A method for detecting non-intrusion area alarm of multiple airports in remote tower mode

Abstract

The embodiment of the application provides a multi-airport non-intrusion area alarm detection method in a remote tower mode, relates to the technical field of airport non-intrusion area alarm detection, and comprises the following steps: acquiring airport data, historical flight data of an airplane and real-time flight data of the airplane; based on takeoff / landing information of the airplane at the airport, the airport data is associated with the historical flight data and the real-time flight data to obtain associated training data and associated detection data; based on the associated training data, a model self-supervision training data set is constructed; an integrated neural network model is constructed, and the integrated neural network model is trained based on the model self-supervision training data set; data feature extraction is performed based on the associated detection data to obtain feature data; and the feature data is input into the trained integrated neural network model for detection to obtain a detection result. The application realizes alarm detection on a multi-airport non-intrusion area in a remote tower mode.

Description

Technical Field

[0001] This application relates to the field of airport non-intrusive zone alarm detection technology, and more specifically, to a method for detecting non-intrusive zone alarms in multiple airports under remote tower mode. Background Technology

[0002] With economic development, the demand for aircraft is constantly increasing, and the scale of airport construction is expanding. More and more airports are opening multiple runways to cope with the increasing number of flights in various regions. The tasks of airport tower control are becoming increasingly heavy. How to maintain good traffic order and ensure the safety of people's lives and property while improving flight efficiency is a key focus of traffic control work. Remote tower technology transfers traditional tower management from the airport to a remote command center. Compared with traditional towers, remote towers are easier to build and less expensive, and can significantly reduce airport operating costs.

[0003] Currently, existing non-intrusive zone alarm detection methods are not applicable to remote tower mode and cannot guarantee non-intrusive zone alarm detection services in complex airspace environments with multiple airports and runways. Therefore, how to achieve alarm detection in non-intrusive zones of multiple airports under remote tower mode is a technical problem that urgently needs to be solved. Summary of the Invention

[0004] The embodiments of this application provide a method for alarm detection in non-intrusive zones of multiple airports under remote tower mode, so as to solve the technical problem that the prior art cannot perform alarm detection in non-intrusive zones of multiple airports under remote tower mode.

[0005] Other features and advantages of this application will become apparent from the following detailed description, or may be learned in part from practice of this application.

[0006] According to a first aspect of the embodiments of this application, a method for detecting alarms in non-intrusive zones of multiple airports in a remote tower mode is provided, including:

[0007] Acquire airport data, historical flight data of aircraft, and real-time flight data of aircraft;

[0008] Based on the aircraft's takeoff / landing information at the airport, the airport data is correlated with the historical flight data and the real-time flight data to obtain correlated data, which includes correlated training data and correlated detection data.

[0009] Based on the aforementioned associated training data, a self-supervised training dataset for the model is constructed;

[0010] Construct an ensemble neural network model and train the ensemble neural network model based on the model's self-supervised training dataset;

[0011] Based on the aforementioned associated detection data, data features are extracted to obtain feature data;

[0012] The feature data is input into a trained ensemble neural network model for detection, and the detection results are obtained.

[0013] In some embodiments of this application, based on the foregoing scheme, the step of associating the airport data with the historical flight data and the real-time flight data based on the aircraft's takeoff / landing information at the airport to obtain associated data includes:

[0014] Determine the associated fields based on the aircraft's takeoff / landing information at the airport;

[0015] Based on the association field, multiple sub-data in the airport data are associated one by one with multiple sub-data in the historical flight data and the real-time flight data to obtain multiple associated sub-data;

[0016] By integrating and verifying multiple related sub-data, the related data is obtained.

[0017] In some embodiments of this application, based on the foregoing scheme, constructing a model self-supervised training dataset based on the associated training data includes:

[0018] Extract non-intrusive zone parameter data, flight parameter data, and alarm result data related to non-intrusive zone alarms from the airport data and the historical flight data;

[0019] The extracted flight parameter data, non-intrusive zone parameter data, and alarm result data are integrated to obtain a model self-supervised training dataset, which includes a training set, a test set, and a validation set.

[0020] In some embodiments of this application, based on the foregoing scheme, before extracting flight parameter data related to aircraft flight and non-intrusive zone parameter data related to non-intrusive zone alarms from the airport data and the historical flight data, the method further includes:

[0021] Outlier handling and missing value filling are performed on the airport data and the historical flight data.

[0022] In some embodiments of this application, based on the foregoing scheme, the construction of the ensemble neural network model includes:

[0023] Construct multiple subnetworks for detecting multiple non-intrusive zones;

[0024] The multiple sub-networks are integrated to construct the ensemble neural network model.

[0025] In some embodiments of this application, based on the foregoing scheme, the construction of multiple sub-networks for detecting multiple non-intrusive zones includes:

[0026] Construct multiple non-intrusive zone detection networks for a given non-intrusive zone;

[0027] Based on the test evaluation results, the optimal network is selected from the multiple non-intrusive area detection networks as the sub-network used to detect the non-intrusive areas corresponding to the multiple non-intrusive area detection networks.

[0028] In some embodiments of this application, based on the foregoing scheme, the construction of multiple non-intrusive zone detection networks for a non-intrusive zone includes:

[0029] Determine the model input for the non-intrusive zone detection network;

[0030] Determine the model output of the non-intrusive zone detection network;

[0031] Calculate the model structure of the non-intrusive zone detection network;

[0032] Define the forward propagation computation of the non-intrusive zone detection network;

[0033] Define the loss function for the non-intrusive zone detection network.

[0034] In some embodiments of this application, based on the foregoing scheme, the model structure for calculating the non-intrusive zone detection network includes:

[0035] The range of network layer numbers is set according to the number of neurons in the output layer of the non-invasive region detection network and the data types processed by the non-invasive region detection network;

[0036] The number of neurons is calculated based on the number of neurons in the input layer of the non-invasive region detection network, the number of neurons in the output layer of the non-invasive region detection network, and the number of samples in the training set.

[0037] In some embodiments of this application, based on the foregoing scheme, the step of selecting the optimal network from the plurality of non-intrusive area detection networks as a sub-network for detecting the non-intrusive areas corresponding to the plurality of non-intrusive area detection networks based on test evaluation results includes:

[0038] The performance of the trained non-intrusive region detection networks is evaluated based on the validation set.

[0039] The non-invasive region detection network with the best performance is selected as the sub-network used to detect the non-invasive regions corresponding to the multiple non-invasive region detection networks.

[0040] In some embodiments of this application, based on the foregoing scheme, after obtaining the detection results, the method further includes:

[0041] Based on the correlation characteristics in the associated data, the airport data and related flight data where the alarm result is located can be traced.

[0042] According to a second aspect of the embodiments of this application, a multi-airport non-intrusive zone alarm detection device in remote tower mode is provided, comprising:

[0043] The acquisition unit is used to acquire airport data, historical flight data of aircraft, and real-time flight data of aircraft.

[0044] The association unit is used to associate the airport data with the historical flight data and the real-time flight data based on the aircraft's takeoff / landing information at the airport to obtain associated data, which includes associated training data and associated detection data.

[0045] The first extraction unit is used to construct a model self-supervised training dataset based on the associated training data;

[0046] Model building unit, constructing ensemble neural network models;

[0047] The training unit is used to train the ensemble neural network model based on the model's self-supervised training dataset.

[0048] The second extraction unit is used to extract data features based on the associated detection data to obtain feature data;

[0049] The detection unit is used to input the feature data into the trained integrated neural network model for detection and obtain the detection result.

[0050] According to a third aspect of the embodiments of this application, an electronic device is provided, including: a memory and a processor;

[0051] The memory is used to store computer instructions;

[0052] The processor is configured to invoke the computer instructions to cause the electronic device to execute the method described in the first aspect above.

[0053] According to a fourth aspect of the embodiments of this application, a computer-readable storage medium is provided, the storage medium storing computer instructions, which, when executed on a computer, cause the computer to perform the method described in the first aspect above.

[0054] The technical solution of this application first trains an integrated neural network model based on airport data and historical flight data, and then uses the integrated neural network model to perform alarm detection on real-time flight data and airport data, realizing alarm detection in the non-intrusive zone of multiple airports under remote tower mode. Moreover, the trained integrated neural network model has strong robustness, ensuring the reliability of the prediction results.

[0055] It should be understood that the above general description and the following detailed description are exemplary and explanatory only, and do not limit this application. Attached Figure Description

[0056] The accompanying drawings, which are incorporated in and form part of this specification, illustrate embodiments consistent with this application and, together with the description, serve to explain the principles of this application. It is obvious that the drawings described below are merely some embodiments of this application, and those skilled in the art can obtain other drawings based on these drawings without any inventive effort. In the drawings:

[0057] Figure 1 A flowchart illustrating a method for detecting alarms in non-intrusive zones of multiple airports in a remote control tower mode, according to an embodiment of this application, is shown.

[0058] Figure 2 A schematic diagram of an integrated neural network model architecture according to an embodiment of this application is shown;

[0059] Figure 3 A schematic diagram of a multi-airport non-intrusive zone alarm detection device in a remote tower mode according to an embodiment of this application is shown.

[0060] Figure 4 A schematic diagram of an electronic device structure according to an embodiment of this application is shown;

[0061] Figure 5 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown. Detailed Implementation

[0062] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of the embodiments of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of this invention, and not all of them. 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.

[0063] Furthermore, the described features, structures, or characteristics can be combined in any suitable manner in one or more embodiments. Numerous specific details are provided in the following description to give a thorough understanding of embodiments of this application. However, those skilled in the art will recognize that the technical solutions of this application can be practiced without one or more of the specific details, or other methods, components, apparatuses, steps, etc., can be employed. In other instances, well-known methods, apparatuses, implementations, or operations are not shown or described in detail to avoid obscuring various aspects of this application.

[0064] The following detailed description of some embodiments of this application will be provided in conjunction with the accompanying drawings. Unless otherwise specified, the following embodiments and features can be combined with each other.

[0065] See Figure 1 The diagram illustrates a flowchart of a multi-airport non-intrusive zone alarm detection method in a remote tower mode according to an embodiment of this application.

[0066] like Figure 1 As shown, a method for detecting alarms in non-intrusive zones of multiple airports in a remote tower mode is illustrated, including steps S100 to S700.

[0067] Step S100: Obtain airport data, historical flight data of the aircraft, and real-time flight data of the aircraft.

[0068] It is understood that this embodiment acquires airport data from multiple airports. Each airport's data includes, but is not limited to, the following: airport name, airport code, airport location, flight information, runway information, NTZ area information, etc. Of course, each airport also needs to be numbered for differentiation.

[0069] The NTZ area refers to the non-invasive zone.

[0070] Continue to refer to Figure 1 In step S200, based on the aircraft's takeoff / landing information at the airport, the airport data is associated with the historical flight data and the real-time flight data to obtain associated data, which includes associated training data and associated detection data.

[0071] It is understood that the takeoff / landing information includes, but is not limited to: the airport number at which the aircraft takes off / landed, and the airport runway at which the aircraft takes off / landed.

[0072] It is understandable that linking flight data with airport data is to determine which airport the aircraft is currently landing at, making it easier to trace the source of the problem after knowing the alarm detection results.

[0073] In some feasible embodiments, based on the foregoing scheme, the association of airport data with historical flight data and real-time flight data based on aircraft takeoff / landing information at the airport to obtain associated data includes:

[0074] Determine the associated fields based on the aircraft's takeoff / landing information at the airport;

[0075] Based on the association field, multiple sub-data in the airport data are associated one by one with multiple sub-data in the historical flight data and the real-time flight data to obtain multiple associated sub-data;

[0076] By integrating and verifying multiple related sub-data, the related data is obtained.

[0077] It should be noted that after obtaining the associated detection data, the associated detection data needs to be updated and maintained in real time; the associated data needs to be verified and adjusted to ensure the accuracy and completeness of the association.

[0078] Understandably, data correlation can integrate airport data (including NTZ area data and runway data) and flight data (including flight data) to build a data network, forming a comprehensive perspective. This allows for a better understanding of the underlying meaning and relationships of the data, supports data tracking, and provides strong support for alarm information generation and alarm solution development. Simultaneously, it allows for data verification and calibration, thereby improving data reliability and accuracy, and helping to avoid erroneous alarm information and solutions based on incorrect or incomplete data.

[0079] Of course, the constructed data network can be updated in real time based on airport data and flight data, ensuring the timeliness and accuracy of the data network.

[0080] It is understood that the multiple sub-data in the airport data refers to airport-related data such as airport number, NTZ area data, and runway data, while the multiple sub-data in the real-time flight data and historical flight data refers to flight data and other data related to aircraft flight.

[0081] For example, if the flight data contains the field "landed on the first runway of airport 1" and the airport data contains the field "airport 1, first runway", then the two will be associated.

[0082] It should be noted that the associated training data is obtained by associating airport data with historical flight data, and the associated detection data is obtained by associating airport data with real-time flight data.

[0083] For example, the process of obtaining associated training data by associating airport data with historical flight data is as follows:

[0084] Determine the associated fields based on the aircraft's takeoff / landing information at the airport;

[0085] Based on the association field, multiple sub-data in the airport data are associated one by one with multiple sub-data in the historical flight data to obtain multiple associated sub-data;

[0086] By integrating and verifying multiple related sub-data, the related training data is obtained.

[0087] For example, the process of correlating airport data with real-time flight data to obtain correlated detection data is as follows:

[0088] Determine the associated fields based on the aircraft's takeoff / landing information at the airport;

[0089] Based on the association field, multiple sub-data in the airport data are associated one by one with multiple sub-data in the real-time flight data to obtain multiple associated sub-data;

[0090] By integrating and verifying multiple related sub-data, the related detection data is obtained.

[0091] Continue to refer to Figure 1 Step S300: Based on the associated training data, construct a model self-supervised training dataset.

[0092] It is understood that the associated training data includes historical flight data, which includes historical alarm results. The model is trained based on the historical alarm results to ensure that the model can have alarm detection capabilities.

[0093] In some feasible embodiments, based on the foregoing scheme, constructing a model self-supervised training dataset based on the associated training data includes:

[0094] Extract non-intrusive zone parameter data, flight parameter data, and alarm result data related to non-intrusive zone alarms from the airport data and the historical flight data;

[0095] The extracted flight parameter data, non-intrusive zone parameter data, and alarm result data are integrated to obtain a model self-supervised training dataset, which includes a training set, a test set, and a validation set.

[0096] It is understandable that when two aircraft are simultaneously approaching a parallel runway, if one aircraft enters the NTZ area, it means that there is a dangerous approach to the other aircraft approaching parallel to it. Therefore, based on the warning critical conditions of the aircraft in the non-intrusion zone, non-intrusion zone parameters are extracted from the airport data and flight parameter data are extracted from the historical flight data.

[0097] The parameters of the non-invasive zone are defined as follows:

[0098] The NTZ (Nearest Runway Zone) is a specific airspace located between the extended centerlines of two runways. The mathematical model of an NTZ is a three-dimensional shape with a rectangular base and right-angled trapezoidal sides. Therefore, by determining the width and length of the base rectangle, as well as the inner and outer heights of the right-angled trapezoids, the shape of an NTZ can be determined. Then, by defining the coordinates of the specific points contained within the rectangle, the parameters of the entire NTZ can be calculated by substituting these coordinates. Each NTZ is numbered and associated with its adjacent runways. The NTZ parameters include at least: 1) the location of the NTZ origin, represented by latitude and longitude coordinates; 2) the length of the NTZ; 3) the width of the NTZ; 4) the inner height of the NTZ; and 5) the outer height of the NTZ.

[0099] The flight parameters are defined as follows:

[0100] The target aircraft is numbered and associated with runway information. Various parameters of the aircraft during flight are defined. By defining and measuring these parameters, the aircraft's flight status can be monitored and evaluated, providing a basis for flight safety and control. Selecting appropriate aircraft performance parameters for monitoring allows for the timely detection of aircraft anomalies through calculations, enabling appropriate measures to be taken. The flight parameters involved in the non-intrusion zone alarm detection task include at least: 1) the aircraft's current position, expressed in latitude and longitude coordinates; 2) the aircraft's speed; 3) the aircraft's acceleration; 4) the aircraft's heading angle; and 5) the aircraft's bank angle.

[0101] For example, this step can be specifically described as follows:

[0102] Read flight parameter data and non-intrusive zone parameter data;

[0103] The extracted historical flight parameter data is merged and integrated with the parameter data of the current aircraft's adjacent NTZ area to obtain comprehensive feature data. The longitude of the NTZ origin is denoted as x0, the latitude as x1, the NTZ length as x2, the NTZ width as x3, the altitude within the NTZ as x4, the altitude outside the NTZ as x5, the longitude of the aircraft's current position as x6, the latitude as x7, the aircraft speed as x8, the aircraft acceleration as x9, and the aircraft heading angle as x. 10 The aircraft's bank angle is denoted as x. 11 .

[0104]

[0105] The actual alarm results extracted from historical flight data are used as label data and correspond one-to-one with the feature data. The alarm result is recorded as y, that is, y is labeled as 1 when there is an alarm and y is labeled as 0 when there is no alarm.

[0106] If the classes in the dataset are imbalanced (the number of samples in a certain class is small), measures can be taken to balance the dataset, including undersampling, oversampling, or synthesizing samples.

[0107] Standardizing data ensures that different features have the same scale. Standardization methods include Z-score standardization and Min-Max standardization.

[0108] All labeled historical datasets are divided into training, validation, and test sets in a 6:2:2 ratio. The training set is used for model training, the validation set is used for parameter tuning, and the test set is used to evaluate model performance.

[0109] In some feasible embodiments, based on the foregoing scheme, before extracting flight parameter data related to aircraft flight and non-intrusive zone parameter data related to non-intrusive zone alarms from the airport data and the historical flight data, the method further includes:

[0110] Outlier handling and missing value filling are performed on the airport data and the historical flight data.

[0111] For example, the outlier handling process in this embodiment can be specifically described as follows:

[0112] This is achieved by replacing outlier values ​​with reasonable ones.

[0113] For example, the missing value imputation process in this embodiment can be specifically described as follows:

[0114] Missing values ​​are filled using an interpolation method.

[0115] Understandably, handling outliers and imputing missing values ​​in data can reduce errors and biases in the data analysis and modeling process, and improve the accuracy and reliability of the model.

[0116] Continue to refer to Figure 1 Step S400: Construct an integrated neural network model.

[0117] Understandably, in this technical solution, the integrated neural network model is jointly modeled by a set of sub-networks and used for alarm detection in multiple NTZ regions.

[0118] In some feasible embodiments, based on the foregoing scheme, the construction of the ensemble neural network model includes:

[0119] Construct multiple subnetworks for detecting multiple non-intrusive zones;

[0120] The multiple sub-networks are integrated to construct the ensemble neural network model.

[0121] Understandably, this solution treats alarm detection as a classification problem and proposes an ensemble learning approach to construct a parallel ensemble network model. This involves using a set of sub-networks to jointly model the alarms, with each sub-network responsible for alarm calculation in an NTZ region. This overcomes the problem that traditional calculation methods cannot handle multiple tasks.

[0122] For example, this step can be specifically described as follows:

[0123] Defining the subnetwork structure, fitting nonlinear relationships, and calculating the forward propagation result of the neural network requires three parts of information: The first part is the connection structure of the neural network; the number of hidden layers and neurons needs to be adjusted based on the actual training data. The second part is the input of the neural network, which includes NTZ parameters and aircraft parameters, etc. The third part is the parameters of each neuron, namely the weight W and bias b. The parameters of the neural network are initialized using a random number generator, and then continuously updated during training. Given the input of the neural network, the structure of the neural network, and the edge weight parameters, the output of the neural network can be calculated using the forward propagation algorithm. In this paper, the output of the model is a label for the alarm result, i.e., 1 for alarm and 0 for no alarm.

[0124] In some feasible embodiments, based on the foregoing scheme, the construction of multiple sub-networks for detecting multiple non-intrusive zones includes:

[0125] Construct multiple non-intrusive zone detection networks for a given non-intrusive zone;

[0126] Based on the test evaluation results, the optimal network is selected from the multiple non-intrusive area detection networks as the sub-network used to detect the non-intrusive areas corresponding to the multiple non-intrusive area detection networks.

[0127] Understandably, a non-intrusive region corresponds to multiple non-intrusive region detection networks. The optimal one is selected from these networks to ensure that the model's performance is optimal, thereby improving the model's detection accuracy.

[0128] In some feasible embodiments, based on the foregoing scheme, constructing multiple non-invasive zone detection networks for a single non-invasive zone includes:

[0129] Determine the model input for the non-intrusive zone detection network;

[0130] Determine the model output of the non-intrusive zone detection network;

[0131] Calculate the model structure of the non-intrusive zone detection network;

[0132] Define the forward propagation computation of the non-intrusive zone detection network;

[0133] Define the loss function for the non-intrusive zone detection network.

[0134] Understandably, this step represents the construction process of each non-intrusive zone detection network. Through this step, non-intrusive zone detection networks for multiple non-intrusive zones can be constructed.

[0135] For example, the construction of a non-intrusive zone detection network mainly includes the following process:

[0136] (1) Determine the input of the non-intrusive zone detection network model, which is the feature dataset X;

[0137] (2) Determine the output of the non-intrusive zone detection network model, which is the feature dataset Y;

[0138] (3) Determine the structure of the non-intrusive zone detection network model, select the feedforward neural network model, and configure the model structure parameters using an automatic generation method;

[0139] (4) Define the forward propagation calculation of the non-intrusive zone detection network;

[0140] In this process, n represents the number of samples, and the input is x. i The weight of the l-th layer is W. l The bias is b l The activation function is denoted as f(x), and the output is... The calculation process is as follows:

[0141]

[0142] (5) Define the loss function

[0143] The predicted value represents the final output, and the true value represents the actual value. The loss function is defined to represent the model's performance on the current task. The mean squared error loss function can be used, but is not limited to it, and is defined as follows:

[0144]

[0145] Where n represents the number of samples. Indicates the prediction result, y i Represents the actual value.

[0146] In some feasible embodiments, based on the foregoing scheme, the model structure for computing the non-intrusive zone detection network includes:

[0147] The range of network layer numbers is set according to the number of neurons in the output layer of the non-invasive region detection network and the data types processed by the non-invasive region detection network;

[0148] The number of neurons is calculated based on the number of neurons in the input layer of the non-invasive region detection network, the number of neurons in the output layer of the non-invasive region detection network, and the number of samples in the training set.

[0149] It should be noted that the structure of traditional neural networks needs to be designed manually, and parameter design is usually based on the designer's experience and professional knowledge, determined through subjective judgment or analogical inference. When applying the model to a real system, the network structure needs to be designed according to the business scenario and real data, and the model needs to be trained in real time. This requires a high level of professional skills from the operators and is not easy to use.

[0150] To address the shortcomings of the aforementioned network structure design, this application provides the steps outlined above. These steps can generate suitable network architectures based on different task objectives and constraints, enabling training and evaluation. This saves time and costs associated with manual design and debugging, and improves the performance and efficiency of the model.

[0151] It should be noted that, theoretically, the deeper the network layers, the stronger the ability to fit the function and the better the results should be. However, in practice, deeper layers may lead to overfitting and increase the difficulty of training, making it difficult for the model to converge.

[0152] No hidden layers: can only represent linearly separable functions or decisions.

[0153] A hidden layer of 1 allows fitting of any function that contains a continuous mapping from one finite space to another.

[0154] With 2 hidden layers, and with an appropriate activation function, it can represent any decision boundary with arbitrary precision and fit any smooth mapping with any precision.

[0155] The number of hidden layers is greater than 2: the additional hidden layers can learn complex descriptions (a kind of automatic feature engineering).

[0156] For example, NTZ alarm tasks do not involve complex image, audio, or video information and lack multi-dimensional data. Therefore, the hidden layer range recommended by this invention is set to W = [1, 2, 3].

[0157] It's important to note that too few neurons in the hidden layer can lead to underfitting, while too many can lead to overfitting. When a neural network has excessive information processing power, the limited information in the training set may not be sufficient to train all the neurons in the hidden layers, resulting in overfitting. Even if the training data contains enough information, too many neurons in the hidden layers will increase training time, making it difficult to achieve the desired results. Typically, the first layer can learn many low-order features, which are then transferred to subsequent layers to extract higher-order features. Therefore, setting a larger first layer followed by smaller layers will lead to better performance.

[0158] For example, the number of neurons N in the hidden layer h The calculation formula is as follows:

[0159]

[0160] Where, N i N is the number of neurons in the input layer, and N0 is the number of neurons in the output layer. S N is the number of samples in the training set, and α is a variable value ranging from [2, 10]. Then, the calculated results are filtered. h It should satisfy N0 < N h <N i .

[0161] Continue to refer to Figure 1 Step S500: Train the ensemble neural network model based on the model's self-supervised training dataset.

[0162] Understandably, in the specific training process, each non-intrusive region detection network is trained using the training set in the model self-supervised training dataset.

[0163] In some feasible embodiments, based on the foregoing scheme, the step of selecting the optimal network from the plurality of non-invasive region detection networks as a sub-network for detecting the non-invasive regions corresponding to the plurality of non-invasive region detection networks, based on the test evaluation results, includes:

[0164] The performance of the trained non-intrusive region detection networks is evaluated based on the validation set.

[0165] The non-invasive region detection network with the best performance is selected as the sub-network used to detect the non-invasive regions corresponding to the multiple non-invasive region detection networks.

[0166] Understandably, models can be adaptively and intelligently rebuilt and retrained based on changes in application scenarios.

[0167] Continue to refer to Figure 1Step S600: Based on the associated detection data, perform data feature extraction to obtain feature data.

[0168] Continue to refer to Figure 1 In step S700, the feature data is input into the trained integrated neural network model for detection to obtain the detection result.

[0169] Understandably, the model's output is the detection result.

[0170] In some feasible embodiments, based on the foregoing scheme, after obtaining the detection results, the method further includes:

[0171] Based on the correlation characteristics in the associated data, the airport data and related flight data where the alarm result is located can be traced.

[0172] The following describes an embodiment of the apparatus described in this application, which can be used to execute a multi-airport non-intrusion zone alarm detection method in a remote tower mode as described in the above embodiments of this application. For details not disclosed in the apparatus embodiments of this application, please refer to the embodiments of the method described in the above applications.

[0173] Reference Figure 3 As shown, a multi-airport non-intrusive zone alarm detection device 300 in a remote tower mode according to an embodiment of this application includes: an acquisition unit 301, an association unit 302, a first extraction unit 303, a model building unit 304, a training unit 305, a second extraction unit 306, and a detection unit 307.

[0174] The system includes: an acquisition unit 301 for acquiring airport data, historical flight data of the aircraft, and real-time flight data of the aircraft; an association unit 302 for associating the airport data with the historical flight data and the real-time flight data based on the aircraft's takeoff / landing information at the airport to obtain associated data, which includes associated training data and associated detection data; a first extraction unit 303 for constructing a model self-supervised training dataset based on the associated training data; a model construction unit 304 for constructing an ensemble neural network model; a training unit 305 for training the ensemble neural network model based on the model self-supervised training dataset; a second extraction unit 306 for extracting data features based on the associated detection data to obtain feature data; and a detection unit 307 for inputting the feature data into the trained ensemble neural network model for detection to obtain detection results.

[0175] like Figure 4As shown, this application embodiment also provides an electronic device 400, including a memory 410, a processor 420, and a computer program 411 stored in the memory 410 and executable on the processor. When the processor 420 executes the computer program 411, it implements the above-mentioned method for detecting alarms in non-intrusive zones of multiple airports under remote tower mode.

[0176] Since the electronic device described in this embodiment is the device used to implement the multi-airport non-intrusive zone alarm detection device in the remote tower mode of this application embodiment, those skilled in the art can understand the specific implementation method and its various variations of the electronic device in this embodiment based on the method described in this application embodiment. Therefore, how the electronic device implements the method in this application embodiment will not be described in detail here. As long as those skilled in the art implement the method in this application embodiment, the device used is within the scope of protection of this application.

[0177] In practice, when the computer program 411 is executed by the processor, it can implement any of the embodiments corresponding to the first aspect.

[0178] Figure 5 A schematic diagram of the structure of a computer system suitable for implementing the electronic device of the present application is shown.

[0179] It should be noted that, Figure 5 The computer system 500 of the electronic device shown is merely an example and should not impose any limitation on the functionality and scope of use of the embodiments of this application.

[0180] like Figure 5 As shown, the computer system 500 includes a Central Processing Unit (CPU) 501, which can perform various appropriate actions and processes based on programs stored in Read-Only Memory (ROM) 502 or programs loaded from storage portion 508 into Random Access Memory (RAM) 503, such as performing the methods described in the above embodiments. The RAM 503 also stores various programs and data required for system operation. The CPU 501, ROM 502, and RAM 503 are interconnected via a bus 504. An Input / Output (I / O) interface 505 is also connected to the bus 504.

[0181] The following components are connected to I / O interface 505: an input section 506 including a keyboard, mouse, etc.; an output section 507 including a cathode ray tube (CRT), liquid crystal display (LCD), etc., and speakers, etc.; a storage section 508 including a hard disk, etc.; and a communication section 509 including a network interface card such as a LAN (Local Area Network) card, modem, etc. The communication section 509 performs communication processing via a network such as the Internet. A drive 510 is also connected to I / O interface 505 as needed. Removable media 511, such as a disk, optical disk, magneto-optical disk, semiconductor memory, etc., are installed on drive 510 as needed so that computer programs read from them can be installed into storage section 508 as needed.

[0182] Specifically, according to embodiments of this application, the processes described above with reference to the flowcharts can be implemented as computer software programs. For example, embodiments of this application include a computer program product comprising a computer program carried on a computer-readable medium, the computer program containing program code for performing the methods shown in the flowcharts. In such embodiments, the computer program can be downloaded and installed from a network via communication section 509, and / or installed from removable medium 511. When the computer program is executed by central processing unit (CPU) 501, it performs various functions defined in the system of this application.

[0183] It should be noted that the computer-readable medium shown in the embodiments of this application can be a computer-readable signal medium or a computer-readable storage medium, or any combination of the two. A computer-readable storage medium can be, for example,—but not limited to—an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, apparatus, or device, or any combination thereof. More specific examples of a computer-readable storage medium may include, but are not limited to: an electrical connection having one or more wires, a portable computer disk, a hard disk, random access memory (RAM), read-only memory (ROM), erasable programmable read-only memory (EPROM), flash memory, optical fiber, portable compact disc read-only memory (CD-ROM), optical storage device, magnetic storage device, or any suitable combination thereof. In this application, a computer-readable storage medium can be any tangible medium containing or storing a program that can be used by or in conjunction with an instruction execution system, apparatus, or device. In this application, a computer-readable signal medium can include a data signal propagated in baseband or as part of a carrier wave, carrying computer-readable program code. Such transmitted data signals can take various forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination thereof. The computer-readable signal medium can also be any computer-readable medium other than a computer-readable storage medium, which can send, propagate, or transmit a program for use by or in connection with an instruction execution system, apparatus, or device. The program code contained on the computer-readable medium can be transmitted using any suitable medium, including but not limited to wireless, wired, etc., or any suitable combination thereof.

[0184] The flowcharts and block diagrams in the accompanying drawings illustrate the architecture, functionality, and operation of possible implementations of systems, methods, and computer program products according to various embodiments of this application. Each block in a flowchart or block diagram may represent a module, segment, or portion of code, which contains one or more executable instructions for implementing a specified logical function. It should also be noted that in some alternative implementations, the functions indicated in the blocks may occur in a different order than those indicated in the drawings. For example, two consecutively indicated blocks may actually be executed substantially in parallel, and they may sometimes be executed in reverse order, depending on the functions involved. It should also be noted that each block in a block diagram or flowchart, and combinations of blocks in a block diagram or flowchart, can be implemented using a dedicated hardware-based system that performs the specified function or operation, or using a combination of dedicated hardware and computer instructions.

[0185] The units described in the embodiments of this application can be implemented in software or hardware, and the described units can also be located in a processor. The names of these units do not necessarily limit the specific unit itself.

[0186] In another aspect, this application also provides a computer program product or computer program including computer instructions stored in a computer-readable storage medium. A processor of a computer device reads the computer instructions from the computer-readable storage medium and executes the computer instructions, causing the computer device to perform the multi-airport non-intrusive zone alarm detection method described in the above embodiments.

[0187] In another aspect, this application also provides a computer-readable medium, which may be included in the electronic device described in the above embodiments; or it may exist independently and not assembled into the electronic device. The computer-readable medium carries one or more programs, which, when executed by the electronic device, cause the electronic device to implement the multi-airport non-intrusive zone alarm detection method in remote tower mode described in the above embodiments.

[0188] It should be noted that although several modules or units for the device used to perform actions have been mentioned in the detailed description above, this division is not mandatory. In fact, according to the embodiments of this application, the features and functions of two or more modules or units described above can be embodied in one module or unit. Conversely, the features and functions of one module or unit described above can be further divided and embodied by multiple modules or units.

[0189] Through the above description of the embodiments, those skilled in the art will readily understand that the exemplary embodiments described herein can be implemented by software or by combining software with necessary hardware. Therefore, the technical solutions according to the embodiments of this application can be embodied in the form of a software product, which can be stored in a non-volatile storage medium (such as a CD-ROM, USB flash drive, external hard drive, etc.) or on a network, including several instructions to cause a computing device (such as a personal computer, server, touch terminal, or network device, etc.) to execute the method according to the embodiments of this application.

[0190] Other embodiments of this application will readily conceive of by those skilled in the art upon consideration of the specification and practice of the embodiments disclosed herein. This application is intended to cover any variations, uses, or adaptations of this application that follow the general principles of this application and include common knowledge or customary techniques in the art not disclosed herein. It should be understood that this application is not limited to the precise structures described above and shown in the accompanying drawings, and various modifications and changes can be made without departing from its scope. The scope of this application is limited only by the appended claims.

Claims

1. A method for detecting alarms in non-intrusive zones of multiple airports under remote tower mode, characterized in that, include: Acquire airport data, historical flight data of aircraft, and real-time flight data of aircraft; Based on the aircraft's takeoff / landing information at the airport, the airport data is associated with the historical flight data and the real-time flight data to obtain associated data, which includes associated training data and associated detection data. Based on the aforementioned associated training data, a self-supervised training dataset for the model is constructed; Construct an ensemble neural network model and train the ensemble neural network model based on the model's self-supervised training dataset; Based on the aforementioned associated detection data, data features are extracted to obtain feature data; The feature data is input into a trained ensemble neural network model for detection, and the detection result is obtained. The construction of the model self-supervised training dataset based on the associated training data includes: Extract non-intrusive zone parameter data, flight parameter data, and alarm result data related to non-intrusive zone alarms from the airport data and the historical flight data; The extracted flight parameter data, non-intrusive zone parameter data, and alarm result data are integrated to obtain a model self-supervised training dataset, which includes a training set, a test set, and a validation set. The construction of the integrated neural network model includes: Construct multiple subnetworks for detecting multiple non-intrusive zones; The multiple sub-networks are integrated to construct the ensemble neural network model; The construction of multiple sub-networks for detecting multiple non-intrusive zones includes: Construct multiple non-intrusive zone detection networks for a given non-intrusive zone; Based on the test evaluation results, the optimal network is selected from the multiple non-intrusive area detection networks as the sub-network used to detect the non-intrusive areas corresponding to the multiple non-intrusive area detection networks.

2. The method according to claim 1, characterized in that, The method involves associating airport data with historical flight data and real-time flight data based on aircraft takeoff / landing information at the airport, resulting in associated data, including: Determine the associated fields based on the aircraft's takeoff / landing information at the airport; Based on the association field, multiple sub-data in the airport data are associated one by one with multiple sub-data in the historical flight data and the real-time flight data to obtain multiple associated sub-data; By integrating and verifying multiple related sub-data, the related data is obtained.

3. The method according to claim 1, characterized in that, Before extracting flight parameter data related to aircraft flight and non-intrusive zone parameter data related to non-intrusive zone alarms from the airport data and the historical flight data, the process also includes: Outlier handling and missing value filling are performed on the airport data and the historical flight data.

4. The method according to claim 1, characterized in that, The construction of multiple non-invasive zone detection networks for a single non-invasive zone includes: Determine the model input for the non-intrusive zone detection network; Determine the model output of the non-intrusive zone detection network; Calculate the model structure of the non-intrusive zone detection network; Define the forward propagation computation of the non-intrusive zone detection network; Define the loss function for the non-intrusive zone detection network.

5. The method according to claim 4, characterized in that, The model structure of the computational non-intrusive zone detection network includes: The range of network layer numbers is set according to the number of neurons in the output layer of the non-invasive region detection network and the data types processed by the non-invasive region detection network; The number of neurons is calculated based on the number of neurons in the input layer of the non-invasive region detection network, the number of neurons in the output layer of the non-invasive region detection network, and the number of samples in the training set.

6. The method according to claim 5, characterized in that, The step of selecting the optimal network from the multiple non-invasive region detection networks based on test evaluation results as the sub-network for detecting the non-invasive regions corresponding to the multiple non-invasive region detection networks includes: The performance of the trained non-intrusive region detection networks is evaluated based on the validation set. The non-invasive region detection network with the best performance is selected as the sub-network used to detect the non-invasive regions corresponding to the multiple non-invasive region detection networks.

7. The method according to claim 1, characterized in that, After obtaining the test results, the following is also included: Based on the correlation characteristics in the associated data, the airport data and related flight data where the alarm result is located can be traced.

Images