Method and system for predicting and alerting landing distance of an aircraft based on approach flight data
By using multi-time-window prediction and convolutional neural network analysis of aircraft approach data, the instability problem of aircraft landing distance assessment in existing technologies has been solved, enabling more accurate and continuous landing distance prediction and warning.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- CIVIL AVIATION UNIV OF CHINA
- Filing Date
- 2026-06-26
- Publication Date
- 2026-07-31
AI Technical Summary
Existing aircraft landing distance prediction methods rely too heavily on a single time window, making them susceptible to instantaneous data fluctuations. They lack time window stability assessment and runway threshold time prediction distance inference mechanisms, leading to unstable assessment results and false alarms or omissions.
A convolutional neural network-based method is used to predict flight data during the aircraft's approach phase through multiple time windows, calculate the predicted landing distance, and infer the predicted landing distance at the runway threshold by judging the stability and trend of adjacent time windows. Constraint judgment is then made in conjunction with the available runway length.
It improves the continuity and stability of landing distance assessment, reduces false alarms and missed alarms, and provides a more reliable reference for operational decision-making.
Smart Images
Figure CN122493697A_ABST
Abstract
Description
Technical Field
[0001] This invention belongs to the field of aviation flight data processing and flight safety auxiliary decision technology, and particularly relates to a method and system for predicting and warning aircraft landing distance based on approach flight data. Background Technology
[0002] The approach and landing phases are the highest-risk phases of flight operations, as aircraft must complete touchdown, deceleration, and stopping within a limited runway length. Changes in approach speed, wind field, attitude, or braking conditions can increase the actual landing distance, thus affecting runway margin. Therefore, assessing the landing distance in advance before the aircraft reaches the runway threshold and providing corresponding operational reference information based on the available runway length is crucial for improving safety during the approach and landing phases.
[0003] With the widespread application of flight data recording devices such as Quick Access Recorders (QARs), data on aircraft speed, altitude, attitude, wind field, and engine status during the approach phase can be continuously acquired, providing a data foundation for landing distance assessment. Existing methods typically rely on flight status data at a fixed moment, employing empirical models, dynamic models, or neural network models to obtain a single landing distance prediction, and then make a judgment based on the relationship between this result and runway length. However, most existing solutions depend on a single time window or a single prediction result, making them susceptible to instantaneous changes in flight status, wind field disturbances, data noise, or control changes. When a short-term deviation occurs in the prediction result at a certain moment, the system may directly output an inaccurate risk assessment, leading to false alarms or missed alarms. Furthermore, existing solutions typically lack assessment of the stability of continuous prediction results and lack a processing mechanism to infer runway threshold status based on multiple prediction results at critical decision points, making it difficult to fully reflect the continuous changing characteristics of the aircraft during the approach process.
[0004] Therefore, it is necessary to propose a landing distance calculation and alarm method based on multi-time window prediction. This method continuously generates multiple predicted landing distance results during the approach process and judges the prediction stability by observing the changes in the prediction results of adjacent time windows. If the prediction results are stable, the predicted landing distance at the runway threshold time is further inferred by combining multiple prediction results before the decision height position. This is then constrained by the available runway length, thereby improving the continuity, stability, and operational decision-making reference value of the landing distance assessment results. Summary of the Invention
[0005] This invention aims to at least solve one of the technical problems existing in related technologies. To this end, this invention provides a method and system for predicting and warning aircraft landing distance based on approach flight data, which solves the problems of existing landing distance assessment methods that rely too much on the prediction results of a single time window, are easily affected by instantaneous data fluctuations, lack time window stability judgment, and lack a mechanism for inferring the predicted distance at the runway threshold.
[0006] This invention provides a method for predicting and alerting aircraft landing distance based on approach flight data, comprising: S1: Acquire the target aircraft's approach flight data, preprocess the approach flight data to obtain approach phase time series samples, input the approach phase time series samples into a convolutional neural network model to obtain the approach speed prediction value, and calculate the predicted landing distance under the current time window based on the approach speed prediction value. S2: Determine the stability and overall trend of the predicted landing distance under two adjacent time windows; S3: Once the predicted landing distances of two adjacent time windows are determined to be stable, the predicted landing distances for subsequent time windows are obtained until the aircraft reaches the position corresponding to the decision height. After reaching the decision height, based on the overall trend of the predicted landing distances of the previous multiple time windows, a fitting function between the predicted time and the predicted landing distance is selected, and the average predicted landing distance of the time window is calculated. Based on the fitting function and the average predicted landing distance of the time window, the predicted landing distance of the aircraft at the time of arrival at the runway threshold is inferred. S4: Calculate the predicted landing distance with safety margin based on the predicted landing distance when the aircraft arrives at the runway threshold. Perform constraint judgment based on the predicted landing distance with safety margin and the available runway length. When the predicted landing distance with safety margin is less than or equal to the available runway length, output the predicted landing distance with safety margin; when the predicted landing distance with safety margin is greater than the available runway length, output a risk alarm signal.
[0007] Furthermore, the convolutional neural network model includes an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, and a fully connected regression output layer. The first convolutional layer is used to extract short-term local features, the second convolutional layer is used to extract high-level abstract features, the first pooling layer and the second pooling layer are used to reduce feature dimensionality and suppress local noise, and the fully connected regression output layer is used to output the approach velocity prediction value.
[0008] Furthermore, calculating the predicted landing distance for the current time window based on the approach speed prediction includes: S11: Calculate the inlet vacuum velocity based on the predicted approach velocity; S12: Calculate the grounding speed and braking start speed based on the inlet vacuum speed; S13: Calculate the ground speed at which braking begins based on the braking initiation speed and wind speed; S14: Calculate the distance of the air section based on the grounding speed and the inlet vacuum speed; calculate the distance of the ground transition section based on the grounding speed and the braking start speed; calculate the distance of the ground deceleration section based on the ground speed at the start of braking. S15: Calculate and predict the landing distance based on the distance in the air, the distance in the ground transition section, and the distance in the ground deceleration section.
[0009] Furthermore, stability assessment based on the predicted landing distances in two adjacent time windows includes: Using the runway entrance as the time reference, The second is the time window for the aircraft's distance from the runway threshold. The time window is divided into minutes and decision-high approach flights; Calculate the predicted landing distance for the same flight in different time windows and arrange them in chronological order. Calculate the relative rate of change of the predicted landing distance between two adjacent time windows. Determine the stability based on the relative rate of change of the predicted landing distance between two adjacent time windows.
[0010] Furthermore, the stability criterion is: in, To predict stability between two adjacent time windows, for Time window and The time window predicts the relative rate of change of landing distance; when If the prediction result of the current window is small in difference from the prediction result of the previous window, the prediction result has good continuity, and the prediction result of the current window is determined to be stable. when If the forecast results of adjacent windows change significantly, the current forecast result may be affected by changes in flight status, wind field disturbances, or data fluctuations, and the forecast result of the current window is determined to be unstable.
[0011] Furthermore, judging the overall trend based on the predicted landing distance changes between two adjacent time windows includes: The trend of predicted landing distance changes between two adjacent time windows is determined by the change values of predicted landing distance between two adjacent time windows; The percentage of landing distances showing an upward trend between adjacent time windows is calculated based on the predicted landing distance change trends between adjacent time windows. The overall trend of change between two adjacent time windows can be judged by the percentage of those two windows showing an upward trend.
[0012] Furthermore, the predicted landing distance at the moment the aircraft arrives at the runway threshold is inferred as: in, The predicted landing distance is calculated from a distance of 0 seconds from the runway threshold. The average predicted landing distance is calculated over the time window. This represents the overall trend between two adjacent time windows. This is the fitting cutoff constant.
[0013] The present invention also provides an aircraft landing distance prediction and warning system based on approach flight data, for executing the above-mentioned aircraft landing distance prediction and warning method based on approach flight data, comprising: The predicted landing distance calculation module acquires the target aircraft's approach flight data, preprocesses the approach flight data to obtain approach phase time series samples, inputs the approach phase time series samples into a convolutional neural network model to obtain approach speed prediction values, and calculates the predicted landing distance under the current time window based on the approach speed prediction values. The judgment module performs stability determination and overall trend determination on the predicted landing distance under two adjacent time windows; The inference module, after determining that the predicted landing distance of two adjacent time windows is stable, obtains the predicted landing distance of the subsequent time windows until the aircraft reaches the position corresponding to the decision height. After reaching the decision height, based on the overall trend of the predicted landing distance of the previous multiple time windows, selects the fitting function between the predicted time and the predicted landing distance, calculates the average predicted landing distance of the time window, and infers the predicted landing distance of the aircraft at the time of arrival at the runway threshold based on the fitting function and the average predicted landing distance of the time window. The output module calculates a predicted landing distance with a safety margin based on the predicted landing distance at the time the aircraft arrives at the runway threshold. It then performs constraint judgment based on the predicted landing distance with a safety margin and the available runway length. When the predicted landing distance with a safety margin is less than or equal to the available runway length, it outputs the predicted landing distance with a safety margin; when the predicted landing distance with a safety margin is greater than the available runway length, it outputs a risk alarm signal.
[0014] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the program to implement the steps of the above-described method for predicting and warning aircraft landing distance based on approach flight data.
[0015] The present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements the steps of the above-described method for predicting and warning aircraft landing distance based on approach flight data.
[0016] The above-described one or more technical solutions in the embodiments of the present invention have at least one of the following technical effects: This invention generates multiple predicted landing distances continuously during the approach process and assesses prediction stability by observing changes in predictions across adjacent time windows. When predictions are stable, it further combines multiple predictions before the decision height to infer the predicted landing distance at the runway threshold, then compares this with the available runway length for constraint assessment. This improves the continuity, stability, and operational decision-making reference value of the landing distance assessment results. This invention addresses the problems of existing landing distance assessment methods, such as over-reliance on single-time-window predictions, susceptibility to instantaneous data fluctuations, lack of time-window stability assessment, and lack of a mechanism for inferring predicted distances at the runway threshold.
[0017] Additional aspects and advantages of the invention will be set forth in part in the description which follows, and in part will be obvious from the description, or may be learned by practice of the invention. Attached Figure Description
[0018] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.
[0019] Figure 1 This is a flowchart illustrating an aircraft landing distance prediction and warning method based on approach flight data provided by the present invention.
[0020] Figure 2 This is a schematic diagram of the structure of an aircraft landing distance prediction and warning system based on approach flight data provided by the present invention.
[0021] Figure 3 This is a block diagram of the electronic device provided by the present invention.
[0022] Figure label: 101. Predicted landing distance calculation module; 102. Judgment module; 103. Inference module; 104. Output module; 201. Processor; 202. Communication bus; 203. Communication interface; 204. Memory. Detailed Implementation
[0023] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. Based on the embodiments of this invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of this invention. The following embodiments are used to illustrate this invention but cannot be used to limit the scope of this invention.
[0024] In the description of this specification, the references to terms such as "one embodiment," "some embodiments," "example," "specific example," or "some examples," etc., refer to specific features, structures, materials, or characteristics described in connection with that embodiment or example, which are included in at least one embodiment or example of the present invention. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.
[0025] The following is combined with Figures 1 to 3 This invention describes a method and system for predicting and warning aircraft landing distance based on approach flight data.
[0026] like Figure 1 As shown, a method for predicting and warning aircraft landing distance based on approach flight data includes: S1: Acquire target aircraft approach flight data (QAR), preprocess the approach flight data to obtain approach phase time series samples, input the approach phase time series samples into a convolutional neural network model to obtain approach speed prediction values, and calculate the predicted landing distance under the current time window based on the approach speed prediction values.
[0027] The target aircraft's approach flight data includes ground speed, weight, wind speed, wind direction, vacuum speed, Mach number, indicated airspeed, longitudinal overload, barometric altitude, total temperature, left and right engine speeds, heading, pitch angle, vertical speed, and radio altitude, as shown in Table 1.
[0028] Table 1 Target aircraft approach flight data parameters
[0029] The target aircraft approach flight data includes both speed variables reflecting the aircraft's current motion state and environmental and configuration-related variables reflecting its dynamic and attitude characteristics, which can comprehensively describe the aircraft's operational status during the approach phase.
[0030] The preprocessing includes handling outliers and missing values in the target aircraft approach flight data. Data records with obvious errors are removed, and missing values can be filled in using nearest-time interpolation, sample mean, or other reasonable methods. Then, time period data relevant to modeling are selected according to the landing process of the target flight. Finally, the selected features are normalized to obtain the time series samples of the approach phase.
[0031] The convolutional neural network model includes an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, and a fully connected regression output layer. The first convolutional layer extracts short-term local features, the second convolutional layer further extracts high-level abstract features, the first and second pooling layers reduce feature dimensionality and suppress local noise, and the fully connected regression output layer outputs the final approach velocity prediction.
[0032] In preferred embodiments, the kernel size, number of output channels in the convolutional layer, and training optimization algorithm of the convolutional neural network can be adaptively selected based on the training sample size, QAR data sampling frequency, time window length, and prediction accuracy requirements. The convolutional neural network model extracts local features and deep correlation features of multidimensional flight parameters within a short time window during the approach phase, thereby improving the stability and accuracy of landing distance prediction results. The mapping relationship of the convolutional neural network can be written as: in, This is the predicted approach speed. For convolutional neural network mapping functions, The input is a QAR time series sample.
[0033] The predicted landing distance for the current time window is calculated based on the predicted approach velocity, including: S11: Calculate the inlet vacuum velocity based on the predicted approach velocity; in, Vacuum speed, The specific heat ratio of the standard atmosphere. For pressure, air density, This is the standard atmospheric sea-level pressure. This is the standard atmospheric sea-level density.
[0034] S12: Calculate the grounding velocity and braking initiation velocity based on the inlet vacuum velocity. The expression for calculating the grounding velocity is: in, For grounding speed, This is a grounding velocity correction factor; The formula for calculating the braking starting speed is: in, The initial braking speed, This is a correction factor for the braking start speed; S13: Calculate the ground speed at the start of braking based on the initial braking speed and wind speed. The calculation expression is: in, This represents the wind speed component along the runway direction. The ground speed at which braking begins.
[0035] S14: Calculate the distance of the air section based on the grounding speed and the inlet vacuum speed; calculate the distance of the ground transition section based on the grounding speed and the braking start speed; calculate the distance of the ground deceleration section based on the ground speed at the start of braking. The formula for calculating the distance in the air segment is: in, The distance in the air segment, Indicates a period of time in the air; The formula for calculating the distance of the ground transition section is: in, This refers to the distance of the ground transition section. This is the transition period from grounding to the automatic braking taking effect; The formula for calculating the distance of the ground deceleration section is: in, This refers to the distance of the ground deceleration section. To automatically brake the target to reduce speed.
[0036] S15: Calculate and predict the landing distance based on the distance in the air, the distance in the ground transition phase, and the distance in the ground deceleration phase; The formula for predicting the landing distance is: in, To predict the landing distance; The formula for calculating the predicted landing distance with a safety margin is: in, To predict the landing distance with a safety margin, This is the safety margin coefficient.
[0037] S2: Determine the stability and overall trend of the predicted landing distance under two adjacent time windows; S21: Stability is determined based on the predicted landing distances in two adjacent time windows; S211: Based on the runway threshold as the time base, The second is the time window for the aircraft's distance from the runway threshold. The time window is divided into minutes for approach flights from the decision point to the critical point. , This indicates the number of time windows used for trend determination.
[0038] Preferred, , .
[0039] S212: Calculate the predicted landing distance for the same flight in different time windows and arrange them in chronological order. Calculate the relative rate of change of the predicted landing distance between two adjacent time windows based on the predicted landing distance between two adjacent time windows. Determine the stability based on the relative rate of change of the predicted landing distance between two adjacent time windows. The formula for calculating the relative rate of change of the predicted landing distance between two adjacent time windows is: in, for Time window and The time window predicts the relative rate of change of landing distance. For the first Predicted landing distance for each time window, For the first Predicted landing distance for each time window.
[0040] The stability criterion is: in, To predict stability for two adjacent time windows.
[0041] when If the prediction result of the current window is small in difference from the prediction result of the previous window, the prediction result has good continuity, and the prediction result of the current window is determined to be stable. when If the forecast results in adjacent windows change significantly, the current forecast result may be affected by changes in flight status, wind field disturbances, or data fluctuations. The forecast result in the current window is determined to be unstable, and an unstable approach process warning is output. Pilots are advised to make a comprehensive judgment based on the actual flight status and operating environment.
[0042] S22: Judge the overall trend of change based on the predicted landing distance changes between two adjacent time windows; S221: Determine the trend of predicted landing distance changes between two adjacent time windows based on the changes in predicted landing distance between the two adjacent time windows; Changes in predicted landing distance between adjacent windows It can be represented as: in, This indicates that the predicted landing distance has increased compared to the previous window; This indicates that the predicted landing distance changes relatively gradually. This indicates that the predicted landing distance is decreasing relative to the previous window; the trend of predicted landing distance changes between adjacent windows. It can be represented as: in, This indicates that the predicted landing distance for the current window is trending upwards. This indicates that the predicted landing distance for the current window remains essentially unchanged or shows a downward trend.
[0043] S222: Calculation of the percentage of landing distance changes predicted between two adjacent time windows showing an upward trend. The calculation expression is as follows: in, The percentage of adjacent windows showing an upward trend; S223: Overall trend of change between two adjacent time windows It can be represented as: when hour, This indicates that the overall trend between two adjacent time windows is upward. , If the trend changes smoothly or decline between two adjacent time windows, it indicates that the overall trend is relatively flat or declining.
[0044] S3: Once the predicted landing distances of two adjacent time windows are determined to be stable, the predicted landing distances for subsequent time windows are obtained until the aircraft reaches the position corresponding to the decision height. After reaching the decision height, based on the overall trend of the predicted landing distances of the previous multiple time windows, a fitting function between the predicted time and the predicted landing distance is selected, and the average predicted landing distance of the time window is calculated. Based on the fitting function and the average predicted landing distance of the time window, the predicted landing distance of the aircraft at the time of arrival at the runway threshold is inferred. The fitting function between the predicted time and the predicted landing distance is: in, This represents the time it takes for the aircraft to reach the runway threshold. To fit the cutoff constant, These are the fitting coefficients. for Predicted landing distance at each time point. The function is adjusted based on predictions from multiple time windows. After fitting, Substituting 0 into the function, the predicted landing distance at the moment the aircraft arrives at the runway threshold is: in, This represents the inferred predicted landing distance at a distance of 0 seconds from the runway threshold. The average predicted landing distance is calculated for the time window.
[0045] The calculation expression is: when hour, for ;when hour, Average value This reduces the impact of single-window prediction errors on subsequent runway length constraint judgments. If the inference prediction with a safety margin predicts the landing distance... Less than the available length of the runway Then the output will be the predicted landing distance with a safety margin.
[0046] S4: Calculate the predicted landing distance with safety margin based on the predicted landing distance when the aircraft arrives at the runway threshold. Perform constraint judgment based on the predicted landing distance with safety margin and the available runway length. When the predicted landing distance with safety margin is less than or equal to the available runway length, output the predicted landing distance with safety margin; when the predicted landing distance with safety margin is greater than the available runway length, output a risk alarm signal. The formula for calculating the predicted landing distance with a safety margin is: in, To provide a predicted landing distance with a safety margin, This is the safety margin coefficient.
[0047] like Then, an automatic braking gear adjustment module is constructed, in which the automatic braking gear... Including 1, 2, ... Bmax, where Bmax is the highest gear, which increases the automatic braking gear. The predicted landing distance and the predicted landing distance with safety margin are calculated based on the adjusted braking speed at the time of re-arrival at the runway threshold.
[0048] If the adjusted predicted landing distance with safety margin is less than or equal to the available runway length, then the adjusted predicted landing distance with safety margin and automatic braking gear adjustment suggestions will be output.
[0049] If, after adjusting the automatic braking gear to the maximum level, the predicted landing distance with a safety margin is still greater than the available runway length, then operational handling information such as risk warnings, go-around suggestions, or reassessment of landing decisions will be output.
[0050] like Figure 2 As shown, an aircraft landing distance prediction and warning system based on approach flight data is used to execute the aforementioned aircraft landing distance prediction and warning method based on approach flight data, including: The predicted landing distance calculation module 101 acquires the approach flight data of the target aircraft, preprocesses the approach flight data to obtain the approach phase time series sample, inputs the approach phase time series sample into the convolutional neural network model to obtain the approach speed prediction value, and calculates the predicted landing distance under the current time window based on the approach speed prediction value. The judgment module 102 performs stability and overall trend judgment on the predicted landing distance under two adjacent time windows; Once the predicted landing distance between two adjacent time windows is determined to be stable, the inference module 103 obtains the predicted landing distance under the subsequent time window until the aircraft reaches the position corresponding to the decision height. After reaching the decision height, based on the overall trend of the predicted landing distance in the previous multiple time windows, it selects the fitting function between the predicted time and the predicted landing distance, calculates the average predicted landing distance of the time window, and infers the predicted landing distance of the aircraft at the time of arrival at the runway threshold based on the fitting function and the average predicted landing distance of the time window. The output module 104 calculates the predicted landing distance with safety margin based on the predicted landing distance when the aircraft arrives at the runway threshold. It then performs constraint judgment based on the predicted landing distance with safety margin and the available runway length. When the predicted landing distance with safety margin is less than or equal to the available runway length, it outputs the predicted landing distance with safety margin; when the predicted landing distance with safety margin is greater than the available runway length, it outputs a risk alarm signal.
[0051] Through the collaborative work of the above modules, multiple predicted landing distance results are continuously generated during the approach process, and the stability of the prediction is judged by the changes in the prediction results of adjacent time windows. When the prediction results are stable, the predicted landing distance at the runway threshold is further inferred by combining multiple prediction results before the decision height position, and then constrained by the available runway length. This improves the continuity, stability and operational decision reference value of the landing distance assessment results, and solves the problems of existing landing distance assessment methods that rely too much on the prediction results of a single time window, are easily affected by instantaneous data fluctuations, lack time window stability judgment, and lack a mechanism for inferring the predicted distance at the runway threshold.
[0052] To illustrate the multi-time-window prediction landing distance calculation and alarm system of the present invention, a flight sample is selected as an example. Taking the same flight sample in the embodiment as the object, starting from 240 seconds before the runway threshold, multiple time windows are constructed at 30-second update intervals, and the predicted landing distance corresponding to each time window is obtained. W7 is taken as the last prediction window before the decision height position. As the aircraft continues to approach, at each update time, the newly acquired QAR data is used to verify the consistency of the prediction result of the previous time window, and the subsequent prediction distance inference method and runway length constraint judgment result are determined accordingly. First, the stability of the predicted landing distance of two adjacent time windows is judged. If the relative change rate of the predicted landing distance of the current time window and the predicted landing distance of the previous time window is less than 10%, it is considered that the change of the prediction result of the adjacent window is small, and the prediction result of the window meets the stability requirements; if the relative change rate is greater than or equal to 10%, it is considered that the prediction result fluctuates greatly, and an unstable approach process warning is output. The asymptotic decision-making results under each time window are shown in Table 2. Among them, W1 is the initial window without the reference of the previous window and does not participate in the determination of relative change rate and stability.
[0053] Table 2. Incremental decision-making results under each time window
[0054] As shown in Table 2, the relative rate of change of the predicted landing distance for each adjacent time window is less than 10%, indicating that the prediction results for the flight before the decision height are generally stable, without any obvious abnormal jumps. Therefore, the system uses the predicted landing distances from these multiple stable windows to infer the predicted distance at the runway threshold.
[0055] Further analysis using the trend determination method reveals that the increase or decrease in predicted landing distances between adjacent time windows is statistically analyzed. Table 2 shows that the number of windows with an increasing trend is 2. Therefore, the overall trend determination value for the time windows in this example is: satisfy , If the time window forecast shows a flat or declining overall trend, then in this example, the average time window value is used as the final predicted landing distance at the runway threshold. The system infers the predicted landing distance of the flight at the runway threshold. It is 1699m. Even after considering safety margins, it is still less than the available runway length. This indicates that, after considering the safety margin, the predicted landing distance for the flight is still less than the available runway length, thus meeting the landing distance constraint. In this case, the system will not enter the automatic braking gear adjustment or risk management process, but will directly output the predicted landing distance at the runway threshold.
[0056] Therefore, the final output of the system in this example is: the predicted landing distance at the runway threshold is approximately 1699m, and this prediction satisfies the runway available length constraint. Through the above process, this invention can infer the final predicted landing distance using the prediction information from multiple windows when the prediction results are stable across multiple time windows, and combine this with the safety margin to complete the runway length constraint judgment, thereby providing a relatively stable reference result for the landing distance assessment during the approach and landing phase.
[0057] Figure 3 An example is a block diagram of an electronic device, such as... Figure 3 As shown, the electronic device may include a processor 201, a communication interface 203, a memory 204, and a communication bus 202. The processor 201, communication interface 203, and memory 204 communicate with each other via the communication bus 202. The processor 201 can call logical instructions from the memory 204 to execute an aircraft landing distance prediction and warning method based on approach flight data.
[0058] Furthermore, the logical instructions in the aforementioned memory 204 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, essentially, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0059] On the other hand, the present invention also provides a computer program product, the computer program product including a computer program stored on a non-transitory computer-readable storage medium, the computer program including program instructions, and when the program instructions are executed by a computer, the computer is able to execute a method for predicting and warning an aircraft landing distance based on approach flight data provided by the above methods.
[0060] In another aspect, the present invention also provides a non-transitory computer-readable storage medium storing a computer program thereon, which, when executed by a processor, implements a method for predicting and warning aircraft landing distance based on approach flight data provided by the above methods.
[0061] The device embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.
[0062] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.
[0063] 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 them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.
Claims
1. An aircraft landing distance prediction and alerting method based on approach flight data, characterized in that, include: S1: Acquire the target aircraft's approach flight data, preprocess the approach flight data to obtain approach phase time series samples, input the approach phase time series samples into a convolutional neural network model to obtain the approach speed prediction value, and calculate the predicted landing distance under the current time window based on the approach speed prediction value. S2: Determine the stability and overall trend of the predicted landing distance under two adjacent time windows; S3: Once the predicted landing distances of two adjacent time windows are determined to be stable, the predicted landing distances for subsequent time windows are obtained until the aircraft reaches the position corresponding to the decision height. After reaching the decision height, based on the overall trend of the predicted landing distances of the previous multiple time windows, a fitting function between the predicted time and the predicted landing distance is selected, and the average predicted landing distance of the time window is calculated. Based on the fitting function and the average predicted landing distance of the time window, the predicted landing distance of the aircraft at the time of arrival at the runway threshold is inferred. S4: Calculate the predicted landing distance with safety margin based on the predicted landing distance when the aircraft arrives at the runway threshold. Perform constraint judgment based on the predicted landing distance with safety margin and the available runway length. When the predicted landing distance with safety margin is less than or equal to the available runway length, output the predicted landing distance with safety margin; when the predicted landing distance with safety margin is greater than the available runway length, output a risk alarm signal.
2. The method of claim 1, wherein, The convolutional neural network model includes an input layer, a first convolutional layer, a first pooling layer, a second convolutional layer, a second pooling layer, and a fully connected regression output layer. The first convolutional layer is used to extract short-term local features, the second convolutional layer is used to extract high-level abstract features, the first pooling layer and the second pooling layer are used to reduce feature dimensionality and suppress local noise, and the fully connected regression output layer is used to output the approach velocity prediction value.
3. The method of claim 1, wherein, The predicted landing distance for the current time window is calculated based on the predicted approach velocity, including: S11: Calculate the inlet vacuum velocity based on the predicted approach velocity; S12: Calculate the grounding speed and braking start speed based on the inlet vacuum speed; S13: Calculate the ground speed at which braking begins based on the braking initiation speed and wind speed; S14: Calculate the distance of the air section based on the grounding speed and the inlet vacuum speed; calculate the distance of the ground transition section based on the grounding speed and the braking start speed; calculate the distance of the ground deceleration section based on the ground speed at the start of braking. S15: Calculate and predict the landing distance based on the distance in the air, the distance in the ground transition section, and the distance in the ground deceleration section.
4. The method of claim 1, wherein, Stability assessment based on predicted landing distances between two adjacent time windows includes: Taking the runway entrance as a time reference, a time window of 10 seconds is set for the aircraft approaching the runway entrance The time window is cut for the aircraft approaching the runway entrance 1-10 minutes before the decision height, and a time window is formed The time window is cut for the aircraft approaching the runway entrance 1-10 minutes before the decision height, and a time window is formed Calculate the predicted landing distance for the same flight in different time windows and arrange them in chronological order. Calculate the relative rate of change of the predicted landing distance between two adjacent time windows. Determine the stability based on the relative rate of change of the predicted landing distance between two adjacent time windows.
5. The method of claim 4, wherein, The stability criterion is: wherein, predicting stability for two adjacent time windows, for time windows and the relative change rate of the landing distance predicted by the time windows. When the current window prediction result is smaller than the previous window prediction result, the prediction result has good continuity, and it is determined that the current window prediction result is stable. when If the forecast results of adjacent windows change significantly, the current forecast result may be affected by changes in flight status, wind field disturbances, or data fluctuations, and the forecast result of the current window is determined to be unstable.
6. The method for predicting and warning aircraft landing distance based on approach flight data according to claim 4, characterized in that, Judging the overall trend based on the changes in predicted landing distance between two adjacent time windows includes: The trend of predicted landing distance changes between two adjacent time windows is determined by the change values of predicted landing distance between two adjacent time windows; The percentage of landing distances showing an upward trend between adjacent time windows is calculated based on the predicted landing distance change trends between adjacent time windows. The overall trend of change between two adjacent time windows can be judged by the percentage of those two windows showing an upward trend.
7. The method for predicting and warning aircraft landing distance based on approach flight data according to claim 1, characterized in that, The predicted landing distance at the moment the aircraft arrives at the runway threshold is: in, The predicted landing distance is calculated from a distance of 0 seconds from the runway threshold. The average predicted landing distance is calculated over the time window. This represents the overall trend between two adjacent time windows. This is the fitting cutoff constant.
8. A landing distance prediction and warning system for aircraft based on approach flight data, characterized in that, To perform the aircraft landing distance prediction and warning method based on approach flight data as described in any one of claims 1 to 7, comprising: The predicted landing distance calculation module acquires the target aircraft's approach flight data, preprocesses the approach flight data to obtain approach phase time series samples, inputs the approach phase time series samples into a convolutional neural network model to obtain approach speed prediction values, and calculates the predicted landing distance under the current time window based on the approach speed prediction values. The judgment module performs stability determination and overall trend determination on the predicted landing distance under two adjacent time windows; The inference module, after determining that the predicted landing distance of two adjacent time windows is stable, obtains the predicted landing distance of the subsequent time windows until the aircraft reaches the position corresponding to the decision height. After reaching the decision height, based on the overall trend of the predicted landing distance of the previous multiple time windows, selects the fitting function between the predicted time and the predicted landing distance, calculates the average predicted landing distance of the time window, and infers the predicted landing distance of the aircraft at the time of arrival at the runway threshold based on the fitting function and the average predicted landing distance of the time window. The output module calculates a predicted landing distance with a safety margin based on the predicted landing distance at the time the aircraft arrives at the runway threshold. It then performs constraint judgment based on the predicted landing distance with a safety margin and the available runway length. When the predicted landing distance with a safety margin is less than or equal to the available runway length, it outputs the predicted landing distance with a safety margin; when the predicted landing distance with a safety margin is greater than the available runway length, it outputs a risk alarm signal.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the steps of the aircraft landing distance prediction and warning method based on approach flight data as described in any one of claims 1 to 7.
10. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the aircraft landing distance prediction and warning method based on approach flight data as described in any one of claims 1 to 7.