Method for generating lift-limiting flight forbidding and avoiding area based on density cluster
By applying the DBSCAN clustering algorithm and the external polygon space closure algorithm in aircraft mission planning, density clustering and polygon closure are performed on discrete elevation points, which solves the accuracy, automation and shape limitations of the division of flight-free zones in the existing technology, and achieves higher accuracy and automation division of flight-free zones, improving the scientificity and safety of flight-free zones.
Patent Information
- Application Number
- CN202411932492.8
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-12-26
- Publication Date
- 2025-05-27
AI Technical Summary
The existing methods of escaping flight-free zone division have accuracy problems, the inability to automatically divide the number of escaping flight-free zones, and the shape limitations, making it difficult to meet the safety needs of aircraft in complex mission environments.
The DBSCAN clustering algorithm based on density clusters is used to cluster the elevation discrete points, generate any polygon cluster, and divide the cluster set into a forbidden flight zone through the external polygon space closure algorithm to achieve optimal division.
It improves the accuracy and automation of the division of flight-free zones, can adapt to the division of flight-free zones in complex shapes, and enhances the scientificity and safety of flight missions.
Smart Images
Figure CN120048160A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of aircraft mission planning, and particularly relates to a method for generating a ceiling avoidance flight area based on density clusters. Background Art
[0002] In modern warfare, the setting of avoidance flight areas is usually for ensuring the flight safety of aircraft, especially in complex battlefield environments or dangerous airspace environments. In more complex mission scenarios, the areas where flight missions are executed are very likely to include complex terrains, adverse weather conditions, and even phenomena of electromagnetic interference, as well as areas where the aircraft cannot reach certain flight altitudes.
[0003] To ensure the effectiveness and accuracy of mission planning and make the execution of missions more scientific, the precise construction of avoidance flight areas is crucial for the flight of aircraft. It can provide accurate geospatial information that the aircraft needs to fly around. These information can be used as a basis for adjusting parameters such as the flight trajectory, speed, altitude, and attitude of the aircraft in the flight trajectory planning, enabling the aircraft to execute flight missions more safely, avoiding collisions with ground obstacles, and preventing flight missions from affecting some specific areas. In summary, the division of avoidance flight areas has a profound impact on the improvement of the practicability and applicability of aircraft, and also has high value for the development of mission planning expertise.
[0004] The current commonly used method for dividing the ceiling avoidance flight area has the following process: First, after obtaining the discrete points of the elevation threshold, the division method is to cluster the discrete points using k-means and its derivative algorithms, and close the avoidance flight area using the rectangular corner points. Using this method easily leads to the following problems:
[0005] (1) Precision problem of the confinement flight area: Considering the flight planning distance of the aircraft, when using k-means and its derivative algorithms, since the algorithm is sensitive to the initial clustering center, during the generation process of the avoidance flight area, the clustering result has a deviation, resulting in inaccurate division of the avoidance flight area, problems such as the avoidance flight area being too large, losing the effective planning distance, and thus the planning path cannot generate a scheme, causing waste of the flyable area.
[0006] (2) Automatic division of the number of avoidance flight areas: When using k-means and its derivative algorithms for clustering, the number of clustering clusters needs to be manually specified, and the number value of the clustering clusters cannot be determined from the algorithm, thus increasing the complexity of manual intervention. Moreover, the selection of this value often depends on experience or multiple trials, resulting in the clustering process being unable to achieve automation.
[0007] (3) Limitation of the shape of the no-fly zone: When using the k-means and its derivative algorithms, they are not good at clustering discrete points of convex polygons and cannot meet the requirements for dividing no-fly zones of specific shapes. Summary of the Invention
[0008] (1) Technical problems to be solved
[0009] The technical problem to be solved by the present invention is: how to propose a ceiling no-fly zone generation method based on density clusters, that is, using the DBSCAN clustering algorithm, which can generate one or more arbitrary polygon clusters according to the clustering of discrete elevation points, and complete the division of one or more no-fly zones by closing the circumscribed polygon space of the generated cluster set, and further achieve the optimal division of the no-fly zone while meeting the ceiling requirements.
[0010] (2) Technical solutions
[0011] To solve the above technical problems, the present invention provides a ceiling no-fly zone generation method based on density clusters. The method is based on discrete elevation points, and the process of generating the no-fly zone includes two major parts:
[0012] Step 1: Clustering of discrete points;
[0013] Step 2: Circumscribed closed polygon of the clustering cluster. The specific flow chart is as Figure 1 shown.
[0014] In the method, the related definitions are as follows:
[0015] (1) Core Point: A point whose number of sample points within the neighborhood radius is greater than or equal to the minimum number of sample points is called a core object, also known as a core point;
[0016] (2) Border Point: A point that does not belong to the core point but is within the neighborhood of a certain core point is called a border point;
[0017] (3) Noise Point: A sample that belongs neither to the core point nor to the border point is a noise point;
[0018] (4) Directly Density-Reachable: If p is a core point and q is within the neighborhood radius of p, then p is said to be directly density-reachable to q, and any core point is directly density-reachable to itself;
[0019] (5) Density-Reachable: If there exists a core point p 2 , p 3 , ……, p n , and p 1 to p2 Direct density reachability, p 2 to p 3 Direct density reachability, ……, p n-1 , to p n Direct density reachability, p n If direct density reachability from p to q, then 1 direct density reachability from p to q can be achieved.
[0020] The graphic description of the relevant definitions of this clustering is as Figure 2 shown.
[0021] Among them, in step 1, clustering of discrete points is performed:
[0022] Taking the elevation discrete points that meet the screening conditions as samples, performing adaptive clustering, and grouping the samples according to the relative position relationship; the density-based clustering algorithm (DBSCAN) uses the Euclidean distance to measure the distance between each discrete point;
[0023] The distance formula is:
[0024]
[0025] where p and q are two data points, p i and q i are their coordinate values in each dimension respectively, and n is the dimension of the data.
[0026] The flowchart related to the clustering of discrete points is as Figure 3 shown;
[0027] The clustering process of discrete points in step 1 includes:
[0028] Step 1.1: Finding core objects;
[0029] Step 1.2: Setting up temporary clustering clusters;
[0030] Step 1.3: Merging temporary clustering clusters;
[0031] Step 1.4: Traversing the remaining temporary clusters.
[0032] Among them, in step 1.1, according to the distribution of discrete points, the domain radius and the minimum number of sample points parameters are given in advance. The domain radius parameter is used to set the domain range of the object, and the minimum number of sample points parameter is a threshold for judging whether it is a core object; traversing each object and marking the core objects;
[0033] The formula for the number of data points in the neighborhood is: density(p) = |N Eps (p)|, where N Eps (p) represents the set of all data points in the neighborhood with p as the center and a radius of Eps.
[0034] In step 1.2, set up temporary clustering clusters; for each core object, the set of objects in its domain is used as a temporary clustering cluster. At the positions where noise points are found, a separate temporary cluster is set up for each noise point to ensure that no noise points are lost.
[0035] In step 1.3, merge the temporary clustering clusters; after completing step 1.2, each core object will have a temporary clustering cluster; check whether there are core objects in the temporary clustering cluster. If there are, merge the temporary clustering cluster corresponding to the core object with the current temporary clustering cluster into one temporary cluster, that is, density-reachable. Repeat this method until all density-reachable points in this temporary cluster are within this temporary cluster, then this temporary cluster becomes a clustering cluster.
[0036] In step 1.4, traverse the remaining temporary clusters; use the method in 1.3 to traverse the remaining temporary clusters to complete the clustering of density clusters.
[0037] The completed clustering effect is as Figure 4 shown.
[0038] Among them, in step 2, circumscribed polygons of individual clustering clusters are respectively performed;
[0039] For each sample cluster, connect the outermost points of the sample cluster to form a closed polygon.
[0040] The flow chart related to clustering is as Figure 5 shown;
[0041] Step 2 includes:
[0042] Step 2.1: Convert the coordinate system; convert the sample points within the sample cluster to the Gaussian coordinate system;
[0043] Step 2.2: Find the minimum point in the coordinate system; find the point with the smallest x coordinate when the y coordinate is the largest, denoted as point A;
[0044] Step 2.3: Rotate to find the minimum point; with A as the origin, scan clockwise along the positive x-axis ray to find the point scanned when the rotation angle is the smallest, denoted as point B; with B as the origin, scan clockwise along the AB direction ray to find the point scanned when the rotation angle is the smallest, denoted as point C;
[0045] Step 2.4: Rotate to find the starting point; according to the rotation method in 2.3, until the starting point A is found.
[0046] The completed polygon closing effect is as Figure 6 shown.
[0047] (III) Beneficial effects
[0048] Compared with the prior art, based on the DBSCAN clustering algorithm and the circumscribed polygon space closing algorithm, the present invention classifies the discrete points by density through the DBSCAN clustering algorithm on the basis of obtaining the discrete points above the elevation threshold in the space.
[0049] Innovations to be protected by the present invention:
[0050] (1) When selecting points in the area of the no-fly zone, the elevation data is effectively used to integrate the available geographical environment information, and the point set of the no-fly zone is effectively selected; when dividing the area, the more advanced algorithm idea of density clusters is used, and the DBSCAN clustering algorithm is used to cluster the above point set, realizing the refinement and uniqueness from the selection of points to the construction of surfaces.
[0051] (2) For the classification results of the constructed clusters, the circumscribed polygon closing method is used to construct the area of the no-fly zone, realizing a higher-precision division of the no-fly zone area and improving the rigor of the mission planning results.
[0052] Beneficial effects:
[0053] (1) The present invention is based on the relatively innovative density cluster clustering algorithm to judge and divide the no-fly zone, making the selection of the no-fly zone more reasonable and real-time, effectively avoiding the situation of repeated modification required in previous tasks, and achieving "avoiding as much as possible with the smallest area"; compared with the inaccurate and unrigorous use of traditional clustering algorithms and the inability to apply to the current mission planning situation, the division of the density cluster clustering method used in the present invention makes the execution of the task more scientific and rigorous.
[0054] (2) The advantage of the density cluster clustering algorithm is that it can independently judge the number of clusters in each category by using its internal perfect algorithm, realizing semi-automation of clustering to a great extent; for the division of the no-fly zone, the DBSCAN clustering algorithm can automatically realize the division of any shape. It meets the requirements of using irregular area no-fly zones in current mission planning and is more practical.
[0055] (3) In actual use, due to geographical environment limitations, common no-fly zone divisions often have over-delineation and mis-delineation, making the flight tasks of aircraft become cumbersome. The DBSCAN clustering algorithm used in the present invention is used to select and divide the no-fly zone, improving the accuracy of the no-fly zone division, avoiding the waste of flight areas caused by unreasonable selection of no-fly zones, and making the path planning more adaptable to the region and more scientifically accurate.
[0056] Its advantages can also be summarized as the following points:
[0057] Advantage 1: Improve the accuracy: It can discover clustering clusters of any shape, and can count the number of clusters by itself without inputting the number of clusters to be discovered, which is more flexible when processing data sets, making the clustering more accurate and reasonable;
[0058] Advantage 2: Improve the planning efficiency: On this basis, through the circumscribed polygon space closing algorithm, the discrete points of this shape are generated into a circumscribed polygon, which can reduce the search space during path planning and improve the planning efficiency on the one hand;
[0059] Advantage 3: Stronger scalability: Since the method of dividing the ceiling avoidance flight area based on density clusters is introduced, the size of the avoidance flight area can be flexibly divided to meet the needs of autonomous longitudinal planning on the missile, making the autonomous planning more flexible and controllable. Brief Description of the Drawings
[0060] Figure 1 It is a flow chart for generating a ceiling avoidance flight area.
[0061] Figure 2 It is a diagram for describing clustering-related definitions.
[0062] Figure 3 It is a flow chart of the clustering algorithm.
[0063] Figure 4 It is a schematic diagram of the clustering effect.
[0064] Figure 5 It is a flow chart for generating a polygon no-fly zone.
[0065] Figure 6 It is a schematic diagram for generating a polygon no-fly zone. Detailed Implementation Manner
[0066] To make the objectives, contents, and advantages of the present invention clearer, the following further describes in detail the specific implementation manner of the present invention with reference to the accompanying drawings and embodiments.
[0067] To solve the above technical problems, the present invention provides a method for generating a ceiling avoidance flight area based on density clusters. The method is based on elevation discrete points, and the process of generating the avoidance flight area includes two major parts:
[0068] Step 1: Clustering of discrete points;
[0069] Step 2: Circumscribed closed polygon of the clustering cluster. The specific flow chart is as Figure 1 shown.
[0070] In the method, the related definitions are as follows:
[0071] (1) Core Object: A point whose number of sample points within the neighborhood radius is greater than or equal to the minimum number of sample points is called a core object, also known as a core point;
[0072] (2) Border Point: A point that does not belong to the core point but is within the neighborhood of a certain core point is called a border point;
[0073] (3) Noise Point: A sample that belongs to neither the core point nor the border point is a noise point;
[0074] (4) Directly Density-Reachable: If p is a core point and q is within the neighborhood radius of p, then p is said to be directly density-reachable to q, and any core point is directly density-reachable to itself;
[0075] (5) Density-Reachable: If there exists a core point p 2 , p 3 , ……, p n , and p 1 is directly density-reachable to p 2 , p 2 is directly density-reachable to p 3 , ……, p n-1 , and is directly density-reachable to p n , p n is directly density-reachable to q, then p 1 is density-reachable to q.
[0076] The graphic description of the clustering-related definitions is as Figure 2 shown.
[0077] Among them, in step 1, clustering of discrete points is performed:
[0078] Taking the elevation discrete points that meet the screening conditions as samples, performing adaptive clustering, and grouping the samples according to the relative position relationship; The Density-Based Spatial Clustering of Applications with Noise (DBSCAN) algorithm uses the Euclidean distance to measure the distance between each discrete point;
[0079] The distance formula is:
[0080]
[0081] where p and q are two data points, and p i and q i are their coordinate values in each dimension respectively, and n is the dimension of the data.
[0082] The flow chart related to the clustering of discrete points is as Figure 3 shown;
[0083] The clustering process of discrete points in step 1 includes:
[0084] Step 1.1: Find the core object;
[0085] Step 1.2: Set up temporary clustering clusters;
[0086] Step 1.3: Merge the temporary clustering clusters;
[0087] Step 1.4: Traverse the remaining temporary clusters.
[0088] Among them, in the said Step 1.1, according to the distribution of discrete points, the domain radius and the minimum number of sample points are given in advance. The domain radius parameter is used to set the domain range of the object, and the minimum number of sample points is a threshold for judging whether it is a core object; traverse each object and mark the core object;
[0089] The formula for the number of data points in the neighborhood is: density(p) = |N Eps (p)|, where N Eps (p) represents the set of all data points in the neighborhood with p as the center and a radius of Eps.
[0090] In the said Step 1.2, set up temporary clustering clusters; the set of objects within the domain of each core object is used as a temporary clustering cluster. In the positions where noise points are found, a separate temporary cluster is set for the noise point to ensure that the noise point is not lost.
[0091] In the said Step 1.3, merge the temporary clustering clusters; after completing Step 1.2, each core object will have a temporary clustering cluster; check whether there is a core object in the temporary clustering cluster. If there is, merge the temporary clustering cluster corresponding to the core object with the current temporary clustering cluster into one temporary cluster, that is, density-reachable. Repeat this method until all density-reachable points in this temporary cluster are in this temporary cluster, then this temporary cluster becomes a clustering cluster.
[0092] In the said Step 1.4, traverse the remaining temporary clusters; use the method in 1.3 to traverse the remaining temporary clusters to complete the clustering of density clusters.
[0093] The completed clustering effect is as Figure 4 shown.
[0094] Among them, in the said Step 2, the circumscribed polygons of individual clustering clusters are carried out respectively;
[0095] For each sample cluster, connect the outermost points of the sample cluster to form a closed polygon.
[0096] The flowchart related to clustering is as Figure 5 shown;
[0097] The said Step 2 includes:
[0098] Step 2.1: Convert the coordinate system; convert the sample points within the sample cluster to the Gaussian coordinate system;
[0099] Step 2.2: Find the minimum point in the coordinate system; find the point with the smallest x-coordinate when the y-coordinate is the largest, denoted as point A;
[0100] Step 2.3: Rotate to find the minimum point; with A as the origin, scan clockwise along the positive x-axis ray, and find the point scanned when the rotation angle is the smallest, denoted as point B; with B as the origin, scan clockwise along the AB direction ray, and find the point scanned when the rotation angle is the smallest, denoted as point C;
[0101] Step 2.4: Rotate to find the starting point; according to the rotation method in 2.3, until the starting point A is found.
[0102] The polygon closing effect is completed as Figure 6 shown.
[0103] The above are only the preferred embodiments of the present invention. It should be noted that for those of ordinary skill in the art, without departing from the technical principle of the present invention, several improvements and deformations can still be made, and these improvements and deformations should also be regarded as the protection scope of the present invention.
Claims
1. A method for generating a ceiling no-fly zone based on density clusters, characterized in that: The method is based on discrete points of elevation, and the process of generating no-fly zones includes two parts: Step 1: Clustering of discrete points; Step 2: Cluster the circumscribed closed polygons of the clusters.
2. The method for generating a ceiling no-fly-avoid zone based on density clusters according to claim 1, characterized in that: In the method, the relevant definitions are as follows: (1) Core object: A point whose number of sample points within the neighborhood radius is greater than or equal to the minimum number of sample points is called a core object, also known as a core point; (2) Edge points: Points that are not core points but are in the neighborhood of a core point are called edge points; (3) Noise points: samples that are neither core points nor edge points are noise points; (4) Density direct point: If p is a core point and q is within the neighborhood radius of p, then p is called a density direct point from q. Any core point is density directly accessible to itself; (5) Density reachable: If there are core points p2, p3, ..., p n , and the density from p1 to p2 is directly accessible, the density from p2 to p3 is directly accessible, ..., p n-1 , to p n Density direct, p n If p1 can directly reach density q, then p1 can reach density q.
3. The method for generating a ceiling no-fly-avoid zone based on density clusters according to claim 2, characterized in that: In step 1, clustering of discrete points is performed: The discrete points of elevation that meet the screening conditions are taken as samples, and adaptive clustering is performed to group the samples according to their relative position relationship; the clustering algorithm based on density clusters uses Euclidean distance to measure the distance between each discrete point; The distance formula is: Where p and q are two data points, p i and q i are their coordinate values in each dimension, and n is the dimension of the data.
4. The method for generating a ceiling no-fly avoidance zone based on density clusters according to claim 3, characterized in that: The clustering process of discrete points in step 1 includes: Step 1.1: Find the core object; Step 1.2: Set up a temporary cluster; Step 1.3: Merge temporary clusters; Step 1.4: Traverse the remaining temporary clusters.
5. The method for generating a ceiling no-fly-avoid zone based on density clusters according to claim 4, characterized in that: In the step 1.1, according to the distribution of discrete points, the domain radius and the minimum number of sample points are given in advance. The domain radius parameter is used to set the domain range of the object, and the minimum number of sample points parameter is a threshold for determining whether it is a core object; traverse each object and mark the core object; The formula for the number of data points in the neighborhood is: density(p)=|N Eps (p)|, where N Eps (p) represents the set of all data points in the neighborhood with a radius of Eps and centered at p.
6. The method for generating a ceiling no-fly-avoid zone based on density clusters according to claim 5, characterized in that: In the step 1.2, a temporary cluster is set; each core object takes the set of objects in its domain as a temporary cluster, and at the location where a noise point is found, the noise point is individually set with a temporary cluster to ensure that the noise point is not lost.
7. The method for generating a ceiling no-fly-avoid zone based on density clusters according to claim 6, characterized in that: In the step 1.3, the temporary clusters are merged; after completing step 1.2, each core object will have a temporary cluster; check whether there is a core object in the temporary cluster, if so, merge the temporary cluster corresponding to the core object with the current temporary cluster into one temporary cluster, that is, density reachable, repeat this method until all density direct points in the temporary cluster are in the temporary cluster, then the temporary cluster becomes a cluster cluster.
8. The method for generating a ceiling no-fly-avoid zone based on density clusters according to claim 7, characterized in that: In the step 1.4, the remaining temporary clusters are traversed; the remaining temporary clusters are traversed using the method in step 1.3 to complete the clustering of density clusters.
9. The method for generating a ceiling no-fly-avoid zone based on density clusters according to claim 8, characterized in that: The step 2 is to respectively generate the circumscribed polygons of a single cluster; For each sample cluster, the outermost points of the sample cluster are connected to form a closed polygon.
10. The method for generating a ceiling no-fly-avoid zone based on density clusters according to claim 9, characterized in that: The step 2 comprises: Step 2.1: Convert the coordinate system; convert the sample points in the sample cluster to the Gaussian coordinate system; Step 2.2: Find the minimum point in the coordinate system; find the point with the smallest x coordinate when the y coordinate is the largest, and record it as point A; Step 2.3: Rotate to find the minimum point; take A as the origin, scan the x-axis positive ray clockwise, find the point scanned when the rotation angle is the minimum, record it as point B; take B as the origin, scan the AB direction ray clockwise, find the point scanned when the rotation angle is the minimum, record it as point C; Step 2.4: Rotate to find the starting point; rotate according to 2.3 until the starting point A is found.