Dynamic detection and avoidance method of aircraft taxiing path conflicts at airport surface based on graph topology network

Abstract

The present invention relates to the technical field of traffic situation prediction, and in particular to a method for dynamic detection and avoidance of airport surface aircraft taxi path conflicts based on a graph topology network, the method comprising: constructing a taxi path graph network matrix based on the airport flight area surface road network, and extracting a taxi path subgraph network matrix according to aircraft taxiing rules; extracting a surface taxi path graph network matrix from the taxi path subgraph network matrix according to aircraft real-time motion data; constructing a prediction model, predicting a spatial taxi path graph network matrix based on the surface taxi path graph network matrix; extracting a conflict moment prediction edge set from the spatial taxi path graph network matrix; performing a backoff solution on the conflict moment prediction edge set, and generating a conflict adjustment path with the goal of minimizing the number of conflicts. The purpose is to solve the technical problems of insufficient future conflict prediction capability and poor taxi path optimization in the existing airport surface aircraft taxi path decision method.

Description

Technical Field

[0001] The present invention relates to the technical field of traffic situation prediction, and in particular to a method for dynamically detecting and avoiding conflicts in taxiing paths of aircraft at airports based on a graph topology network. Background Art

[0002] In airport aircraft taxi path command, since it is difficult to predict potential taxi path conflicts in advance, controllers can often only make emergency adjustments when conflicts occur. This not only affects the efficiency of flight arrival and departure, but may also cause airfield congestion and increase additional taxi distance, thereby reducing airport operating efficiency and passenger experience, and causing economic losses to airlines.

[0003] In the existing technology, the analysis method based on historical data uses statistical models to identify common path conflict patterns, but its prediction accuracy is limited when facing complex real-time scenarios; while the rule-based method uses the rules of the flight area surface management rules and expert experience to encode the formed rule set for path conflict detection. Although the rule-based method is easy to implement, it is not good at dealing with future uncertainties and cannot accurately describe the future taxiing behavior of aircraft. On the other hand, for the taxiing path optimization problem, the existing technology can find the theoretical shortest or optimal path through the graph search algorithm, and then plan the path for the aircraft; but because it does not fully consider the actual taxiing route allocation rules and restrictions in actual operation, the taxiing path optimization is poor, and its practicality is limited to a certain extent. Summary of the invention

[0004] In order to solve the technical problems that the existing airport surface aircraft taxi path decision method has insufficient future conflict prediction ability and poor taxi path optimization, the purpose of the present invention is to provide an airport surface aircraft taxi path conflict dynamic detection and avoidance method based on a graph topology network. The technical solution adopted is as follows:

[0005] A method for dynamic detection and avoidance of aircraft taxiing path conflicts at an airport surface based on a graph topology network, the method comprising:

[0006] The taxi path graph network matrix is ​​constructed based on the airport flight area surface road network, and the taxi path sub-graph network matrix is ​​extracted according to the aircraft taxiing rules;

[0007] Extracting the surface taxi path diagram network matrix from the taxi path sub-diagram network matrix according to the real-time aircraft movement data;

[0008] Construct a prediction model to predict the spatial taxi path map network matrix based on the surface taxi path map network matrix;

[0009] Extract the conflict moment prediction edge set from the network matrix of the space glide path graph;

[0010] The predicted edge set at the time of conflict is back-solved, and a conflict adjustment path is generated with the goal of minimizing the number of conflicts.

[0011] Furthermore, a taxi path graph network matrix is ​​constructed based on the airport flight area surface road network, and a taxi path subgraph network matrix is ​​extracted according to the aircraft taxiing rules, including:

[0012] Construct the airport surface aircraft taxi path quaternary graph network matrix, including the taxi path node matrix, the taxi path edge matrix, the edge eigenvalue matrix and the accessibility matrix between edges;

[0013] According to the preset aircraft taxiing rules, the aircraft taxiing path subgraph network matrix is ​​extracted from the quad graph network matrix.

[0014] Further, extracting the surface taxi path graph network matrix from the taxi path sub-graph network matrix according to the real-time aircraft motion data includes:

[0015] According to the real-time motion data of the aircraft, the position and velocity vector of the aircraft are obtained;

[0016] The position matching degree and direction matching degree of the aircraft taxi path subgraph are calculated using Gaussian function to obtain the comprehensive matching degree.

[0017] The network matrix of the aircraft surface taxi path diagram is extracted based on the comprehensive matching degree.

[0018] Furthermore, the expression for calculating the position matching degree of the aircraft taxi path subgraph using the Gaussian function is:

[0019]

[0020] In the formula, Indicates Aircraft The moment The position matching degree of the glide path subgraph; express The total number of path segments; Represents the standard deviation parameter, which is used to control the decay speed of the position matching degree; Indicates Aircraft The running position at the moment The Euclidean distance between the glide path subgraphs.

[0021] Furthermore, the expression for calculating the direction matching degree of the aircraft taxi path subgraph using the Gaussian function is:

[0022]

[0023] In the formula, express Direction matching degree; Indicates Aircraft The velocity vector at the moment; Indicates Aircraft The edge direction vector of the sliding path subgraph; express Model; express Model.

[0024] Furthermore, a prediction model is constructed to predict the spatial taxiing path map network matrix based on the surface taxiing path map network matrix, including:

[0025] The eigenvector of each edge in the network matrix of the aircraft surface taxi path diagram is forward updated and solved using graph neural network.

[0026] According to the predicted position of the aircraft at the future time, the irrelevant edges are removed from the graph structure edge set at the current time to obtain the graph structure edge set of the aircraft at the future time;

[0027] Calculate the eigenvector and eigenvalue of the aircraft at the future time;

[0028] Construct a network matrix of the space taxiing path diagram of the aircraft at future times.

[0029] Furthermore, the expression for calculating the eigenvector of the aircraft at the future time is:

[0030]

[0031] In the formula, Represents edge exist The feature vector at time, represents the total number of time steps, represents the time step; Represents edge exist The feature vector at the moment; represents the activation function; represents the first training weight matrix; represents the second training weight matrix; Indicates the number of iterations; Represents adjacent edges exist The feature vector at the moment; Represents edge The adjacency matrix of .

[0032] Furthermore, the conflict moment prediction edge set extracted from the network matrix of the space glide path graph includes:

[0033] Extract the number of aircraft on each edge of the network matrix of the spatial taxi path diagram;

[0034] The conflict is determined based on the number of aircraft, and the predicted edge set at the conflict moment is obtained.

[0035] Furthermore, the predicted edge set at the time of conflict is backed off to solve the conflict adjustment path with the goal of minimizing the number of conflicts, including:

[0036] Extract the conflicting aircraft objects from the conflict moment prediction edge set to obtain the conflicting aircraft object set;

[0037] Perform graph network fallback calculation on each aircraft in the conflicting aircraft object set, and extract the predicted graph network after fallback;

[0038] Perform maximum similarity processing on the sliding path subgraph network matrix to extract the conflict adjustment path network with the minimum cost of path change;

[0039] Verify that the conflict adjustment path network is effective in reducing conflicts and update aircraft taxi paths.

[0040] Furthermore, the maximum similarity processing is performed on the sliding path subgraph network matrix, and the expression of the conflict adjustment path network with the minimum cost of path change is extracted as follows:

[0041]

[0042] In the formula, Indicates conflict adjustment paths for each aircraft; Represents conflict path nodes; Indicates The first A glide path subgraph node; , indicating that the conflict node is different from the node in the conflict adjustment path; Indicates The first A glide path subgraph.

[0043] The present invention has the following beneficial effects:

[0044] The present invention provides a method for dynamic detection and avoidance of conflicts in taxiing paths of aircraft on the airport surface based on a graph topology network. By constructing a taxiing path graph network matrix that reflects the actual layout and operating rules of the airport flight area and combining the real-time motion data of the aircraft, accurate modeling of the taxiing path of the aircraft is achieved. In addition, individual aircraft are provided with their own specific graph structures, and the evolution process of different aircraft in the taxiing path graph network of the airport surface is described, so as to realize dynamic correction of the taxiing path of the aircraft and reduce the risk of conflict. By constructing a prediction model, the influence of the update of the aircraft behavior state at different times is considered to improve the prediction accuracy. Finally, the predicted edge set at the conflict time is extracted from the taxiing path graph network matrix of the prediction space, the real operating physical space and operating rules in the airport are described, and the fallback calculation of the graph structure is performed based on the predicted conflict to reconstruct the taxiing path of the aircraft, so as to realize rapid identification of surface conflicts and dynamic avoidance of taxiing paths. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] In order to more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings required for use in the embodiments or the prior art descriptions are briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0046] Figure 1 A method flow chart of a method for dynamic detection and avoidance of taxiing path conflicts of aircraft at an airport surface based on a graph topology network provided by one embodiment of the present invention;

[0047] Figure 2 A schematic diagram of constructing a quad graph network matrix provided by an embodiment of the present invention;

[0048] Figure 3 A schematic diagram of extracting an aircraft taxi path subgraph network matrix provided in accordance with an embodiment of the present invention. DETAILED DESCRIPTION

[0049] In order to further explain the technical means and effects adopted by the present invention to achieve the predetermined invention purpose, the following is a detailed description of the specific implementation method, structure, features and effects of a dynamic detection and avoidance method for taxiing path conflicts of aircraft at an airport based on a graph topology network proposed by the present invention in combination with the accompanying drawings and preferred embodiments. In the following description, different "one embodiment" or "another embodiment" does not necessarily refer to the same embodiment. In addition, specific features, structures or characteristics in one or more embodiments may be combined in any suitable form.

[0050] Unless defined otherwise, all technical and scientific terms used herein have the same meaning as commonly understood by one of ordinary skill in the art to which this invention belongs.

[0051] The following is a detailed description of a specific scheme of a method for dynamic detection and avoidance of taxiing path conflicts of aircraft at an airport based on a graph topology network provided by the present invention in conjunction with the accompanying drawings.

[0052] See also Figure 1 , which shows a method flow chart of a method for dynamic detection and avoidance of taxiing path conflicts of aircraft on the airport surface based on a graph topology network provided by an embodiment of the present invention. The method for dynamic detection and avoidance of taxiing path conflicts of aircraft on the airport surface based on a graph topology network comprises:

[0053] Step S100: constructing a taxiing path graph network matrix based on the airport flight area surface road network, and extracting a taxiing path subgraph network matrix according to aircraft taxiing rules;

[0054] Step S100 specifically includes:

[0055] Step S110: construct the airport surface aircraft taxi path quaternary graph network matrix, including the taxi path node matrix, the taxi path edge matrix, the edge eigenvalue matrix and the access matrix between edges; for details, please refer to Figure 2 As shown in the figure, the airport scene graph structure is constructed according to the physical structure of the airport scene, and then the airport scene quad graph network matrix is ​​constructed; the airport scene taxi path specifically refers to the set of path segments of each aircraft from the starting point (arrival point or take-off point) to the end point (take-off point or departure point) on the airport ground, and the path segment set is composed of multiple nodes and edges; a node refers to a key position or intersection in the airport ground layout, such as a parking space, a runway entrance and exit, a taxiway intersection, etc. A node can be regarded as a stop point or turning point for an aircraft during taxiing; an edge refers to the connection between two nodes, that is, the actual taxi path segment of an aircraft from one node to another; each edge corresponds to a specific taxiway or a moving trajectory in a specific area;

[0056] In this embodiment, the quaternary graph network matrix can be expressed as:

[0057]

[0058] In the formula, represents a quaternion graph network matrix; represents the glide path node matrix; represents the glide path edge matrix; represents the edge eigenvalue matrix; represents the access matrix between edges; where an edge can be represented as , Number the rows and columns of the edge matrix, and the eigenvalues ​​of the edges can be expressed as:

[0059]

[0060] In the formula, Represents edge The length of the space; Represents edge The distribution of the taxiing time; Represents edge The edge direction of

[0061] The access matrix between edges can be expressed as:

[0062]

[0063] In the formula, Represents edge With edge The accessibility relationship between the edges refers to whether the edges are directly continuous; if there is a path that allows the aircraft to Slide directly to the edge , then these two edges have a reachable relationship;

[0064] In this embodiment, by creating a quad graph structure based on the actual terrain scene of the airport flight zone, the aircraft taxi path of the airport scene can be truly restored. By introducing edge features and accessibility relationships, the model can more finely describe the airport scene and command rule requirements, providing a more accurate basis for taxi path optimization and conflict prediction.

[0065] Step S120: extracting the aircraft taxiing path subgraph network matrix from the quad graph network matrix according to the preset aircraft taxiing rules; see Figure 3 As shown in the figure, different aircrafts have different taxi paths on the airport surface due to the influence of parking positions, take-off and landing runways, flight information (airlines, routes, aircraft types), weather, etc. The taxi path constraints of each aircraft on the airport surface are determined according to the airport's operating rules and the specific requirements of airlines, such as the fixed path from the parking position to the runway or from the runway to the parking position, the taxiing speed limit of a specific aircraft type, the use priority and restrictions of the taxiway, etc. For each aircraft , extract all possible sliding paths from the quad graph network matrix, which can be expressed as:

[0066]

[0067] In the formula, Indicates The first A set of optional glide paths; Indicates the total number of path segments;

[0068] Based on the set of optional taxi paths, the relevant nodes, edges, eigenvalues ​​and accessibility relationships are extracted from the quad-graph network matrix to construct a single aircraft taxi path subgraph network matrix, which can be expressed as:

[0069]

[0070] In the formula, Indicates The first A glide path subgraph; Indicates The first A set of nodes in the glide path subgraph; Indicates The first The edge set of the glide path subgraph; Indicates The first The set of edge eigenvalues ​​in the glide path subgraph; Indicates The first The set of accessibility relations in the glide path subgraph;

[0071] Finally, all possible path subgraphs are combined to form a single aircraft taxi path subgraph network matrix, which can be expressed as:

[0072]

[0073] In the formula, Indicates The taxi path subgraph network matrix of each aircraft is constructed; through the taxi path subgraph network of individual aircraft, each aircraft has its own specific graph structure, which can manage the taxi path of the airport surface more finely and enhance the detection and avoidance capabilities of potential conflicts.

[0074] Step S200: extracting a surface taxiing path graph network matrix from a taxiing path subgraph network matrix according to the real-time aircraft motion data;

[0075] Step S200 specifically includes:

[0076] Step S210: Obtain the position and velocity vector of the aircraft according to the real-time motion data of the aircraft; specifically, define the aircraft Movement position , the velocity vector , by connecting to the airport real-time monitoring data, the aircraft monitoring coordinates are converted into three-dimensional motion coordinates, where the definition is the horizontal and vertical axis direction of the plane coordinate system, is the direction of the spatial height coordinate axis; motion position ; Velocity vector , Indicates Aircraft in The velocity component in the direction;

[0077] Step S220: using the Gaussian function to calculate the position matching degree and the direction matching degree of the aircraft taxi path subgraph respectively, and obtain a comprehensive matching degree;

[0078] The calculation of the position matching degree of the aircraft taxi path subgraph includes:

[0079] Calculate the Euclidean distance between the current position of the aircraft and the node in the taxi path subgraph, and its expression is:

[0080]

[0081] In the formula, Indicates Aircraft The running position at the moment The Euclidean distance between the glide path subgraphs; Indicates Aircraft The coordinates of the motion position at the moment; Indicates Aircraft Glide Path Subgraph The spatial position of the mid-node element;

[0082] The Gaussian function is used to evaluate the proximity between the current position of the aircraft and the path node. The expression of position matching is:

[0083]

[0084] In the formula, Indicates Aircraft The moment The position matching degree of the glide path subgraph; express The total number of path segments; Represents the standard deviation parameter, which is used to control the decay speed of the position matching degree. The larger the value, the smaller the influence of distance on the matching degree; the smaller the value, the greater the influence of distance on the matching degree;

[0085] The expression for calculating the direction matching degree of the aircraft taxi path subgraph is:

[0086]

[0087] In the formula, express Direction matching degree; Indicates Aircraft The velocity vector at the moment; Indicates Aircraft The edge direction vector of the sliding path subgraph; express Model; express Model;

[0088]

[0089] In the formula, , , Respectively represent Aircraft The coordinates of the motion position at the moment;

[0090]

[0091] In the formula, Representation Node With Node Construct the direction vector of the edge between them;

[0092] The comprehensive matching degree can be expressed as:

[0093]

[0094] In the formula, express The comprehensive matching degree; represents the first weight coefficient; Represents the second weight coefficient.

[0095] Step S230: extracting the aircraft taxiing path diagram network matrix according to the comprehensive matching degree, which can be expressed as:

[0096]

[0097] In the formula, Indicates that the current moment is most suitable for A taxi chart network for each aircraft;

[0098] Will The taxi path diagram network of all aircraft identified at the time and field is matrixed to construct Network matrix of the spatial taxi path diagram of the aircraft group at each moment :

[0099]

[0100] in, is the total number of aircraft on the scene; step S200 combines the real-time position and speed information of the aircraft, and uses the Gaussian function and cosine similarity to calculate the position matching and direction matching of the taxiing path subgraph, so as to determine the taxiing path that best matches the current aircraft status, which not only improves the accuracy of path selection, but also enhances the understanding and prediction capabilities of the dynamic behavior of the aircraft; through real-time data and matching algorithms, it is ensured that the selected path best matches the actual operation of the aircraft; real-time updating of path information helps to timely discover potential conflict points and take measures to avoid collisions.

[0101] Step S300: constructing a prediction model to predict the spatial taxiing path map network matrix based on the surface taxiing path map network matrix;

[0102] Step S300 specifically includes:

[0103] Step S310: For each edge of the aircraft surface taxi path graph network matrix, the eigenvector of each edge is forward updated and solved using graph neural network. The update rule can be expressed as:

[0104]

[0105] In the formula, Indicates The edge after iterations The eigenvector of represents the activation function; represents the first training weight matrix; represents the second training weight matrix; Represents edge exist The feature vector at the moment; Represents adjacent edges exist The feature vector at the moment; Represents edge The adjacency matrix of

[0106]

[0107] In the formula, Represents edge exist The feature vector at the moment; Represents edge exist The characteristic value at the moment; side exist Accessibility at all times; represents the exponential function, Is the aircraft nearby? superior;

[0108]

[0109] Step S320: according to the predicted position of the aircraft at the future time, remove irrelevant edges from the graph structure edge set at the current time, and obtain the graph structure edge set of the aircraft at the future time;

[0110] Step S330: Calculate the eigenvector and eigenvalue of the aircraft at the future time, which can be expressed as:

[0111]

[0112] In the formula, in the formula, Represents edge exist The feature vector at time, represents the total number of time steps, represents the time step; Represents edge exist The feature vector at the moment; represents the activation function; represents the first training weight matrix; represents the second training weight matrix; Indicates the number of iterations; Represents edge exist The feature vector at the moment; Represents edge The adjacency matrix of

[0113]

[0114] In the formula, Indicates Aircraft in The edge eigenvalue at time;

[0115] Step S340: constructing a network matrix of the space taxiing path diagram of the aircraft at the future time, which can be expressed as:

[0116]

[0117] In the formula, Indicates Aircraft in The spatial glide path graph network at each moment; Indicates Aircraft in The node set at the moment; Indicates Aircraft in The edge combination of moments; Indicates Aircraft in The set of edge eigenvalues ​​at time; Indicates Aircraft in The relationship of moment to moment;

[0118]

[0119] In the formula, express The network matrix of the space taxiing diagram of the aircraft group at each moment; Represents the total number of aircraft on the scene; step S300 dynamically updates and predicts the aircraft's taxi path graph network through the graph neural network, which can effectively capture the movement patterns and potential conflicts of aircraft in the airport scene; specifically, by utilizing the iterative update mechanism of the graph neural network, it is possible to more accurately predict the position and path of aircraft at future moments, thereby identifying potential conflicts in advance; by predicting the path at future moments, the aircraft's taxi path can be adjusted in time to avoid or reduce the occurrence of conflicts, helping airport managers make more scientific and reasonable scheduling decisions.

[0120] Step S400: extracting a conflict moment prediction edge set from the network matrix of the space taxiing path graph;

[0121] Step S400 specifically includes:

[0122] Step S410: extract the number of aircraft on each edge of the network matrix of the spatial taxiing path diagram, which can be expressed as:

[0123]

[0124] In the formula, express Time Edge Total number of aircraft on board; Indicates the number of aircraft on the surface; represents the indicator function, Is it on the edge superior;

[0125] Step S420: Determine the conflict according to the number of aircraft and obtain the predicted edge set at the conflict time, which can be expressed as:

[0126]

[0127] In the formula, express Time Edge Whether there is a conflict;

[0128] The predicted edge set at the final conflict moment can be expressed as: ,in, for The total number of edges in the graph; Step S400 can identify potential conflict points in advance by calculating the number of aircraft on each edge at a future time, thereby providing important information support for subsequent conflict avoidance;

[0129] Step S500: performing fallback solution on the edge set predicted at the conflict moment, and generating a conflict adjustment path with the goal of minimizing the number of conflicts;

[0130] Step S500 specifically includes:

[0131] Step S510: extract the conflicting aircraft objects from the conflict time prediction edge set to obtain the conflicting aircraft object set; specifically, based on the future conflict edge set , extract edge The conflicting aircraft objects are grouped to obtain a set of conflicting aircraft objects. , which can be expressed as:

[0132]

[0133] In the formula, express Time Edge The set of aircraft objects that collide with each other; represents the total number of conflicts;

[0134] Step S520: Perform graph network fallback calculation on each aircraft in the conflicting aircraft object set, and extract the predicted graph network after fallback; specifically, Each aircraft performs graph network backoff calculation and sets the backoff step length , for each conflicting aircraft , calculate the back-off prediction graph network , represents the backoff step length The predicted graph network state after that;

[0135] Step S530: Perform maximum similarity processing on the sliding path subgraph network matrix to extract the conflict adjustment path network with the minimum path change cost; On the corresponding aircraft The glide path subgraph network matrix Perform maximum similarity processing to obtain the conflict adjustment path network with the minimum cost of path change :

[0136]

[0137] In the formula, Indicates conflict adjustment paths for each aircraft; Represents conflict path nodes; Indicates The first A glide path subgraph node; , indicating that the conflict node is different from the node in the conflict adjustment path; Indicates The first A glide path subgraph;

[0138] Step S540: Verify whether the conflict adjustment path network effectively reduces conflicts and update the aircraft taxi path; Substitute into the graph neural prediction network for future The prediction of the moment and the conflict edge set are calculated again ,like The number of sets is smaller than the original , then the replacement path is valid, then the edge set As an aircraft The taxi path of the aircraft is adjusted The above process is repeated until the paths of all conflicting aircraft are adjusted, thereby reducing the occurrence of surface conflicts.

[0139] In summary, by systematically analyzing potential conflicts, back-calculating, finding the optimal path, and verifying the adjustment results, the actual conflict events in the airport scene can be effectively reduced, effective path planning for potential conflicts can be achieved, and the risk of conflict in airport ground transportation can be reduced. It not only improves safety, but also optimizes the scheduling efficiency of aircraft, reduces unnecessary waiting time, and thus improves the smoothness and reliability of the entire airport operation. In addition, this method supports dynamic adjustment and can adapt to changing flight conditions, making airport management more flexible and efficient.

[0140] It should be noted that the sequence of the above embodiments of the present invention is for description only and does not represent the advantages and disadvantages of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0141] The various embodiments in this specification are described in a progressive manner, and the same or similar parts between the various embodiments can be referenced to each other, and each embodiment focuses on the differences from other embodiments.

Claims

1. A method for dynamic detection and avoidance of aircraft taxiing path conflicts at airports based on a graph topology network, characterized in that: The method comprises: The taxi path graph network matrix is ​​constructed based on the airport flight area surface road network, and the taxi path sub-graph network matrix is ​​extracted according to the aircraft taxiing rules; Extracting the surface taxi path diagram network matrix from the taxi path sub-diagram network matrix according to the real-time aircraft movement data; Construct a prediction model to predict the spatial taxi path map network matrix based on the surface taxi path map network matrix; Extract the conflict moment prediction edge set from the network matrix of the space glide path graph; The predicted edge set at the time of conflict is back-solved, and the conflict adjustment path is generated with the goal of minimizing the number of conflicts, including: Extract the conflicting aircraft objects from the conflict moment prediction edge set to obtain the conflicting aircraft object set; Perform graph network fallback calculation on each aircraft in the conflicting aircraft object set, and extract the predicted graph network after fallback; The maximum similarity processing is performed on the sliding path subgraph network matrix to extract the conflict adjustment path network with the minimum cost of path change. The expression is: In the formula, Indicates conflict adjustment paths for each aircraft; Represents conflict path nodes; Indicates The first A glide path subgraph node; , indicating that the conflict node is different from the node in the conflict adjustment path; Indicates The first A glide path subgraph; Verify that the conflict adjustment path network is effective in reducing conflicts and update aircraft taxi paths.

2. The method for dynamic detection and avoidance of taxiing path conflicts of aircraft at airports based on graph topology networks according to claim 1, characterized in that: The taxi path map network matrix is ​​constructed based on the airport flight area surface road network, and the taxi path sub-map network matrix is ​​extracted according to the aircraft taxiing rules, including: Construct the airport surface aircraft taxi path quaternary graph network matrix, including the taxi path node matrix, the taxi path edge matrix, the edge eigenvalue matrix and the accessibility matrix between edges; According to the preset aircraft taxiing rules, the aircraft taxiing path subgraph network matrix is ​​extracted from the quad graph network matrix.

3. The method for dynamic detection and avoidance of taxiing path conflicts of aircraft at airports based on graph topology network according to claim 2, characterized in that: The network matrix of the surface taxi path diagram is extracted from the taxi path sub-diagram network matrix according to the real-time aircraft movement data, including: According to the real-time motion data of the aircraft, the position and velocity vector of the aircraft are obtained; The position matching degree and direction matching degree of the aircraft taxi path subgraph are calculated using Gaussian function to obtain the comprehensive matching degree. The network matrix of the aircraft surface taxi path diagram is extracted based on the comprehensive matching degree.

4. The method for dynamic detection and avoidance of airport surface aircraft taxiing path conflicts based on a graph topology network as claimed in claim 3, characterized in that: The expression for calculating the position matching degree of the aircraft taxi path subgraph using the Gaussian function is: In the formula, Indicates Aircraft The moment The position matching degree of the glide path subgraph; express The total number of path segments; Represents the standard deviation parameter, which is used to control the decay speed of the position matching degree; Indicates Aircraft The running position at the moment The Euclidean distance between the glide path subgraphs.

5. The method for dynamic detection and avoidance of taxiing path conflicts of aircraft at airport surfaces based on graph topology network according to claim 4, characterized in that: The expression for calculating the direction matching degree of the aircraft taxi path subgraph using the Gaussian function is: In the formula, express Direction matching degree; Indicates Aircraft The velocity vector at the moment; Indicates Aircraft The edge direction vector of the sliding path subgraph; express Model; express Model.

6. The method for dynamic detection and avoidance of taxiing path conflicts of aircraft at airports based on graph topology network according to claim 1, characterized in that: Construct a prediction model to predict the spatial taxi path map network matrix based on the surface taxi path map network matrix, including: The eigenvector of each edge in the network matrix of the aircraft surface taxi path diagram is forward updated and solved using graph neural network. According to the predicted position of the aircraft at the future time, the irrelevant edges are removed from the graph structure edge set at the current time to obtain the graph structure edge set of the aircraft at the future time; Calculate the eigenvector and eigenvalue of the aircraft at the future time; Construct a network matrix of the space taxiing path diagram of the aircraft at future times.

7. The method for dynamic detection and avoidance of taxiing path conflicts of aircraft at airports based on graph topology network according to claim 6, characterized in that: The expression for calculating the eigenvector of an aircraft at a future time is: In the formula, Represents edge exist The feature vector at time, represents the total number of time steps, represents the time step; Represents edge exist The feature vector at the moment; represents the activation function; represents the first training weight matrix; represents the second training weight matrix; Indicates the number of iterations; Represents adjacent edges exist The feature vector at the moment; Represents edge The adjacency matrix of .

8. The method for dynamic detection and avoidance of taxiing path conflicts of aircraft at airports based on graph topology networks according to claim 1, characterized in that: The conflict moment prediction edge set extracted from the network matrix of the space glide path graph includes: Extract the number of aircraft on each edge of the network matrix of the spatial taxi path diagram; The conflict is determined based on the number of aircraft, and the predicted edge set at the conflict moment is obtained.

Images