An autonomous landing site selection method for an electric vertical take-off and landing aircraft in a complex urban environment

Abstract

The application discloses a kind of electric vertical take-off and landing aircraft city complex environment autonomous landing site selection method, it is related to aircraft technical field, comprising: circumscribe aircraft pre-landing area, collection pre-landing area in evaluation data, through evaluation data creation landing evaluation coefficient, obtain first target landing area;Real-time flight training data and slope identification data of aircraft in first target landing area are collected, and real-time flight training data is generated real-time ground obstacle elevation by trained aerial triangulation model;Through ArcGIS, ground obstacle elevation and DEM data are fused, and then ground obstacle three-dimensional data DSM is obtained;Second target landing area of aircraft landing point is circumscribed in combination with ground obstacle three-dimensional data DSM and slope identification data;Through obstacle avoidance bounding box tree, intersection detection is carried out to second target landing area, and aircraft preliminary site landing point coordinates meeting that bounding box primitive is not intersected are obtained.

Description

Technical Field

[0001] This invention relates to the field of aircraft landing site selection, and more specifically, to a method for autonomous landing site selection of electric vertical take-off and landing aircraft in complex urban environments. Background Technology

[0002] In the complex urban environment, autonomous landing site selection for electric vertical takeoff and landing (EVL) aircraft is a key technology that ensures that the aircraft lands in a suitable location to meet the needs of urban air traffic. However, due to the complexity and uncertainty of the urban environment, autonomous landing site selection for aircraft still faces some technical challenges. The invention disclosed in CN116070812A presents a method for selecting take-off and landing points for manned electric vertical take-off and landing (EVTOL) aircraft. This invention imports relevant geographic information data of a city into ArcGIS to generate a 3D image map of the city, excludes airspace-restricted areas, and filters candidate points that meet the height and area requirements using a raster grid. Then, it estimates the number of suitable take-off and landing points for urban construction based on construction costs and establishes a busy coefficient for each demand point. Finally, it combines a P-median model with the shortest weighted path from the demand point to the take-off and landing point as the optimization objective to establish a mathematical model, analyzes the service radius of each point, and uses a greedy algorithm to perform segmented calculations on the model to obtain the final site selection scheme. However, this invention only considers height and area conditions when filtering candidate take-off and landing points, neglecting 3D data of ground obstacles and slope recognition data, and failing to consider the flight restriction surface during aircraft landing. It cannot determine the aircraft landing point through 3D collision detection. Therefore, this invention innovatively addresses these problems. Summary of the Invention

[0003] To overcome the aforementioned deficiencies of the prior art, the present invention provides the following technical solution: a method for autonomous landing site selection of an electric vertical takeoff and landing aircraft in complex urban environments, comprising: Step S1: Delineate the pre-landing area of ​​the aircraft, collect evaluation data within the pre-landing area, create landing evaluation coefficients based on the evaluation data, and obtain the first target landing area; Step S2: Collect real-time flight characteristic data and slope identification data of the aircraft in the first target landing area. The real-time flight characteristic data generates real-time ground obstacle elevation through a trained aerial triangulation model. The ground obstacle elevation and DEM data are fused through ArcGIS to obtain ground obstacle 3D data (DSM). The second target landing area of ​​the aircraft landing point is delineated by combining the ground obstacle 3D data (DSM) and slope identification data. Step S3: By avoiding the bounding box tree through obstacles, perform intersection detection on the landing area of ​​the second target to obtain the coordinates of the initial landing point of the aircraft that satisfies the condition that the bounding box primitives do not intersect. Step S4: Construct the flight restriction surface for the aircraft landing, establish the bounding box tree of the flight restriction surface, and perform collision detection on the coordinates of the initial landing point using the ray and triangle intersection detection algorithm. The coordinates of the flight restriction surface that satisfy the condition that primitives do not intersect are the coordinates of the final landing point of the aircraft.

[0004] Furthermore, the pre-landing area of ​​the aircraft is set up by staff, and the evaluation data includes the height of ground buildings, the wind speed when the aircraft lands, the runway length of the pre-landing area, and the flight restriction area; the wind speed when the aircraft lands is collected by an anemometer, and the height of ground buildings, the runway length of the pre-landing area, and the flight restriction area are obtained by radar equipment; When the landing evaluation coefficient reaches the preset landing evaluation coefficient threshold, the pre-landing area of ​​the aircraft is determined to be the first target landing area. Furthermore, the aerial triangulation model is trained using historical data, which includes historical flight characteristic data, including aircraft position and attitude data and ground reference data. The aircraft position and attitude data includes the altitude, forward overlap rate, lateral overlap rate, and resolution imagery of the airborne photography equipment. The ground reference data includes DEM (Digital Elevation Model) data, which is acquired through radar equipment. The slope identification data includes slope values ​​and slope ranges, which are collected by a laser scanner. The historical flight training data is collected through a flight data recorder installed inside the aircraft fuselage. Furthermore, the training of the aerial triangulation model includes: Collect k sets of flight training data in advance, where k is an integer greater than 1; Flight training data is divided into training and test sets. A classifier is constructed using real-time flight feature data from the training set as input data and ground obstacle elevation data from the training set as output data. The classifier is trained to obtain an initial classifier. The initial classifier is then tested using the test set. When the root mean square error between the predicted ground obstacle elevation and the actual ground obstacle elevation is within 0.01, the classifier that meets the preset accuracy is output as the aerial triangulation model. The aerial triangulation model can be one of the following: Naive Bayes algorithm, random forest regression, or neural network regression model. The root mean square error expression is as follows:

[0005] In the formula, The root mean square error, The number of ground obstacles by elevation. For the predicted ground obstacle elevation, This represents the actual elevation of the ground obstacle. Furthermore, the method for constructing the flight limiting surface includes: By acquiring the 3D ground obstacle data (DSM), obstacles in the second target landing area are identified. Based on the aircraft's flight mission, the operational limitations of the aircraft under given terrain and obstacle conditions are assessed, thereby constructing the flight restriction surface. Within the flight restriction surface, the Final Approach and Takeoff Area (FATO), Clearway, Touchdown and Departure Area (TLOF), Safety Zone, and Aircraft Parking Position are sequentially set. Furthermore, the bounding box tree detection method includes: setting up a bounding box tree and performing bounding box intersection detection, with the following steps: Based on the landable points within the second target landing area, establish landable point bounding boxes; based on the ground obstacle elevations, establish obstacle bounding boxes; and construct an obstacle avoidance bounding box tree, combining the landable point bounding boxes and obstacle bounding boxes into an obstacle avoidance bounding box tree. The obstacle avoidance bounding box tree consists of a root node and a series of child nodes, each node corresponding to a bounding box. Starting from the root node of the obstacle avoidance bounding box tree, bounding box intersection detection is performed layer by layer. For each node, it is first checked whether the landable bounding box intersects with the obstacle bounding box. When the landable bounding box intersects with the obstacle bounding box, the hierarchical bounding box tree is updated, and the child nodes of the node are checked again. When the landable bounding box does not intersect with the primitive, the coordinates of the landable bounding box point that satisfies the primitive non-intersection is the initial landing point coordinate of the helipad. Furthermore, the flight restriction surface is a three-dimensional space volume. A bounding box tree of the flight restriction surface is established, and the coordinates of the initial landing point of the parking apron are traversed. For each traversed initial landing point coordinate, the collision detection of the intersecting initial landing point coordinates of the parking apron is performed by the ray and triangle intersection detection algorithm until the coordinates of the flight restriction surface that satisfy the condition that the primitives do not intersect are found. The coordinates of the flight restriction surface that satisfy the condition that the primitives do not intersect are output as the final landing point of the aircraft. Furthermore, the ray-triangle intersection detection algorithm method includes: Step S10: Define the ray and triangle. The ray is defined with the aircraft's landing point coordinates as the origin and the normal vector of the flight restriction surface as the direction vector. The triangle is defined by the coordinates of the three vertices of the flight restriction surface. The Möller-Trumbore algorithm is used to calculate the intersection of the ray and the triangle, and the parameters of the ray and the triangle are calculated. ; Step S11: Determine if the ray intersects the triangle: If the calculated parameters satisfy If , it means that the ray intersects the triangle; if or If the intersection point is located on the extension of the ray, it means that the ray does not intersect the triangle; For all flight restriction surfaces, repeat steps S10 to S11 to check the collision between the initial landing point of the parking apron and each flight restriction surface. If any flight restriction surface intersects with the initial landing point of the parking apron, it means that the landing point does not meet the final landing point of the aircraft. If none of the flight restriction surfaces intersect with the initial landing point of the parking apron, it means that the landing point meets the final landing point of the aircraft. An autonomous landing site selection system for electric vertical takeoff and landing (EVTOL) aircraft in complex urban environments is provided, which is based on the aforementioned autonomous landing site selection method for EVTOL aircraft in complex urban environments, and includes: The first target landing area setting module for the aircraft delineates the pre-landing area of ​​the aircraft, collects evaluation data within the pre-landing area, creates landing evaluation coefficients based on the evaluation data, and obtains the first target landing area; The second target landing area setting module collects real-time flight characteristic data and slope identification data of the aircraft within the first target landing area. The real-time flight characteristic data generates real-time ground obstacle elevations through a trained aerial triangulation model. The ground obstacle elevations and DEM data are fused using ArcGIS to obtain three-dimensional ground obstacle data (DSM). The second target landing area of ​​the aircraft's landing point is delineated by combining the three-dimensional ground obstacle data (DSM) and the slope identification data. The initial landing point setting module for the aircraft uses an obstacle avoidance bounding box tree to perform intersection detection on the second target landing area and obtain the coordinates of the initial landing point for the aircraft that satisfies the condition that the bounding box primitives do not intersect. The final landing point setting module for the aircraft constructs the flight restriction surface during the aircraft's landing, establishes a bounding box tree for the flight restriction surface, and performs collision detection on the coordinates of the initial landing point using a ray and triangle intersection detection algorithm. The output coordinates of the flight restriction surface that satisfy the condition that primitives do not intersect are the coordinates of the final landing point of the aircraft.

[0006] The technical effects and advantages of the present invention regarding an autonomous landing site selection method for electric vertical take-off and landing aircraft in complex urban environments are as follows: This invention delineates the pre-landing area of ​​an aircraft, collects evaluation data within the pre-landing area, creates landing evaluation coefficients based on the evaluation data, and obtains a first target landing area. It then collects real-time flight characteristic data and slope identification data of the aircraft within the first target landing area. The real-time flight characteristic data is used to generate real-time ground obstacle elevations through a trained aerial triangulation model. The ground obstacle elevations and DEM data are fused using ArcGIS to obtain three-dimensional ground obstacle data (DSM). Finally, the second target landing area of ​​the aircraft's landing point is delineated by combining the three-dimensional ground obstacle data (DSM) and the slope identification data. By using an obstacle avoidance bounding box tree to perform intersection detection on the second target landing area, the coordinates of the initial landing point of the aircraft that satisfy the condition that the bounding box primitives do not intersect are obtained. A flight restriction surface is constructed for the aircraft during landing, and a bounding box tree of the flight restriction surface is established. Collision detection is performed on the coordinates of the initial landing point using a ray and triangle intersection detection algorithm. The coordinates of the flight restriction surface that satisfy the condition that the primitives do not intersect are the coordinates of the final landing point of the aircraft. It is capable of identifying and landing in safe landing areas without landmarks, realizing emergency landing of the aircraft, enabling the aircraft to perceive the surrounding environment, and giving the aircraft more intelligent judgment. Attached Figure Description

[0007] Figure 1 This is a schematic diagram of an autonomous landing site selection method for an electric vertical take-off and landing aircraft in a complex urban environment according to the present invention; Figure 2 This is a schematic diagram of an autonomous landing site selection system for an electric vertical take-off and landing aircraft in complex urban environments, according to the present invention. Detailed Implementation

[0008] 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.

[0009] Example 1 Please see Figure 1 As shown in this embodiment, a method for autonomous landing site selection of an electric vertical takeoff and landing aircraft in complex urban environments includes: Step S1: Delineate the pre-landing area of ​​the aircraft, collect evaluation data within the pre-landing area, create landing evaluation coefficients based on the evaluation data, and obtain the first target landing area; Step S2: Collect real-time flight characteristic data and slope identification data of the aircraft in the first target landing area. The real-time flight characteristic data generates real-time ground obstacle elevation through a trained aerial triangulation model. The ground obstacle elevation and DEM data are fused through ArcGIS to obtain ground obstacle 3D data (DSM). The second target landing area of ​​the aircraft landing point is delineated by combining the ground obstacle 3D data (DSM) and slope identification data. Step S3: By avoiding the bounding box tree through obstacles, perform intersection detection on the landing area of ​​the second target to obtain the coordinates of the initial landing point of the aircraft that satisfies the condition that the bounding box primitives do not intersect. Step S4: Construct the flight restriction surface for the aircraft landing, establish the bounding box tree of the flight restriction surface, and perform collision detection on the coordinates of the initial landing point using the ray and triangle intersection detection algorithm. The coordinates of the flight restriction surface that satisfy the condition that primitives do not intersect are the coordinates of the final landing point of the aircraft.

[0010] The pre-landing area for the aircraft is set up by staff, and the evaluation data includes the height of ground buildings, the wind speed when the aircraft lands, the runway length of the pre-landing area, and the restricted flight area; the wind speed when the aircraft lands is collected by an anemometer, and the height of ground buildings, the runway length of the pre-landing area, and the restricted flight area are obtained by radar equipment. Methods for determining whether the aircraft's pre-landing area is the primary target landing area include: When the predicted landing evaluation coefficient is greater than or equal to the preset landing evaluation coefficient threshold, the pre-landing area of ​​the aircraft is determined to be the first target landing area. When the predicted landing evaluation coefficient is less than the preset landing evaluation coefficient threshold, the pre-landing area of ​​the aircraft is determined to be the first target landing area. The preset landing evaluation coefficient threshold was set by the staff during the experimental phase. By collecting data on the height of different ground buildings in the pre-landing area, the wind speed during the aircraft's landing, the runway length in the pre-landing area, and the flight restriction area, the corresponding landing evaluation coefficients were recorded, and the average of multiple landing evaluation coefficients was taken as the preset landing evaluation coefficient threshold. Similarly, the ground obstacle 3D data DSM threshold and the slope recognition data threshold were set. The aerial triangulation model is trained using historical data, which includes historical flight characteristic data, including aircraft position and attitude data and ground reference data. The aircraft position and attitude data includes the altitude of the airborne photography equipment, forward overlap rate, lateral overlap rate, and resolution imagery. The ground reference data includes DEM (Digital Elevation Model) data, which is acquired through radar equipment. The slope identification data includes slope values ​​and slope ranges, acquired through a laser scanner. Historical flight training data is acquired through a flight data recorder installed inside the aircraft fuselage. The airborne photography equipment, at an altitude of 300m, acquires orthophotos of the first target landing area with a forward overlap rate of 75% and a lateral overlap rate of 70%, automatically obtaining high-resolution images with a size of 1km*1km and a resolution of 0.2cm. The training of the aerial triangulation model includes: Collect k sets of flight training data in advance, where k is an integer greater than 1; Flight training data is divided into training and test sets. A classifier is constructed using real-time flight feature data from the training set as input data and ground obstacle elevation data from the training set as output data. The classifier is trained to obtain an initial classifier. The initial classifier is then tested using the test set. When the root mean square error between the predicted ground obstacle elevation and the actual ground obstacle elevation is within 0.01, the classifier that meets the preset accuracy is output as the aerial triangulation model. The aerial triangulation model can be one of the following: Naive Bayes algorithm, random forest regression, or neural network regression model. The root mean square error expression is as follows:

[0011] In the formula, The root mean square error, The number of ground obstacles by elevation. For the predicted ground obstacle elevation, This represents the actual elevation of the ground obstacle. The method for delineating the second target landing area of ​​the aircraft landing point by combining three-dimensional ground obstacle data (DSM) and slope recognition data includes: By setting the three-dimensional obstacle data (DSM) threshold and slope recognition data threshold by staff, obstacles exceeding the threshold in height and areas with ground slope exceeding a certain limit are removed, and potential landing areas that meet the requirements for safe landing are selected as the second target landing areas for the aircraft landing point. The method for constructing the flight restriction surface includes: By acquiring the 3D ground obstacle data (DSM), obstacles in the second target landing area are identified. Based on the aircraft's flight mission, the operational limitations of the aircraft under given terrain and obstacle conditions are assessed, thereby constructing the flight restriction surface. Within the flight restriction surface, the Final Approach and Takeoff Area (FATO), Clearway, Touchdown and Departure Area (TLOF), Safety Zone, and Aircraft Parking Position are sequentially set. Final Approach and Takeoff Area (FATO): This is the main area for aircraft approach and takeoff; it is usually located at the end of the runway and needs to be long and wide enough to accommodate aircraft approach and takeoff operations; the flight restriction surfaces of the FATO can be used to restrict the location of the landing point and ensure that the aircraft has enough space during approach and takeoff. Clearance area: The clearance area is the area connecting the FATO and the TLOF (Touch and Takeoff Area); it needs to provide sufficient space to allow aircraft to have safe vertical and horizontal separation during approach and takeoff; the flight restriction surface of the clearance area is used to limit the height of obstacles around the landing point to ensure the safe operation of the aircraft. Touch-off and Take-off Area (TLOF): This is the area where the aircraft touches down and takes off, located between the FATO and the safe zone. It needs to provide a flat, unobstructed surface to ensure smooth operation of the aircraft during touch-off and take-off. The flight limiting surface of the TLOF is used to limit the height of obstacles around the landing point to ensure the safe operation of the aircraft. Safety zones and aircraft parking positions: A safety zone is the area around an aircraft landing point, which needs to provide sufficient space to ensure the safe operation of the aircraft; an aircraft parking position is the area where aircraft are parked and take-off and landing operations are performed, which needs to have sufficient area and support facilities; the flight restriction surfaces of the safety zone and aircraft parking positions are used to limit the height of surrounding obstacles to ensure the safe parking and take-off and landing operations of the aircraft. The design of these areas ensures that aircraft have sufficient space and safety during landing and takeoff. The bounding box tree detection method includes: setting up a bounding box tree and performing bounding box intersection detection, with the following steps: Based on the landable points within the second target landing area, establish landable point bounding boxes; based on the ground obstacle elevations, establish obstacle bounding boxes; and construct an obstacle avoidance bounding box tree, combining the landable point bounding boxes and obstacle bounding boxes into an obstacle avoidance bounding box tree. The obstacle avoidance bounding box tree consists of a root node and a series of child nodes, each node corresponding to a bounding box. Starting from the root node of the obstacle avoidance bounding box tree, bounding box intersection detection is performed layer by layer. For each node, it is first checked whether the landable bounding box intersects with the obstacle bounding box. When the landable bounding box intersects with the obstacle bounding box, the hierarchical bounding box tree is updated, and the child nodes of the node are checked again. When the landable bounding box does not intersect with the primitive, the coordinates of the landable bounding box point that satisfies the primitive non-intersection is the initial landing point coordinate of the helipad. The flight restriction surface is a three-dimensional space volume. A bounding box tree of the flight restriction surface is established. The coordinates of the initial landing point of the parking apron are traversed. For each initial landing point coordinate of the parking apron that is traversed, the collision detection of the intersecting initial landing point coordinates of the parking apron is performed by the ray and triangle intersection detection algorithm until the coordinates of the flight restriction surface that satisfy the condition that the primitives do not intersect are found. The coordinates of the flight restriction surface that satisfy the condition that the primitives do not intersect are output as the final landing point of the aircraft. The ray-triangle intersection detection algorithm includes: Step S10: Define the ray and triangle. The ray is defined with the aircraft's landing point coordinates as the origin and the normal vector of the flight restriction surface as the direction vector. The triangle is defined by the coordinates of the three vertices of the flight restriction surface. The Möller-Trumbore algorithm is used to calculate the intersection of the ray and the triangle, and the parameters of the ray and the triangle are calculated. ; Step S11: Determine if the ray intersects the triangle: If the calculated parameters satisfy If , it means that the ray intersects the triangle; if or If the intersection point is located on the extension of the ray, it means that the ray does not intersect the triangle; For all flight restriction surfaces, repeat steps S10 to S11 to check the collision between the initial landing point of the parking apron and each flight restriction surface. If any flight restriction surface intersects with the initial landing point of the parking apron, it means that the landing point does not meet the final landing point of the aircraft. If none of the flight restriction surfaces intersect with the initial landing point of the parking apron, it means that the landing point meets the final landing point of the aircraft. In this embodiment, by delineating the pre-landing area of ​​the aircraft, collecting evaluation data within the pre-landing area, creating landing evaluation coefficients based on the evaluation data, and obtaining the first target landing area, real-time flight characteristic data and slope identification data of the aircraft within the first target landing area are collected. The real-time flight characteristic data generates real-time ground obstacle elevations through a trained aerial triangulation model. The ground obstacle elevations and DEM data are fused using ArcGIS to obtain three-dimensional ground obstacle data (DSM). The second target landing area of ​​the aircraft landing point is delineated by combining the three-dimensional ground obstacle data (DSM) and the slope identification data. By using an obstacle avoidance bounding box tree to perform intersection detection on the second target landing area, the coordinates of the initial landing point of the aircraft that satisfy the condition that the bounding box primitives do not intersect are obtained. A flight restriction surface is constructed for the aircraft during landing, and a bounding box tree of the flight restriction surface is established. Collision detection is performed on the coordinates of the initial landing point using a ray and triangle intersection detection algorithm. The coordinates of the flight restriction surface that satisfy the condition that the primitives do not intersect are the coordinates of the final landing point of the aircraft. It is capable of identifying and landing in safe landing areas without landmarks, realizing emergency landing of the aircraft, enabling the aircraft to perceive the surrounding environment, and giving the aircraft more intelligent judgment.

[0012] Example 2 Please see Figure 2 As shown, for parts not described in detail in this embodiment, please refer to the description in Embodiment 1. An autonomous landing site selection system for electric vertical takeoff and landing (EVTOL) aircraft in complex urban environments is provided, comprising: The first target landing area setting module for the aircraft delineates the pre-landing area of ​​the aircraft, collects evaluation data within the pre-landing area, creates landing evaluation coefficients based on the evaluation data, and obtains the first target landing area; The second target landing area setting module collects real-time flight characteristic data and slope identification data of the aircraft within the first target landing area. The real-time flight characteristic data generates real-time ground obstacle elevations through a trained aerial triangulation model. The ground obstacle elevations and DEM data are fused using ArcGIS to obtain three-dimensional ground obstacle data (DSM). The second target landing area of ​​the aircraft's landing point is delineated by combining the three-dimensional ground obstacle data (DSM) and the slope identification data. The initial landing point setting module for the aircraft uses an obstacle avoidance bounding box tree to perform intersection detection on the second target landing area and obtain the coordinates of the initial landing point for the aircraft that satisfies the condition that the bounding box primitives do not intersect. The final landing point setting module for the aircraft constructs the flight restriction surface during landing, establishes a bounding box tree for the flight restriction surface, and performs collision detection on the coordinates of the initially selected landing point using a ray and triangle intersection detection algorithm. The coordinates of the flight restriction surface points that satisfy the condition of non-intersecting primitives are the coordinates of the final landing point of the aircraft. All modules are connected via wired and / or wireless means to achieve data transmission between modules.

[0013] Since the electronic device described in this embodiment is the electronic device used to implement the autonomous landing site selection method for electric vertical take-off and landing aircraft in complex urban environments according to the embodiments of this application, those skilled in the art can understand the specific implementation method and various variations of the electronic device in this embodiment based on the autonomous landing site selection method for electric vertical take-off and landing aircraft in complex urban environments described in the embodiments of this application. Therefore, how the electronic device implements the method in the embodiments of this application will not be described in detail here. Any electronic device used by those skilled in the art to implement the electricity meter operation status management method according to the embodiments of this application falls within the scope of protection of this application.

[0014] The above formulas are all dimensionless calculations. The formulas are derived from software simulations based on a large amount of collected data to obtain the most recent real-world results. The preset parameters and thresholds in the formulas are set by those skilled in the art according to the actual situation.

[0015] The above description is merely a preferred embodiment of the present invention, and the scope of protection of the present invention is not limited to the above embodiments. All technical solutions falling within the scope of the present invention's concept are within the scope of protection of the present invention. It should be noted that for users of ordinary technical skills, any improvements and modifications made without departing from the principles of the present invention should also be considered within the scope of protection of the present invention.

Claims

1. A method for autonomous landing site selection in complex urban environments for electric vertical takeoff and landing aircraft, characterized in that: The method for autonomous landing site selection in complex urban environments using electric vertical takeoff and landing aircraft includes the following steps: Step S1: Delineate the pre-landing area of ​​the aircraft, collect evaluation data within the pre-landing area, create landing evaluation coefficients based on the evaluation data, and obtain the first target landing area; Step S2: Collect real-time flight characteristic data and slope identification data of the aircraft in the first target landing area. The real-time flight characteristic data generates real-time ground obstacle elevation through a trained aerial triangulation model. The ground obstacle elevation and DEM data are fused through ArcGIS to obtain ground obstacle 3D data (DSM). The second target landing area of ​​the aircraft landing point is delineated by combining the ground obstacle 3D data (DSM) and slope identification data. Step S3: By avoiding the bounding box tree through obstacles, perform intersection detection on the landing area of ​​the second target to obtain the coordinates of the initial landing point of the aircraft that satisfies the condition that the bounding box primitives do not intersect. Step S4: Construct the flight restriction surface for the aircraft landing, establish the bounding box tree of the flight restriction surface, and perform collision detection on the coordinates of the initial landing point using the ray and triangle intersection detection algorithm. The coordinates of the flight restriction surface that satisfy the condition that primitives do not intersect are the coordinates of the final landing point of the aircraft.

2. The method for autonomous landing site selection of an electric vertical takeoff and landing aircraft in complex urban environments according to claim 1, characterized in that, The pre-landing area for the aircraft is set up by staff, and the evaluation data includes the height of ground buildings, wind speed at the time of aircraft landing, runway length of the pre-landing area, and flight restriction area. The wind speed at the time of aircraft landing is collected by an anemometer, and the height of ground buildings, runway length of the pre-landing area, and flight restriction area are obtained by radar equipment.

3. The method for autonomous landing site selection of an electric vertical takeoff and landing aircraft in complex urban environments according to claim 1, characterized in that, The aerial triangulation model is trained using historical data, which includes historical flight characteristic data, including aircraft position and attitude data and ground reference data. The aircraft position and attitude data includes the altitude, forward overlap rate, lateral overlap rate, and resolution imagery of the airborne photography equipment. The ground reference data includes DEM (Digital Elevation Model) data, which is acquired through radar equipment. The slope identification data includes slope values ​​and slope ranges, which are collected by a laser scanner. Historical flight training data is collected through a flight data recorder installed inside the aircraft fuselage.

4. The method for autonomous landing site selection of an electric vertical takeoff and landing aircraft in complex urban environments according to claim 1, characterized in that, The training of the aerial triangulation model includes: Collect k sets of flight training data in advance, where k is an integer greater than 1; Flight training data is divided into training and test sets. A classifier is constructed using real-time flight feature data from the training set as input data and ground obstacle elevation data from the training set as output data. The classifier is trained to obtain an initial classifier. The initial classifier is then tested using the test set. When the root mean square error between the predicted ground obstacle elevation and the actual ground obstacle elevation is within 0.01, the classifier that meets the preset accuracy is output as the trained aerial triangulation model. The aerial triangulation model can be one of the following: Naive Bayes algorithm, random forest regression, or neural network regression model. The root mean square error expression is as follows: ； In the formula, The root mean square error, The number of ground obstacles by elevation. For the predicted ground obstacle elevation, This represents the actual elevation of the ground obstacle.

5. The method for autonomous landing site selection of an electric vertical takeoff and landing aircraft in complex urban environments according to claim 1, characterized in that, The method for constructing the flight restriction surface includes: By acquiring the 3D ground obstacle data (DSM), obstacles in the second target landing area are identified. Based on the aircraft's flight mission, the operational limitations of the aircraft under given terrain and obstacle conditions are assessed, thereby constructing the flight restriction surface. Within the flight restriction surface, the final approach and takeoff area (FATO), clear lane, touchdown and departure area (TLOF), safety zone, and aircraft parking position are sequentially set.

6. The method for autonomous landing site selection of an electric vertical takeoff and landing aircraft in complex urban environments according to claim 1, characterized in that, The bounding box tree detection method includes: setting up a bounding box tree and performing bounding box intersection detection, with the following steps: Based on the landable points within the second target landing area, establish landable point bounding boxes; based on the ground obstacle elevations, establish obstacle bounding boxes; and construct an obstacle avoidance bounding box tree, combining the landable point bounding boxes and obstacle bounding boxes into an obstacle avoidance bounding box tree. The obstacle avoidance bounding box tree consists of a root node and a series of child nodes, each node corresponding to a bounding box. Starting from the root node of the obstacle avoidance bounding box tree, bounding box intersection detection is performed layer by layer. For each node, it is first checked whether the landable bounding box intersects with the obstacle bounding box. When the landable bounding box intersects with the obstacle bounding box, the hierarchical bounding box tree is updated, and the child nodes of the node are checked again. When the landable bounding box does not intersect with the primitive, the coordinates of the landable bounding box point that satisfies the primitive non-intersection is the initial landing point coordinate of the helipad.

7. The method for autonomous landing site selection of an electric vertical takeoff and landing aircraft in complex urban environments according to claim 6, characterized in that, The flight restriction surface is a three-dimensional space volume. A bounding box tree of the flight restriction surface is established, and the coordinates of the initial landing point of the apron are traversed. For each traversed initial landing point coordinate, a collision detection algorithm is used to detect the intersecting initial landing point coordinates of the apron until the coordinates of the flight restriction surface that satisfy the condition that the primitives do not intersect are found. The coordinates of the flight restriction surface that satisfy the condition that the primitives do not intersect are output as the coordinates of the initial landing point of the aircraft.

8. The method for autonomous landing site selection of an electric vertical takeoff and landing aircraft in complex urban environments according to claim 7, characterized in that, The ray-triangle intersection detection algorithm includes: Step S10: Define the ray and triangle. The ray is defined with the aircraft's landing point coordinates as the origin and the normal vector of the flight restriction surface as the direction vector. The triangle is defined by the coordinates of the three vertices of the flight restriction surface. The Möller-Trumbore algorithm is used to calculate the intersection of the ray and the triangle, and the parameters of the ray and the triangle are calculated. ; Step S11: Determine if the ray intersects the triangle: If the calculated parameters satisfy If , it means that the ray intersects the triangle; if or If the intersection point is located on the extension of the ray, it means that the ray does not intersect the triangle; For all flight restriction surfaces, repeat steps S10 to S11 to detect the collision between the initial landing point of the parking apron and each flight restriction surface. If any flight restriction surface intersects with the initial landing point of the parking apron, it means that the landing point does not meet the final landing point of the aircraft. If none of the flight restriction surfaces intersect with the initial landing point of the parking apron, it means that the landing point meets the final landing point of the aircraft.

9. A system for autonomous landing site selection in complex urban environments for electric vertical takeoff and landing (EVTOL) aircraft, used to implement the method for autonomous landing site selection in complex urban environments for EVTOL aircraft as described in any one of claims 1 to 8, characterized in that, include: The first target landing area setting module for the aircraft delineates the pre-landing area of ​​the aircraft, collects evaluation data within the pre-landing area, creates landing evaluation coefficients based on the evaluation data, and obtains the first target landing area; The second target landing area setting module for the aircraft collects real-time flight characteristic data and slope identification data of the aircraft within the first target landing area. The real-time flight characteristic data generates real-time ground obstacle elevation through a trained aerial triangulation model. The ground obstacle elevation and DEM data are then fused using ArcGIS to obtain three-dimensional ground obstacle data (DSM). By combining 3D ground obstacle data (DSM) and slope identification data, a second target landing area was delineated for the aircraft's landing point. The initial landing point setting module for the aircraft uses an obstacle avoidance bounding box tree to perform intersection detection on the second target landing area and obtain the coordinates of the initial landing point for the aircraft that satisfies the condition that the bounding box primitives do not intersect. The final landing point setting module for the aircraft constructs the flight restriction surface during the aircraft's landing, establishes a bounding box tree for the flight restriction surface, and performs collision detection on the coordinates of the initial landing point using a ray and triangle intersection detection algorithm. The output coordinates of the flight restriction surface that satisfy the condition that primitives do not intersect are the coordinates of the final landing point of the aircraft.

Images