the continuity assessment results; these points are marked with indicators indicating turning angles exceeding safe limits. The system loads spatial data of wind direction abrupt change zones from the spatial constraint label set. This data is stored in a polygonal grid, with each polygon associated with a wind speed abrupt change intensity parameter. The spatial relationship detection algorithm determines the location of anomalous trajectory points and abrupt change zones, using a ray-mapping method for point-to-surface inclusion testing. The detection process is accelerated using spatial indexing, rapidly locating potentially intersecting areas through a 3D R-tree structure. For anomalous trajectory points determined to be within abrupt change zones, the system extracts their spatiotemporal coordinates and environmental parameters, constructing a candidate conflict node set. Each node in the set records its spatial location, timestamp, turning angle deviation, and the code of its respective abrupt change zone. This data organization supports efficient range queries and nearest neighbor searches, facilitating subsequent feature analysis.

[0070] The conflict feature analysis submodule performs a refined evaluation of candidate nodes. This module first establishes a spatiotemporal correlation query, retrieving meteorological monitoring data for the corresponding time based on the node's timestamp. The search scope includes specific time windows before and after the node to capture abrupt wind speed changes. Wind speed data processing employs a moving average filter to eliminate high-frequency noise while preserving trend changes. The abrupt change intensity calculation considers the rate of change of the wind speed vector's modulus and the angle of change of direction, quantifying the degree of wind field disturbance through a composite index. The safe radius offset calculation combines aircraft dynamic parameters, converting the turning angle deviation into a spatial offset distance. The conflict determination model integrates the wind speed abrupt change intensity and the safe radius offset, using a nonlinear function to calculate the conflict risk value. Nodes with risk values ​​exceeding a dynamic threshold are marked as confirmed conflict points, and conflict types are categorized and recorded. The final generated set of conflict nodes is stored in a spatiotemporal database, supporting multi-dimensional retrieval and analysis. The data model includes spatial coordinates, time information, environmental parameters, conflict features, and type labels, providing a decision-making basis for path optimization.

[0071] Formula Explanation: The steering angle is calculated using the following expression.

[0072]

[0073] in, This is the displacement vector of the previous trajectory segment; This is the displacement vector of the current segment of the trajectory; The spatial turning angle between the two vectors.

[0074] Example 4: Execution begins with the obstacle avoidance path calculation submodule. This module first loads the full data of the set of conflict nodes in the flight path, including the 3D spatial coordinates, timestamp, wind speed change characteristics, and turning angle offset parameters of each conflict node. The system establishes a spatial query interface to retrieve the building obstacle density level data of the airspace grid cell where the conflict node is located. The density data is organized in an octree structure, supporting fast 3D range queries. In the environmental parameter preprocessing stage, the building density values ​​and wind speed characteristics are normalized to form a unified obstacle influence coefficient. The path search algorithm adopts an improved 3D A* method, and the heuristic function design considers the dual constraints of spatial distance and obstacle distribution. The cost function weights are dynamically adjusted during the search process to achieve a balance between path length and safety margin. The node expansion strategy introduces turning angle constraints to ensure that the generated path complies with the UAV's maneuverability limitations. For high-density obstacle areas, the system automatically triggers a refined search mode, reducing the grid step size to improve path accuracy. The calculated obstacle avoidance path segments are output in the form of a spatial coordinate sequence, with environmental parameter labels and turning constraint suggestions attached to each path point.

[0075] The global path integration submodule initiates the trajectory reassembly process after receiving the obstacle avoidance path segment. This module first accesses the raw output data from the trajectory segmentation module, extracting historical flight trajectory points from non-conflict sections. Data alignment ensures the continuity of the time series, and interpolation algorithms are used to fill any minor gaps. The path fusion process performs spatial smoothing, inserting transition curves at the junctions of obstacle avoidance path segments and historical trajectory segments. Transition curve generation considers aircraft dynamics, using cubic spline interpolation to ensure second-order continuity of the trajectory. The reassembled complete path undergoes global optimization, adjusting the positions of key waypoints through dynamic programming to reduce overall flight energy consumption. The system output includes two data formats: a discrete waypoint sequence for the flight control system and a parametric curve description for visualization. The data encapsulation format supports multiple transmission protocols to meet the needs of different application scenarios. Path conflict avoidance labels employ a layered labeling mechanism, overlaying safety level markers and alternative path suggestions onto the basic trajectory.

[0076] The real-time dynamic correction module establishes a data channel with the UAV flight control system, continuously receiving real-time status information from the aircraft. The status data packet includes the current 3D position, velocity vector, attitude angle, and onboard sensor readings. The system maintains a dynamic environmental database, periodically updating airspace weather conditions and obstacle status. The trajectory monitoring process compares the actual flight path with the planned path in real time, detecting segments where deviations exceed a threshold. The environmental change response mechanism analyzes the difference between the current wind speed and the predicted value, triggering local path replanning when a significant change is detected. The replanning range is dynamically determined based on the degree of environmental change, prioritizing high-risk segments. Correction command generation considers flight control delay characteristics, and lead calculation is based on the UAV dynamic response model. The system output is an incremental trajectory update command, reducing data transmission volume and improving response speed.

[0077] Formula Explanation: The obstacle avoidance path cost function is expressed by the following expression.

[0078]

[0079] The meanings of each character are as follows: This represents the spatial distance of the i-th path segment; This represents the obstacle influence coefficient of the i-th path segment; This indicates the change in turning angle between adjacent path segments; This represents the distance weighting parameter; Represents the obstacle weight parameters; This represents the steering penalty coefficient.

[0080] This system demonstrates adaptive path planning capabilities in dynamic environments, achieving a balance between computational efficiency and path quality through a layered processing architecture. The system design emphasizes modularity, with functional components collaborating through standardized interfaces. The data processing flow employs a streaming computing model, supporting real-time operations in large-scale airspace environments. Algorithm implementation considers practical engineering constraints, striking a balance between theoretical optimization and real-time performance. The storage solution is optimized for spatiotemporal data characteristics, ensuring fast access and efficient updates. The entire system runs on a distributed computing platform, with load balancing mechanisms guaranteeing responsiveness in high-concurrency scenarios. Implementation details fully consider the actual conditions of UAV applications, including communication latency, sensor errors, and computational resource limitations, ensuring the solution's engineering practicality.

[0081] Example 5: The data acquisition process begins with the environmental data matching submodule. This submodule establishes a UAV communication data stream interface to continuously receive real-time status data packets transmitted from the aircraft. After decoding the data packets, key fields are extracted: global positioning coordinates including longitude, latitude, and altitude values; airborne meteorological sensor readings including three-dimensional wind speed vectors and atmospheric temperature; and an additional timestamp accurate to the millisecond level. The system synchronously calls the flight path conflict node set database, which stores historical conflict feature records by airspace grid partitioning. Each feature record includes the spatial coordinates of the conflict point, a timestamp, wind speed change intensity value, and steering offset ratio parameter. After the similarity calculation engine is started, a multi-dimensional matching model is established: the spatial dimension uses a weighted spherical distance algorithm, comprehensively considering horizontal distance and vertical height difference; the environmental dimension uses wind speed feature difference analysis, comparing the modulus difference and angle difference between the real-time wind speed vector and the conflict feature vector. The matching results generate a similarity score, which reflects the degree of conformity between the current flight environment and the historical conflict scenario.

[0082] The path replanning submodule monitors the environment matching results in real time and triggers the path update procedure under specific conditions. The system maintains a dynamic decision state machine, which switches to replanning mode when the environment similarity score continuously exceeds the warning threshold and the real-time wind speed deviation ratio approaches the preset conflict boundary value. The replanning operation uses the UAV's current position as the calculation starting point and reads the updated three-dimensional dynamic feature set data of the airspace. The query range for the building obstacle density level is limited to the neighborhood grid of the current flight airspace, and the query efficiency is improved by using a spatial sliding window. The obstacle avoidance path generation reuses the core algorithm of the path optimization output module, but the calculation constraints are dynamically strengthened: the weight of the safety distance is increased in high-risk areas of sudden wind speed changes, and the priority of turning angle restrictions is increased in densely built areas. After the new path sequence is calculated, a smoothing process is performed, and a parametric curve transition is used to ensure the continuity of the path derivative. The system uses an incremental replacement strategy to perform the path update operation: the current track point position node is located in the global optimized path planning results, subsequent path segments are deleted, and new planning segments are inserted. The update data packet is encapsulated and transmitted through the flight control protocol, including a sequence number verification mechanism to avoid transmission errors.

[0083] The airspace map update module runs as an independent periodic task, with its execution frequency configured based on the rate of change in urban planning. Data acquisition tasks utilize multi-source interfaces: newly added building point cloud data is periodically retrieved from the municipal geographic information system, maintaining consistency with historical data formats; airspace control rule change information is updated via subscription to the Civil Aviation Administration's announcement push service, automatically resolving geofence coordinate changes. The historical data processing stage retrieves the current system's airspace 3D dynamic feature set and extracts its spatial grid coding structure. New data fusion employs a version control strategy: for unchanged airspace grid units, historical obstacle distribution density values ​​are directly inherited; for newly added or changed units, the point cloud density calculation process is re-executed. The obstacle density update algorithm supports incremental calculation, processing only grid units within the affected area of ​​the changed region. After iterative calculation, a new feature set file is generated, stored using a time-version naming convention, while retaining the three most recent historical versions for retrospective analysis. Database switching operations are performed during low-load periods to ensure service continuity.

[0084] The flight status feedback loop calibrates environmental matching accuracy at fixed intervals: it compares the actual flight parameters of the UAV with the predicted path, records deviation exceeding limits, and analyzes the causes. An environmental database update listener detects meteorological data refresh events and automatically recalculates the similarity score. An anomaly handling mechanism is activated upon detecting a data stream interruption, using backup cached data to maintain short-term decision-making capabilities. All operation logs record the status of key nodes, including environmental similarity score curves, replanning trigger records, and map update execution reports. Log data includes aircraft identification codes and airspace block codes to support subsequent playback analysis.

[0085] Emphasizing adaptive response capabilities in dynamic environments, the system comprises three sub-modules forming a cascaded processing chain. Environmental data matching establishes the foundation for real-time risk warning, path replanning enables dynamic flight correction, and airspace map updates ensure the timeliness of basic data. The system deployment employs a microservice architecture, with each sub-module independently deployed and scaled. A publish-subscribe communication mechanism ensures efficient data processing under high concurrency scenarios. The dynamic allocation strategy for computing resources adjusts based on task priority: environmental matching tasks receive the highest real-time priority, path replanning tasks are allocated medium computing resources, and map update tasks run in the background with low priority. The implementation process strictly separates control logic from the data storage layer, and all state changes are managed through transaction management to ensure consistency.

[0086] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0087] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A system for urban airspace map construction and UAV path planning, characterized in that, The system includes: The airspace feature fusion module collects point cloud data of building outlines, real-time wind direction values ​​from meteorological sensors, and boundary coordinates of airspace control areas within the urban airspace. It then associates spatial location information and calculates the distribution density of airspace obstacles to generate a three-dimensional dynamic feature set of the airspace. The constraint modeling module extracts the obstacle density and wind direction change values ​​from the three-dimensional dynamic feature set of the airspace, divides the airspace into units according to the spatial grid, marks the building height range and wind direction change segments in each unit, and generates an airspace constraint label set. The trajectory segmentation processing module obtains the airspace units of the non-abrupt segments in the airspace constraint label set, extracts the historical flight trajectory point sequence, analyzes the turning angle between trajectory points in combination with the UAV flight altitude limit, evaluates the path continuity strength in the dynamic environment, and generates trajectory continuity analysis results. The conflict node screening module identifies trajectory points in the trajectory continuity analysis results whose turning angle deviation is greater than the safety threshold and are located in the wind direction change section, and marks them as potential conflict nodes to form a trajectory conflict node set. The path optimization output module obtains the coordinates of all nodes and corresponding environmental parameters in the set of conflicting flight paths, calculates the shortest obstacle avoidance path between nodes, and generates an airspace map and the UAV's global optimized path planning results.

2. The urban airspace map construction and UAV path planning system according to claim 1, characterized in that: The three-dimensional dynamic feature set of the airspace includes obstacle density hierarchy, spatial grid coding, and normalized meteorological factors; The specific set of spatial constraint conditions includes building height range labels, wind direction change zone labels, and density difference rates of adjacent spatial units. The results of the track continuity analysis include the cumulative deviation of the turning angle, the influence coefficient of the altitude change rate, and the comparison of the path breakage response under dynamic interference conditions. The set of flight path conflict nodes includes the spatial coordinates of the conflict points, the characteristics of sudden wind speed changes at the conflict points, and the ratio of the turning angle of the conflict points to the safety radius offset. The airspace map and the UAV global optimized path planning results include a three-dimensional airspace grid model, an optimized path node sequence, and path conflict avoidance labels.

3. The urban airspace map construction and UAV path planning system according to claim 1, characterized in that, The spatial feature fusion module includes: The multi-source data acquisition submodule acquires urban building laser point cloud data and real-time wind direction values ​​from meteorological monitoring stations, extracts the coordinates of building vertices and the locations of meteorological data acquisition points from the point cloud data, maps the coordinates to a unified spatial reference system, and generates a spatial basic data set. The dynamic factor weighting submodule dynamically assigns weights based on the building vertex density and wind direction change rate in the aforementioned airspace basic data set. It then spatially superimposes the weighted obstacle distribution density with the real-time wind direction intensity value to generate a three-dimensional dynamic feature set for the airspace.

4. The urban airspace map construction and UAV path planning system according to claim 3, characterized in that, The constraint modeling module includes: The mesh division calculation submodule extracts the spatial location code from the three-dimensional dynamic feature set of the spatial domain, divides the spatial domain into cubic units with a preset mesh size, counts the maximum and minimum number of building vertices in each unit, and calculates the height range within the unit. The abrupt change section detection submodule extracts the gradient of wind direction sensor data change between adjacent units based on the height range of the airspace units, marks the unit boundaries where the gradient value exceeds the abrupt change threshold, and integrates the building height range label and the wind direction abrupt change section label to generate an airspace constraint condition label set.

5. The urban airspace map construction and UAV path planning system according to claim 4, characterized in that, The trajectory segmentation processing module includes: The flight track data extraction submodule filters airspace units labeled as non-wind direction change zones based on the airspace constraint label set, calls the timestamps and three-dimensional coordinate data of historical UAV flight track points, and constructs a spatial sequence of track points in chronological order. The continuity assessment submodule calculates the three-dimensional turning angle and altitude change rate between adjacent trajectory points based on the spatial sequence of trajectory points, and identifies trajectory segments whose cumulative turning angle exceeds the continuity interruption threshold by combining the flight altitude limit threshold within the airspace unit, thereby generating track continuity analysis results.

6. The urban airspace map construction and UAV path planning system according to claim 5, characterized in that, The conflict node filtering module includes: The node positioning submodule extracts the coordinates of trajectory points with turning angle deviations greater than the safety threshold from the trajectory continuity analysis results, overlays the airspace constraint conditions to mark the spatial range of concentrated wind direction change sections, and filters the deviation trajectory points located within the change sections. The conflict feature analysis submodule calls the real-time wind speed value and UAV safe hovering radius data of the filtered trajectory points, calculates the offset ratio between the wind speed change intensity and the safe radius, marks the trajectory points whose offset ratio exceeds the conflict judgment standard as potential conflict nodes, and generates a set of flight path conflict nodes.

7. The urban airspace map construction and UAV path planning system according to claim 6, characterized in that, The path optimization output module includes: The obstacle avoidance path calculation submodule obtains the three-dimensional coordinates and corresponding wind speed offset ratios of all nodes in the set of conflicting flight paths. Using the building obstacle density level as a constraint, it uses a spatial topology algorithm to calculate the shortest obstacle avoidance path between nodes. The global path integration submodule, based on the shortest obstacle avoidance path sequence, integrates historical flight path point data of non-conflict sections, reconstructs the complete flight path according to time continuity, and generates a three-dimensional grid model of the airspace map and the global optimized path planning results of the UAV.

8. The urban airspace map construction and UAV path planning system according to claim 7, characterized in that, The system also includes: The real-time dynamic correction module accesses the real-time position coordinates of the UAV and the updated airspace weather data, extracts the next flight point from the global optimized path planning results, compares the real-time wind speed value with the wind speed change characteristics in the planned path, and dynamically adjusts the turning angle threshold of the path nodes.

9. The urban airspace map construction and UAV path planning system according to claim 8, characterized in that, The real-time dynamic correction module includes: The environmental data matching submodule obtains the real-time location coordinates and wind speed sensor values ​​transmitted by the UAV, calls the conflict feature data in the set of flight path conflict nodes, and calculates the environmental similarity between the current location and the nearest conflict node. Based on the environmental similarity results, when the real-time wind speed offset ratio approaches the conflict threshold, the path replanning submodule recalculates the obstacle avoidance path sequence starting from the current node and updates the track point data in the global optimized path planning results.

10. The urban airspace map construction and UAV path planning system according to claim 9, characterized in that, The system also includes: The airspace map update module periodically collects newly added building point cloud data and airspace control rule change information, extracts the historical spatial grid codes from the airspace three-dimensional dynamic feature set, recalculates the obstacle distribution density after the addition of new data, and iteratively generates a new version of the airspace three-dimensional dynamic feature set.

Images