Airport scene aircraft taxiing conflict real-time prediction and dynamic scheduling method and system

Abstract

The invention relates to the technical field of air traffic management, and particularly provides an airport scene aircraft taxiing conflict real-time prediction and dynamic scheduling method, which comprises the following steps: collecting the motion state of an aircraft and airport road network data, and generating the original trajectory of the aircraft by adopting a hidden Markov model; predicting a future taxiing trajectory of the aircraft to obtain a predicted trajectory; screening potential conflict aircrafts by adopting a dynamic R-Tree spatial index technology, judging whether tracks between the potential conflict aircrafts are intersected or not, and generating a dynamic risk value; executing a dynamic scheduling decision based on the dynamic risk value, and generating an aircraft bypass path by adopting a path generation algorithm; rolling time domain control is adopted, and a local path is re-planned based on data obtained in real time. According to the method, prediction of the taxiing trajectory of the aircraft is fully considered, the risk value of the taxiing conflict of the aircraft is pre-judged in advance, and dynamic path adjustment is performed, so that the labor cost is reduced, and the efficiency and safety of airport ground operation are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of air traffic management, and in particular to a method and system for real-time prediction and dynamic scheduling of taxiing conflicts between aircraft at an airport. Background Art

[0002] With the continuous growth of air transport volume, the scale and operational complexity of modern airports have increased significantly. Airport ground taxiing conflicts have become a key bottleneck restricting operational efficiency and flight safety. Although computer simulation-based airport operational efficiency assessment models are currently widely used, existing conflict detection technologies still have limitations.

[0003] Trajectory prediction involves integrating real-time aircraft status data with historical operational patterns to construct future taxi paths, thereby enabling a priori assessment of conflict risk. Existing research on airport conflict detection and warning falls into two main categories: one is based on real-time aircraft position data for conflict assessment. This approach can only identify overt conflicts that have already occurred, lacks foresight into potential conflicts, and provides insufficient advance warning time for conflicts, failing to provide controllers with sufficient decision-making buffer time. The other is based on historical operational data, utilizing machine learning or mathematical statistics to extract common conflict patterns from a historical trajectory library. Historical data-based analysis methods can identify common path conflict patterns through statistical models, but their prediction accuracy is limited in real-time dynamic scenarios and often overlook the actual control operations at the airport.

[0004] Most current research lacks the ability to dynamically predict taxiing conflicts in real time and replan routes. Most systems remain at the conflict alarm level and fail to form a closed-loop linkage with dynamic route replanning technology, making it difficult to meet the needs of large hub airports for real-time conflict resolution and proactive scheduling. Summary of the Invention

[0005] In view of this, the present invention proposes a method and system for real-time prediction and dynamic scheduling of taxiing conflicts between aircraft at airports to solve the problems existing in the above-mentioned prior art.

[0006] On one hand, to achieve the above-mentioned object, the present invention proposes a method for real-time prediction and dynamic scheduling of taxiing conflicts between aircraft on an airport surface. The method is characterized by comprising:

[0007] Acquire airport digital road network data, real-time aircraft signals, air traffic control instructions, and flight plan data, perform data preprocessing, and use hidden Markov models to generate original aircraft trajectories;

[0008] Based on the original trajectory, as well as pre-processed real-time aircraft signals and airport digital road network data, the future taxiing trajectory of the aircraft is predicted to obtain a predicted trajectory;

[0009] Based on the predicted trajectories, a dynamic R-Tree spatial indexing technique is used to screen potential conflicting aircraft, and whether the trajectories of the potential conflicting aircraft intersect is determined to generate a dynamic risk value.

[0010] Executing dynamic scheduling decisions based on the dynamic risk value and generating an aircraft detour path based on a path generation algorithm;

[0011] Adopt rolling horizon control to replan local paths based on real-time acquired data.

[0012] Furthermore, the airport digital road network data includes the latitude and longitude, structure, and length of the taxiways and runways; the air traffic control instructions include waiting instructions and path change instructions; and the flight plan data includes the pre-stored taxi paths for each aircraft.

[0013] Furthermore, the process of generating the original trajectory of the aircraft using the hidden Markov model includes:

[0014] The deviation distance between the trajectory point and the centerline of each taxiway is input into the Gaussian probability density function to calculate the observation probability of the hidden Markov model; a state transition probability matrix is constructed based on the transition frequency between nodes in the historical trajectory, and the original trajectory point is mapped to the taxiway centerline to generate a corrected trajectory; and linear interpolation is performed on the corrected trajectory to generate the original trajectory.

[0015] Furthermore, the process of predicting the future glide path of the aircraft includes:

[0016] When the predicted duration is not greater than the remaining travel time of the current taxiway segment, the trajectory is predicted based on the uniform linear motion model;

[0017] When the predicted duration is greater than the remaining travel time of the current taxiway segment, multiple topologically connected paths are selected for parallel prediction based on the airport network topology according to the following priorities: a. Pre-stored paths that match the flight plan database, b. Paths specified by real-time air traffic control instructions, and c. Feasible paths that meet the aircraft's minimum turning radius. Prediction stops when the target runway threshold or target parking stand entrance is detected.

[0018] Based on the prediction results, the future trajectory of each aircraft is aggregated to generate a predicted trajectory.

[0019] Furthermore, the process of screening potential conflicting aircraft using dynamic R-Tree spatial indexing technology includes:

[0020] Based on the dynamically expanding R-Tree index, the minimum bounding rectangle is used to cover the future trajectory of the aircraft within the prediction time window;

[0021] If the minimum bounding rectangles of the future trajectories of the two aircraft are detected to overlap, it is determined that there is a potential conflict between the two aircraft; otherwise, there is no potential conflict;

[0022] The dynamically expanding R-Tree index is constructed by time slices, and each slice is updated independently to adapt to dynamic movement.

[0023] Furthermore, in the process of determining whether the trajectories of the potential conflicting aircraft intersect, if so, the absolute value of the time difference between the two aircraft reaching the intersection point is calculated; if not, the minimum separation distance between the two aircraft and the absolute value of the time difference between the two aircraft reaching the closest point on their trajectories are calculated;

[0024] The calculation method of the dynamic risk value is as follows:

[0025]

[0026] Where ΔT is the absolute value of the time difference between the two aircraft arriving at the intersection point or the absolute value of the time difference between the two aircraft arriving at the closest point of their trajectories; if ΔT = 0, it is directly regarded as the highest conflict risk and no formula calculation is required; ΔD min is the minimum separation distance between two aircraft; α is the time urgency weight; β is the distance sensitivity weight; γ is the time decay coefficient, which controls the rate of influence of ΔT on risk; ∈ is the smoothing factor parameter; δ is the fixed risk bonus parameter for the intersection scenario; w direction is the direction weight.

[0027] Furthermore, the process of executing a dynamic scheduling decision based on the dynamic risk value includes:

[0028] Based on the dynamic risk value classification, response measures are triggered. In low-risk situations, logs are recorded without intervention. In medium-risk situations, warnings are sent to the control tower for manual monitoring. In high-risk situations, avoidance instructions are automatically generated, including deceleration instructions or the detour path.

[0029] Determine the priority of aircraft avoidance. If two aircraft are traveling in the same direction, the priority is to slow down or adjust the path of the following aircraft. Otherwise, the priority is to slow down or adjust the path of the slower or heavier aircraft. When both aircraft are heavy aircraft, the priority is to adjust the path of the aircraft on the glide path.

[0030] Furthermore, the process of generating an aircraft detour path based on the path generation algorithm includes:

[0031] The calculation is performed based on the comprehensive path length and the real-time conflict risk value based on the cost function, and the cost function is as follows:

[0032] Total cost f(n) = g(n) + h(n) + μ·R(n)

[0033] Where g(n) is the path cost from the current aircraft position to node n; h(n) is the Manhattan distance from node n to the target node; R(n) is the risk cost from the current aircraft position to node n; μ is the risk weight coefficient; the calculation formulas for R(n) and g(n) are as follows:

[0034]

[0035] Among them, Length i is the physical length of the i-th taxiway (meters), Risk i is the dynamic risk value of the taxiway segment, BaseWeight i is the basic weight coefficient of the taxiway segment;

[0036] Based on the calculated total cost, an improved A* algorithm is used to generate a detour path.

[0037] Furthermore, the process of generating a detour path using the improved A* algorithm includes:

[0038] Initialization phase: Load the airport road network topology data and define the path starting point S and target point G; the open list is the set of nodes to be investigated, initially containing only the starting point, and the closed list is the set of nodes that have been investigated, initially empty.

[0039] Path search loop: If the open list is empty, terminate the process and return no solution; otherwise, select the node n with the smallest total cost f(n) from the open list and move it to the closed list; if the current node n is the target point G, backtrack backwards and output the complete path;

[0040] Neighborhood node expansion and filtering: Generate all neighboring nodes of the current node n; filter out neighboring nodes that do not meet the aircraft's minimum turning radius constraint; exclude nodes whose taxiway one-way traffic direction conflicts with the current movement direction;

[0041] Node status update: If the adjacent node is not in the open list or closed list, add it to the open list; if the adjacent node already exists in the open list and the newly calculated f(n) value is better, update its cost value and parent node; then return to the path search loop to continue searching until the target point is found or the open list is exhausted.

[0042] On the other hand, to achieve the above-mentioned purpose, the present invention proposes a real-time prediction and dynamic scheduling system for taxiing conflicts between aircraft at airports, which is characterized by comprising a data processing module, a trajectory prediction engine, a conflict detection engine, a dynamic scheduling decision module, and a visualization and alarm terminal;

[0043] The data processing module is used to receive and perform data preprocessing, and generate the original trajectory of the aircraft using a hidden Markov model;

[0044] The trajectory prediction engine is used to predict the future taxiing trajectory of the aircraft based on the original trajectory and preprocessed data to obtain a predicted trajectory;

[0045] The conflict detection engine is used to screen potential conflicting aircraft using dynamic R-Tree spatial indexing technology, and determine whether the trajectories of the potential conflicting aircraft intersect, thereby generating a dynamic risk value;

[0046] The dynamic scheduling decision module is used to perform dynamic scheduling decisions according to the dynamic risk value and generate an aircraft detour path based on a path generation algorithm;

[0047] The visualization and alarm terminal is used for rendering the airport digital road network heat map, displaying risk levels, and displaying dynamic scheduling decision feedback.

[0048] Compared with the prior art, the present invention has the following beneficial effects:

[0049] The present invention provides a real-time prediction and dynamic scheduling method for taxiing conflicts of aircraft at airports, which includes a multimodal trajectory prediction method. By integrating the airport digital road network topology, air traffic control instructions, flight plan data and aircraft ADS-B real-time signal data, a hidden Markov model (HMM) is used to correct trajectory probabilities, and multi-path parallel prediction is achieved in combination with dynamic taxiing status.

[0050] The real-time detection and warning method for airport taxiing conflicts proposed in this invention includes spatiotemporal conflict detection and risk assessment. Based on the dynamic R-Tree spatial index, aircraft pairs with potential conflicts are screened, and a conflict risk assessment model integrating spatiotemporal factors (including time difference, minimum separation distance and direction weight) is constructed.

[0051] The present invention triggers response measures in a hierarchical manner based on the dynamic risk values of potential conflicting aircraft pairs. For situations requiring path adjustment, an improved A* algorithm and rolling horizon control (RHC) are used to generate a detour path, and the local path is periodically replanned based on the latest data.

[0052] The present invention fully considers the prediction of aircraft taxiing trajectories, predicts the risk value of aircraft taxiing conflicts in advance and performs dynamic path adjustments, which helps to reduce labor costs and improve the efficiency and safety of airport ground operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] Various other advantages and benefits will become apparent to those skilled in the art by reading the detailed description of the preferred embodiment below. The accompanying drawings are only for the purpose of illustrating the preferred embodiment and are not to be considered as limiting the present invention. In the accompanying drawings:

[0054] Figure 1 This is a flow chart of the method for real-time prediction and dynamic scheduling of taxiing conflicts between aircraft at an airport according to the present invention;

[0055] Figure 2 Schematic diagram of aircraft trajectory before and after trajectory processing in an embodiment of the present invention;

[0056] Figure 3 This is a flowchart of conflict screening for an R-Tree index in an embodiment of the present invention;

[0057] Figure 4 Schematic diagram of the R-Tree hierarchical structure in an embodiment of the present invention;

[0058] Figure 5 Schematic diagram of R-Tree time slice MBR overlap in an embodiment of the present invention;

[0059] Figure 6 This is a schematic diagram of a specific implementation process of the improved A* algorithm in an embodiment of the present invention;

[0060] Figure 7 A schematic diagram of the framework of the system provided by an embodiment of the present invention;

[0061] Figure 8 This is a schematic diagram of the hardware deployment architecture of the system provided by an embodiment of the present invention. DETAILED DESCRIPTION

[0062] The exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although exemplary embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, these embodiments are provided to enable a more thorough understanding of the present disclosure and to fully convey the scope of the present disclosure to those skilled in the art. It should be noted that, unless there is a conflict, the embodiments of the present disclosure and the features described in the embodiments can be combined with each other. The present invention will be described in detail below with reference to the accompanying drawings and in combination with the embodiments.

[0063] This embodiment proposes a real-time prediction and dynamic scheduling method for taxiing conflicts between aircraft at an airport. Figure 1 As shown, the specific steps include:

[0064] Step 1: Data reception and trajectory processing: The system acquires airport digital network data, air traffic control instructions, flight plan databases, and real-time aircraft ADS-B signals. Outliers are removed based on preset acceleration thresholds (set to ±3 m / s²) and heading mutation thresholds (set to ±30° / s). The original trajectory points are then mapped to the taxiway centerline using a hidden Markov model (HMM). The HMM state transition probability matrix is obtained by counting the transition frequencies between nodes in the historical trajectory.

[0065] Step 2: Real-time taxi trajectory prediction: Based on the processed aircraft motion status data and the airport digital road network topology, the aircraft's taxi trajectory within the future time window is predicted for two situations: the predicted duration is less than the remaining travel time of the current taxiway segment, and the predicted duration is greater than or equal to the remaining travel time.

[0066] Step 3: Spatiotemporal conflict detection and risk assessment. Based on the predicted aircraft trajectories, dynamic R-Tree spatial indexing technology is used to screen aircraft pairs with potential conflicts. The future trajectory intersection point or closest point of these aircraft pairs is calculated, and a dynamic risk value is generated by combining the time difference, minimum separation distance, and movement direction weight.

[0067] Step 4: Dynamic scheduling decision-making, based on the dynamic risk value of the aircraft pairs with potential conflicts, triggers response measures according to the preset three-level risk value threshold.

[0068] The data reception and trajectory processing specifically include:

[0069] Step 1-1: Obtain airport digital road network data, including the latitude and longitude, structure, and length of taxiways and runways; air traffic control instructions, including holding instructions and route changes (taxiway switching, runway reallocation, and parking position adjustment instructions); flight plan database, including pre-stored taxi paths for each aircraft;

[0070] Step 1-2: Receive real-time ADS-B signals from aircraft, including position, velocity, and heading angle, and remove abnormal points based on preset acceleration thresholds and heading mutation thresholds;

[0071] Step 1-3: Use the taxiway node set S = {s1, s2, s3, ..., s n The original trajectory points are mapped to the taxiway centerline using a hidden Markov model (HMM) in the state space. The observation probability of the hidden Markov model is calculated by inputting the Gaussian probability density function with the deviation distance between the trajectory point and the centerline of each taxiway. Where d is the deviation distance, the mean μ is set to 0 (ideal center line), σ reflects the tolerance of the actual deviation and can be set to 2m; the state transition probability matrix A = [a ij] is obtained by counting the transfer frequencies between nodes in the historical trajectory:

[0072]

[0073] where N i→j Represents the slave node s in the historical data i Transfer to node s j the number of times;

[0074] Step 1-4: Perform linear interpolation on the corrected trajectory data (eg, generate an interpolation point every 0.5 seconds) to improve the accuracy of the trajectory data.

[0075] Figure 2 Schematic diagrams of aircraft trajectories before and after trajectory processing are shown.

[0076] The real-time prediction of the glide trajectory specifically includes:

[0077] Step 2-1: Trajectory prediction based on dynamic gliding state, setting the prediction time to 40s:

[0078] ① When the predicted duration t ≤ the remaining travel time of the current taxiway segment, the trajectory is predicted based on the uniform linear motion model. Assuming that the aircraft maintains uniform linear motion during the taxiing phase, the formula for calculating its future position is:

[0079]

[0080] Where (x0, y0) is the current aircraft coordinate, v is the minimum value of the aircraft's real-time speed and the taxiway speed limit, θ is the heading angle (relative to true north), and t is the future time (seconds);

[0081] ② When the predicted duration t exceeds the remaining travel time of the current taxiway segment, multiple topologically connected paths are selected for parallel prediction based on the airport road network topology according to the following priorities: a. a pre-stored path that matches the flight plan database, b. a path specified by air traffic control instructions received in real time, and c. a feasible path that meets the aircraft's minimum turning radius.

[0082] ③ When the target runway threshold or target parking stand entrance is detected, stop distance prediction;

[0083] Step 2-2: Construct a trajectory time slice set, which includes the trajectory of each aircraft within the next time t (t=40s).

[0084] The spatiotemporal conflict detection and risk assessment specifically includes the following steps:

[0085] Step 3-1: Potential conflict screening: Use dynamic R-Tree spatial indexing technology to quickly screen aircraft pairs with potential conflicts and process trajectory data in batches by time slices;

[0086] Step 3-2: Conflict Detection: Traverse the screened aircraft pairs with potential conflicts and determine whether the predicted trajectories of the two aircraft in each pair at time t in the future intersect. If the trajectories intersect, calculate the absolute value of the time difference between the two aircraft reaching the intersection point. If the trajectories do not intersect, calculate the minimum separation distance between the two aircraft and the absolute value of the time difference between the two aircraft reaching the closest point on the trajectory.

[0087] Step 3-3: Conflict risk assessment: Based on the future trajectory intersection point or closest point of the aircraft pair, combined with the time difference, minimum separation distance and movement direction weight, a dynamic risk value is generated to quantify the risk level.

[0088] Figure 3 The R-Tree conflict screening flow chart is shown. The potential conflict screening is specifically as follows:

[0089] First, input the trajectory time slice set based on the aircraft prediction trajectory obtained in step 2, and then generate the minimum bounding rectangle (MBR) by slicing. The minimum bounding rectangle (MBR) of the R-Tree index covers the future trajectory within the aircraft prediction time window, and constructs a hierarchical R-Tree ( Figure 4 The figure shows the schematic diagram of the R-Tree hierarchical structure. Then, overlap detection is performed. If the R-Tree index detects that the minimum bounding rectangles (MBRs) of the future trajectories of two aircraft overlap, then the aircraft pair has a potential conflict. Otherwise, there is no potential conflict. Figure 5 A schematic diagram showing the overlap of the time-sliced MBRs of the R-Tree is shown. The MBRs formed by the predicted trajectories of two aircraft have an intersecting area, indicating a potential conflict between the two aircraft. Finally, the aircraft pairs with potential conflicts are output.

[0090] The spatial index technology is a dynamically extended R-Tree index. The R-Tree index is updated every 2 seconds and is triggered to be forced updated when the aircraft speed changes by more than ±10% or the heading changes by more than ±5°.

[0091] The conflict risk assessment is specifically as follows:

[0092] The dynamic risk value calculation formula is:

[0093]

[0094] Where ΔT is the absolute value of the time difference between the two aircraft reaching the intersection point (trajectory intersection) or the absolute value of the time difference between the two aircraft reaching the closest point of the trajectory (trajectory non-intersection); if ΔT = 0, it is directly regarded as the highest risk and no formula calculation is required; ΔD minis the minimum separation distance between two aircraft; α is the time urgency weight, which is 1.0 in the embodiment; β is the distance sensitivity weight, which is 1.5 in the embodiment; γ is the time attenuation coefficient, which is 0.1 in the embodiment, controlling the rate of influence of ΔT on risk; ∈ is the smoothing factor parameter, preventing ΔD min →The value overflows when 0, and the value in the embodiment is 0.5; δ is the fixed risk bonus parameter of the intersection scene, and the value in the embodiment is 0.3; w direction is the direction weight, which is set in sections according to the heading angle θ: in the embodiment, it is set to 0.2 when in the same direction (θ<30°), 1.0 when crossing (30°≤θ<150°), and 1.5 when facing (θ≥150°).

[0095] For example, a potential collision is detected between aircraft A and aircraft B, where aircraft A's speed is 15 m / s and its heading is 30°, and aircraft B's speed is 10 m / s and its heading is 210°. Their trajectories are predicted to intersect at point X on the taxiway. Calculate the intermediate quantity:

[0096] ΔT = |(time from A to point X) - (time from B to point X)| = |80s - 120s| = 40s

[0097] ΔD min = 25m (the shortest distance between the two aircraft tracks)

[0098] θ=|210°-30°|=180°→w direction =1.5(opposite direction)

[0099] The final calculated current risk value is 0.67.

[0100] The dynamic scheduling decision specifically includes:

[0101] Step 4-1: Based on the dynamic risk values of potentially conflicting aircraft pairs, response measures are triggered in a graded manner. In low-risk situations (Risk < 0.3), a log is recorded without intervention. In medium-risk situations (0.3 ≤ Risk < 0.7), an alert is sent to the tower for manual monitoring. In high-risk situations (Risk ≥ 0.7), avoidance instructions are automatically generated, including deceleration instructions or generating a detour based on a path generation algorithm. Table 1 shows a comparison table for risk level determination in this embodiment:

[0102] Table 1

[0103] Risk value interval Risk Level Response measures [0,0.3) Low risk Log, don't interfere [0.3,0.7) Medium risk Send yellow warning to the tower ≥0.7 High risk Automatically generate deceleration / path change instructions

[0104] According to Table 1, the risk value of aircraft A and aircraft B is 0.67, and a yellow warning prompt needs to be sent to the tower.

[0105] Step 4-2: Determine the priority of aircraft avoidance. If two aircraft are traveling in the same direction, decelerate or adjust the path of the trailing aircraft first. Otherwise, decelerate or adjust the path of the slower or heavier aircraft first. If both aircraft are heavy, prioritize adjusting the glide path aircraft.

[0106] The path generation algorithm adopts the improved A* algorithm. Figure 6 The specific implementation process of improving the A* algorithm is demonstrated, including:

[0107] Initialization phase: Load the airport road network topology data and define the path starting point S and target point G; the open list is the set of nodes to be investigated, initially containing only the starting point, and the closed list is the set of nodes that have been investigated, initially empty.

[0108] Path search loop: If the open list is empty, terminate the process and return no solution; otherwise, select the node n with the smallest comprehensive cost function value f(n) from the open list and move it to the closed list; if the current node n is the target point G, reverse backtracking to output the complete path;

[0109] Neighborhood node expansion and filtering: Generate all neighboring nodes of the current node n; filter out neighboring nodes that do not meet the aircraft's minimum turning radius constraint; exclude nodes whose taxiway one-way traffic direction conflicts with the current movement direction;

[0110] The cost function integrates the path length and the real-time conflict risk value. The comprehensive cost function formula is:

[0111] Total cost f(n) = g(n) + h(n) + μ·R(n)

[0112] Where h(n) is the Manhattan distance from node n to the target node; R(n) is the risk cost from the current aircraft position to node n; g(n) is the path cost from the current aircraft position to node n; μ is the risk weight coefficient, which is dynamically adjusted according to the airport operation status. In this embodiment, it is set to 1.2 during peak hours (when the airport take-off and landing frequency is greater than 40 flights / hour), 0.8 during off-peak hours, and 2.0 in emergency mode (for example, low visibility (<800m) or runway emergencies). The calculation formulas for R(n) and g(n) are as follows:

[0113]

[0114] Length i is the physical length of the i-th taxiway (meters), Risk i is the real-time risk value of the taxiway segment, provided in real time by the conflict risk assessment method, BaseWeight iis the basic weight coefficient of the taxiway segment, which is set to 1.0 for apron, 1.2 for taxiway, and 1.5 for runway according to the regional importance level in the embodiment.

[0115] Node status update: If the adjacent node is not in the open list or closed list, add it to the open list; if the adjacent node already exists in the open list and the newly calculated f(n) value is better, update its cost value and parent node; then return to the path search loop to continue searching until the target point is found or the open list is exhausted.

[0116] The path generation algorithm integrates Rolling Horizon Control (RHC) to periodically replan local paths based on the latest data, with the replanning period of the RHC being adjustable to 5-30 seconds.

[0117] In this example, during peak hours, the planned path of aircraft C conflicts with that of aircraft D (Risk = 0.8). Aircraft C is currently located on taxiway S1 and its target parking position is G. The candidate paths and their costs for aircraft C are:

[0118] Candidate node 1: S1→S2 (length 200m, Risk=0.8),

[0119] The cost is f(n) = 200 × 1.2 + 350 + 0.8 × 200 = 750

[0120] Candidate node 2: S1→S3 (length 240m, Risk=0),

[0121] The cost is f(n) = 240 × 1.2 + 390 = 678

[0122] Although S2 is shorter, there is a risk of conflict, so aircraft C finally chooses to go from node S1 to node S3, avoiding the intersection during peak hours.

[0123] like Figure 7 , Figure 8 As shown, this embodiment provides a real-time prediction and dynamic scheduling system for taxiing conflicts between aircraft at an airport, which is used to implement the above-mentioned real-time prediction and dynamic scheduling method for taxiing conflicts between aircraft at an airport, and includes the following modules:

[0124] Module 1: Data processing module, which loads airport taxiway configuration data, receives and parses multi-source data streams in real time, including aircraft ADS-B data (aircraft position, speed, heading angle), extracts motion status, performs trajectory map matching, and air traffic control instruction data (holding area allocation, taxiway closure status), and outputs a standardized motion status data set;

[0125] Module 2: Trajectory prediction engine, which performs real-time trajectory prediction based on motion state data and generates the sliding trajectory within the next time t;

[0126] Module 3: Conflict Detection Engine, including a dynamic spatial indexing unit, which filters out aircraft pairs with overlapping trajectory bounding boxes based on a dynamic R-Tree index. The dynamic R-Tree index is implemented through GPU parallel computing, and a single full update takes less than 50ms; a risk calculation unit, which is configured to calculate the time difference (ΔT), the minimum separation distance (ΔD) and the minimum distance between the selected aircraft pairs. min ) and dynamic risk value (Risk);

[0127] Module 4: Dynamic dispatch decision module, including a hierarchical response unit, which triggers three levels of response (log record / tower warning / automatic avoidance command) according to the risk level and risk value; a command issuing unit, which issues commands to pilots or tower terminals via the data link. Avoidance commands require secondary confirmation by the controller before they can be sent to the aircraft. The confirmation timeout is a configurable parameter set to 5 seconds. After the timeout, a voice prompt is issued for manual handling.

[0128] Module 5: Visualization and alarm terminal, which displays conflict heat maps, risk levels and dispatch instructions in real time based on the aggregated risk values of taxiway nodes.

[0129] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, ordinary technicians in the field should understand that the specific implementation methods of the present invention can still be modified or replaced by equivalents. Any modification or equivalent replacement that does not depart from the spirit and scope of the present invention should be covered by the scope of protection of the claims of the present invention.

Claims

1. A method for real-time prediction and dynamic scheduling of aircraft taxiing conflicts at an airport, characterized in that: include: Acquire airport digital road network data, real-time aircraft signals, air traffic control instructions, and flight plan data, perform data preprocessing, and use hidden Markov models to generate original aircraft trajectories; Based on the original trajectory, as well as pre-processed real-time aircraft signals and airport digital road network data, the future taxiing trajectory of the aircraft is predicted to obtain a predicted trajectory; Based on the predicted trajectories, a dynamic R-Tree spatial indexing technique is used to screen potential conflicting aircraft, and whether the trajectories of the potential conflicting aircraft intersect is determined to generate a dynamic risk value. Executing dynamic scheduling decisions based on the dynamic risk value and generating an aircraft detour path based on a path generation algorithm; Adopt rolling horizon control to replan local paths based on real-time acquired data.

2. The method for real-time prediction and dynamic scheduling of airport aircraft taxiing conflicts according to claim 1, characterized in that: The airport digital road network data includes the latitude and longitude, structure, and length of the taxiways and runways; the air traffic control instructions include waiting instructions and path change instructions; and the flight plan data includes the pre-stored taxi paths for each aircraft.

3. The method for real-time prediction and dynamic scheduling of airport aircraft taxiing conflicts according to claim 1, characterized in that: The process of generating the original trajectory of an aircraft using a hidden Markov model includes: The deviation distance between the trajectory point and the centerline of each taxiway is input into the Gaussian probability density function to calculate the observation probability of the hidden Markov model; a state transition probability matrix is constructed based on the transition frequency between nodes in the historical trajectory, and the original trajectory point is mapped to the taxiway centerline to generate a corrected trajectory; and linear interpolation is performed on the corrected trajectory to generate the original trajectory.

4. The method for real-time prediction and dynamic scheduling of airport aircraft taxiing conflicts according to claim 1, characterized in that: The process of predicting the future glide path of an aircraft includes: When the predicted duration is not greater than the remaining travel time of the current taxiway segment, the trajectory is predicted based on the uniform linear motion model; When the predicted duration is greater than the remaining travel time of the current taxiway segment, multiple topologically connected paths are selected for parallel prediction based on the airport network topology according to the following priorities: a. Pre-stored paths that match the flight plan database, b. Paths specified by real-time air traffic control instructions, and c. Feasible paths that meet the aircraft's minimum turning radius. Prediction stops when the target runway threshold or target parking stand entrance is detected. Based on the prediction results, the future trajectory of each aircraft is aggregated to generate a predicted trajectory.

5. The method for real-time prediction and dynamic scheduling of airport aircraft taxiing conflicts according to claim 1, characterized in that: The process of screening potential conflicting aircraft using dynamic R-Tree spatial indexing technology includes: Based on the dynamically expanding R-Tree index, the minimum bounding rectangle is used to cover the future trajectory of the aircraft within the prediction time window; If the minimum bounding rectangles of the future trajectories of the two aircraft are detected to overlap, it is determined that there is a potential conflict between the two aircraft; otherwise, there is no potential conflict; The dynamically expanding R-Tree index is constructed by time slices, and each slice is updated independently to adapt to dynamic movement.

6. The method for real-time prediction and dynamic scheduling of airport aircraft taxiing conflicts according to claim 1, characterized in that: In the process of determining whether the trajectories of the potential conflicting aircraft intersect, if so, calculating the absolute value of the time difference between the two aircraft reaching the intersection point; if not, calculating the minimum separation distance between the two aircraft and the absolute value of the time difference between the two aircraft reaching the closest point on their trajectories; The calculation method of the dynamic risk value is as follows: Where ΔT is the absolute value of the time difference between the two aircraft arriving at the intersection point or the absolute value of the time difference between the two aircraft arriving at the closest point of their trajectories; if ΔT = 0, it is directly regarded as the highest conflict risk and no formula calculation is required; ΔD min is the minimum separation distance between two aircraft; α is the time urgency weight; β is the distance sensitivity weight; γ is the time decay coefficient, which controls the rate of influence of ΔT on risk; ∈ is the smoothing factor parameter; δ is the fixed risk bonus parameter for the intersection scenario; w direction is the direction weight.

7. The method for real-time prediction and dynamic scheduling of airport aircraft taxiing conflicts according to claim 1, characterized in that: The process of performing dynamic scheduling decision based on the dynamic risk value includes: Based on the dynamic risk value classification, response measures are triggered. In low-risk situations, logs are recorded without intervention. In medium-risk situations, warnings are sent to the control tower for manual monitoring. In high-risk situations, avoidance instructions are automatically generated, including deceleration instructions or the detour path. Determine the priority of aircraft avoidance. If two aircraft are traveling in the same direction, the priority is to slow down or adjust the path of the following aircraft. Otherwise, the priority is to slow down or adjust the path of the slower or heavier aircraft. When both aircraft are heavy aircraft, the priority is to adjust the path of the aircraft on the glide path.

8. The method for real-time prediction and dynamic scheduling of airport aircraft taxiing conflicts according to claim 1, characterized in that: The process of generating an aircraft detour path based on the path generation algorithm includes: The calculation is performed based on the comprehensive path length and the real-time conflict risk value based on the cost function, and the cost function is as follows: Total cost f(n) = g(n) + h(n) + μ·R(n) Where g(n) is the path cost from the current aircraft position to node n; h(n) is the Manhattan distance from node n to the target node; R(n) is the risk cost from the current aircraft position to node n; μ is the risk weight coefficient; the calculation formulas for R(n) and g(n) are as follows: Among them, Length i is the physical length of the i-th taxiway (meters), Risk i is the dynamic risk value of the taxiway segment, BaseWeight i is the basic weight coefficient of the taxiway segment; Based on the calculated total cost, an improved A* algorithm is used to generate a detour path.

9. The method for real-time prediction and dynamic scheduling of airport aircraft taxiing conflicts according to claim 8, characterized in that: The process of generating a detour path using the improved A* algorithm includes: Initialization phase: Load the airport road network topology data and define the path starting point S and target point G; the open list is the set of nodes to be investigated, initially containing only the starting point, and the closed list is the set of nodes that have been investigated, initially empty. Path search loop: If the open list is empty, terminate the process and return no solution; otherwise, select the node n with the smallest total cost f(n) from the open list and move it to the closed list; if the current node n is the target point G, backtrack backwards and output the complete path; Neighborhood node expansion and filtering: Generate all neighboring nodes of the current node n; filter out neighboring nodes that do not meet the aircraft's minimum turning radius constraint; exclude nodes whose taxiway one-way traffic direction conflicts with the current movement direction; Node status update: If the adjacent node is not in the open list or closed list, add it to the open list; if the adjacent node already exists in the open list and the newly calculated f(n) value is better, update its cost value and parent node; then return to the path search loop to continue searching until the target point is found or the open list is exhausted.

10. A real-time prediction and dynamic scheduling system for taxiing conflicts between aircraft at an airport based on the method according to any one of claims 1 to 9, characterized in that: It includes data processing module, trajectory prediction engine, conflict detection engine, dynamic scheduling decision module, visualization and alarm terminal; The data processing module is used to receive and perform data preprocessing, and generate the original trajectory of the aircraft using a hidden Markov model; The trajectory prediction engine is used to predict the future taxiing trajectory of the aircraft based on the original trajectory and preprocessed data to obtain a predicted trajectory; The conflict detection engine is used to screen potential conflicting aircraft using dynamic R-Tree spatial indexing technology, and determine whether the trajectories of the potential conflicting aircraft intersect, thereby generating a dynamic risk value; The dynamic scheduling decision module is used to perform dynamic scheduling decisions according to the dynamic risk value and generate an aircraft detour path based on a path generation algorithm; The visualization and alarm terminal is used for rendering the airport digital road network heat map, displaying risk levels, and displaying dynamic scheduling decision feedback.

Images