Method and device for generating checklist based on large model
Through the large-model-based checklist generation method, the deep learning model is used to automatically generate and adjust checklists, which solves the problems of traditional manual checklist time-consuming and labor-intensive and human errors, real-time and dynamic checklist generation, and improves the Flight safety.
Patent Information
- Application Number
- CN202411113558.0
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-08-14
- Publication Date
- 2024-11-12
AI Technical Summary
The traditional way of generating checklists relies on manual creation, which is time-consuming and labor-intensive, and has human errors, so it cannot be adjusted in time, especially when an emergency occurs after the aircraft takes off, it is difficult to update dynamically.
The large-model-based checklist generation method is adopted to obtain flight data and flight mission information in real time, and use deep learning models to automatically generate checklists, and adjust them based on real-time data, including checklists before, during and after flight.
It improves the efficiency and accuracy of the checklist, enhances dynamic adaptability, ensures flight safety, reduces human errors, and can respond to emergencies in a timely manner.
Smart Images

Figure CN118940739A_ABST
Abstract
Description
Technical Field
[0001] The present application relates to the field of computer technology, and in particular to a method and device for generating a checklist based on a large model. Background Art
[0002] At present, the traditional checklist generation method is mostly manual creation, which requires a lot of time and manpower, and there are also errors caused by human factors. In addition, the existing checklist creation method is only created before the aircraft takes off. If an emergency occurs after the aircraft takes off, the checklist cannot be changed in time, which greatly tests the crew's experience. Summary of the invention
[0003] In view of this, the present application provides a method and device for generating a checklist based on a large model, which automatically generates a checklist and can adjust the checklist in real time according to flight data and flight mission information, thereby improving efficiency and accuracy while enhancing the dynamic adaptability of the checklist.
[0004] The first aspect of the present application provides a method for generating a checklist based on a large model, comprising:
[0005] Acquire flight data and flight mission information in real time; wherein the flight data and flight mission information include pilot information, airport information, route information, weather information and aircraft information;
[0006] The flight data and the flight mission information are input into a checklist generation model, and a checklist for this flight mission is output; wherein the checklist generation model includes a standard checklist template; the checklist generation model is obtained by training a deep learning model with a training sample set; the training sample set includes at least one training sample, and the training sample includes historical flight data and historical flight mission information as well as historical checklists for historical flight missions; the checklist for this flight mission includes a pre-flight checklist, an in-flight checklist, and a post-flight checklist.
[0007] Optionally, the method for constructing the checklist generation model includes:
[0008] Constructing a training sample set; wherein the training sample set includes at least one training sample; the training sample includes historical flight data and historical flight mission information and historical checklists of historical flight missions;
[0009] Inputting the historical flight data and the historical flight mission information into a deep learning model, and outputting predicted inspection information of the historical flight mission;
[0010] generating a predicted checklist for the historical flight mission according to the predicted checklist information of the historical flight mission and the standard checklist template;
[0011] The hyperparameters in the deep learning model are adjusted using the error between the predicted checklist of the historical flight mission and the historical checklist of the historical flight mission until the error between the predicted checklist of the historical flight mission and the historical checklist of the historical flight mission meets a preset convergence condition, and the deep learning model is used as a checklist generation model.
[0012] Optionally, after inputting the flight data and the flight mission information into a checklist generation model and outputting a checklist for the current flight mission, the method further includes:
[0013] Receive the emergency instruction, immediately execute the real-time acquisition of flight data and flight mission information, input the flight data and the flight mission information into the checklist generation model, output the checklist steps of this flight mission, and obtain the emergency checklist.
[0014] Optionally, after inputting the flight data and the flight mission information into a checklist generation model and outputting a checklist for the current flight mission, the method further includes:
[0015] Receive checklist query instructions and display the checklist for this flight mission;
[0016] Receive user's review information;
[0017] If the user's audit information indicates that no adjustment is required, no action will be taken;
[0018] If the user's review information indicates that there is an adjustment, collecting the user's feedback information and actual adjustment information;
[0019] The checklist generation model is optimized according to the user's feedback information and the actual adjustment information to obtain an optimized checklist generation model.
[0020] Optionally, after inputting the flight data and the flight mission information into a checklist generation model and outputting a checklist for the current flight mission, the method further includes:
[0021] Real-time monitoring of flight data and checklist execution;
[0022] If the checklist is not followed or the flight data is abnormal, an abnormal feedback message will be generated.
[0023] Optionally, the method for generating a checklist based on a large model further includes:
[0024] The standard checklist templates in the checklist generation model are automatically updated regularly based on the latest aviation regulations, safety standards, and aircraft upgrade information.
[0025] Optionally, after acquiring the flight data and flight mission information in real time, the method further includes:
[0026] A risk assessment result of the flight mission is generated based on the historical flight record data, the route information, the flight data and the external intelligence data; wherein the historical flight record data at least includes historical accident reports and mechanical failures; and the external intelligence data at least includes controlled traffic density and severe weather forecasts.
[0027] Optionally, after acquiring the flight data and flight mission information in real time, the method further includes:
[0028] Acquiring the pilot's personality data; wherein the pilot's personality data includes: flight experience, flight records, pilot's pre-set preferences, and pilot's feedback information;
[0029] constructing a user profile of the pilot according to the personality data of the pilot;
[0030] The step of inputting the flight data and the flight mission information into a checklist generation model and outputting a checklist for the flight mission includes:
[0031] The pilot's user portrait, the flight data, and the flight mission information are input into a checklist generation model, and a checklist for this flight mission is output.
[0032] Optionally, each check item in the checklist of this flight mission is added with corresponding self-explanatory information.
[0033] The second aspect of the present application provides a device for generating a checklist based on a large model, comprising:
[0034] A first acquisition unit is used to acquire flight data and flight mission information in real time; wherein the flight data and flight mission information include pilot information, airport information, route information, weather information and aircraft information;
[0035] A first input unit is used to input the flight data and the flight mission information into a checklist generation model, and output a checklist for this flight mission; wherein the checklist generation model includes a standard checklist template; the checklist generation model is obtained by training a deep learning model with a training sample set; the training sample set includes at least one training sample, and the training sample includes historical flight data and historical flight mission information as well as historical checklists for historical flight missions; the checklist for this flight mission includes a pre-flight checklist, an in-flight checklist, and a post-flight checklist.
[0036] Optionally, the construction unit of the checklist generation model includes:
[0037] A training sample set construction unit, used to construct a training sample set; wherein the training sample set includes at least one training sample; the training sample includes historical flight data and historical flight mission information and historical checklists of historical flight missions;
[0038] A second input unit, used to input the historical flight data and the historical flight mission information into a deep learning model, and output predicted inspection information of the historical flight mission;
[0039] A generating unit, configured to generate a predicted checklist for the historical flight mission according to the predicted checklist information of the historical flight mission and the standard checklist template;
[0040] An adjustment unit is used to adjust the hyperparameters in the deep learning model by using the error between the predicted checklist of the historical flight mission and the historical checklist of the historical flight mission, until the error between the predicted checklist of the historical flight mission and the historical checklist of the historical flight mission meets a preset convergence condition, and the deep learning model is used as a checklist generation model.
[0041] Optionally, the device for generating the checklist based on the large model further includes:
[0042] The emergency processing unit is used to receive the emergency instruction, immediately execute the real-time acquisition of flight data and flight mission information, input the flight data and the flight mission information into the checklist generation model, output the checklist steps of this flight mission, and obtain the emergency checklist.
[0043] Optionally, the device for generating the checklist based on the large model further includes:
[0044] A first receiving unit is used to receive a checklist query instruction and display the checklist for this flight mission;
[0045] A second receiving unit is used to receive the user's audit information; if the user's audit information indicates that no adjustment is required, no processing is performed;
[0046] A collecting unit, configured to collect user feedback information and actual adjustment information if the user's review information indicates that there is an adjustment;
[0047] The optimization unit is used to optimize the checklist generation model according to the user's feedback information and the actual adjustment information to obtain an optimized checklist generation model.
[0048] Optionally, the device for generating the checklist based on the large model further includes:
[0049] Monitoring unit, used to monitor flight data and checklist execution in real time;
[0050] The abnormality feedback unit is used to generate abnormality feedback information if the checklist is not executed or the flight data is abnormal.
[0051] Optionally, the device for generating the checklist based on the large model further includes:
[0052] An updating unit is used to automatically update the standard checklist template in the checklist generation model according to the latest aviation regulations, safety standards, and aircraft upgrade information on a regular basis.
[0053] Optionally, the device for generating the checklist based on the large model further includes:
[0054] The risk assessment unit is used to generate a risk assessment result of the current flight mission based on historical flight record data, the route information, the flight data and external intelligence data; wherein the historical flight record data at least includes historical accident reports and mechanical failures; and the external intelligence data at least includes controlled traffic density and severe weather forecasts.
[0055] Optionally, the device for generating the checklist based on the large model further includes:
[0056] A second acquisition unit is used to acquire the pilot's personality data; wherein the pilot's personality data includes: flight experience, flight records, pilot's preset preference settings and pilot's feedback information;
[0057] A portrait construction unit, configured to construct a user portrait of the pilot according to the personality data of the pilot;
[0058] Wherein, the first input unit is used to:
[0059] The pilot's user portrait, the flight data, and the flight mission information are input into a checklist generation model, and a checklist for this flight mission is output.
[0060] Optionally, each check item in the checklist of this flight mission is added with corresponding self-explanatory information.
[0061] A third aspect of the present application provides an electronic device, including:
[0062] one or more processors;
[0063] a storage device having one or more programs stored thereon;
[0064] When the one or more programs are executed by the one or more processors, the one or more processors implement the method for generating a checklist based on a large model as described in any one of the first aspects.
[0065] A fourth aspect of the present application provides a computer storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for generating a checklist based on a large model as described in any one of the first aspects is implemented.
[0066] It can be seen from the above scheme that the present application provides a method and device for generating a checklist based on a large model, by acquiring flight data and flight mission information in real time; then, the flight data and flight mission information are input into the checklist generation model, and the checklist for this flight mission is output; wherein, the checklist generation model includes a standard checklist template; the checklist generation model is obtained by training a deep learning model with a training sample set; the checklist for this flight mission includes a pre-flight checklist, an in-flight checklist, and a post-flight checklist. The present application automatically generates checklists through a pre-trained model, and can adjust the checklists in real time according to flight data and flight mission information, thereby improving efficiency and accuracy while enhancing the dynamic adaptability of the checklists. BRIEF DESCRIPTION OF THE DRAWINGS
[0067] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying creative work.
[0068] Figure 1 A flowchart of a method for generating a checklist based on a large model provided in an embodiment of the present application;
[0069] Figure 2 A flowchart of a method for constructing a checklist generation model provided in another embodiment of the present application;
[0070] Figure 3An architecture diagram of data integration and preprocessing provided by another embodiment of the present application;
[0071] Figure 4 A flowchart of a method for training a checklist generation model provided in another embodiment of the present application;
[0072] Figure 5 A flowchart of a method for generating a checklist based on a large model provided in another embodiment of the present application;
[0073] Figure 6 A specific implementation flow chart of human-computer interaction and verification provided by another embodiment of the present application;
[0074] Figure 7 A flowchart of real-time monitoring of a checklist and updating of a standard checklist template provided by another embodiment of the present application;
[0075] Figure 8 A flow chart of a risk assessment provided for another embodiment of the present application;
[0076] Fig. 9 Another embodiment of the present application provides an architecture diagram of cross-platform compatibility and seamless integration;
[0077] Fig.10 Another embodiment of the present application provides a flowchart of a method for making personalized recommendations based on pilot habits;
[0078] Fig.11 A schematic diagram of a device for generating a checklist based on a large model provided by another embodiment of the present application;
[0079] Fig.12 Another embodiment of the present application provides a schematic diagram of an electronic device for implementing a method for generating a checklist based on a large model. DETAILED DESCRIPTION
[0080] The following will be combined with the drawings in the embodiments of the present application to clearly and completely describe the technical solutions in the embodiments of the present application. Obviously, the described embodiments are only part of the embodiments of the present application, not all of the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of this application.
[0081] The term "including" and its variations used herein are open inclusions, i.e., "including but not limited to". The term "based on" means "based at least in part on". The term "one embodiment" means "at least one embodiment"; the term "another embodiment" means "at least one additional embodiment"; the term "some embodiments" means "at least some embodiments". The relevant definitions of other terms will be given in the following description.
[0082] It is understandable that before using the technical solutions disclosed in the embodiments of the present disclosure, the types, scope of use, usage scenarios, etc. of the personal information involved in the present disclosure should be informed to the user and the user's authorization should be obtained in an appropriate manner in accordance with relevant laws and regulations.
[0083] It should be noted that the concepts such as "first" and "second" mentioned in this application are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.
[0084] It should be noted that the modifications of "one" and "plurality" mentioned in the present application are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".
[0085] The present application embodiment provides a method for generating a checklist based on a large model, such as Figure 1 As shown, the specific steps include:
[0086] S101. Acquire flight data and flight mission information in real time.
[0087] The flight mission information includes pilot information, airport information, route information, weather information and aircraft information. Flight data includes but is not limited to the aircraft's speed, altitude, range, flight time, etc., which are not limited here.
[0088] In the specific implementation process of this application, pilot information includes but is not limited to pilot experience (such as the pilot's total flight hours, flight hours for specific aircraft models, number of take-offs and landings, etc.), aircraft information includes but is not limited to aircraft maintenance history (aircraft maintenance records, extraction of maintenance frequency, the date of the most recent maintenance, whether a failure has occurred and its type, etc.), route information includes but is not limited to route characteristics (route distance, busyness, complexity of the passing area, etc.), weather information includes but is not limited to weather conditions (real-time weather data for the departure and destination airports, including wind speed, visibility, temperature, etc.), and is not limited here.
[0089] S102: Input the flight data and flight mission information into a checklist generation model, and output a checklist for this flight mission.
[0090] Among them, the checklist generation model includes a standard checklist template; the checklist generation model is obtained by training the deep learning model with a training sample set; the training sample set includes at least one training sample, and the training sample includes historical flight data and historical flight mission information as well as historical checklists for historical flight missions; the checklists for this flight mission include a pre-flight checklist, an in-flight checklist, and a post-flight checklist.
[0091] In the specific implementation of this application, the checklist is a checklist generated based on the current aircraft model, airport, weather, etc. to help the flight crew quickly count the items to ensure a smooth flight. This application helps users quickly identify and prepare all necessary inspection items through intelligent checklist generation, effectively prevents potential risks, fundamentally supports aviation safety, and provides a more solid guarantee for aviation safety.
[0092] Optionally, in another embodiment of the present application, each inspection item in the checklist of this flight mission is added with corresponding self-explanatory information, which helps the executor understand the logic and necessity behind the inspection and improves the execution efficiency and accuracy.
[0093] Optionally, in another embodiment of the present application, an implementation method of constructing a checklist generation model is as follows: Figure 2 As shown, including:
[0094] S201: Construct a training sample set.
[0095] The training sample set includes at least one training sample; the training sample includes historical flight data and historical flight mission information and historical checklists of historical flight missions.
[0096] In the specific implementation process of this application, data integration and preprocessing are required to construct a training sample set. Figure 3 As shown, data collection is first performed, which can be performed in but not limited to internal systems (aircraft operation systems, maintenance management systems, meteorological information systems, etc.), external data sources (air traffic control bureau, meteorological bureau, global positioning system, etc.) and aircraft data sources (black box, health management system, etc.), and then data cleaning is performed to remove invalid, duplicate or erroneous data, standardize the data format, ensure data quality, and finally perform feature extraction to obtain training samples.
[0097] S202: Input historical flight data and historical flight mission information into a deep learning model, and output predicted inspection information of historical flight missions.
[0098] In the specific implementation process of this application, deep learning models include but are not limited to recurrent neural networks (RNNs), Transformers, etc., which are not limited here.
[0099] When inputting data into the model, if the data cannot be fed into the model, adjust the data processing strategy according to the specific reasons:
[0100] 1. Data cleaning:
[0101] Check whether there are missing values, outliers, or data in the wrong format in the data, and handle them accordingly;
[0102] Standardize the data format to ensure that all data fields meet the requirements of the model.
[0103] 2. Data preprocessing:
[0104] Divide the data into training, validation and test sets;
[0105] Normalize or standardize numerical features to ensure that the features are on a similar scale.
[0106] 3. Data conversion:
[0107] If the data format does not meet the model requirements, try converting the data format, for example from text to vector representation.
[0108] 4. Exception handling:
[0109] Design exception handling mechanisms, such as retry logic and skip logic when data reading fails.
[0110] S203: Generate a prediction checklist for the historical flight mission based on the prediction checklist information of the historical flight mission and the standard checklist template.
[0111] Specifically, the generated predicted inspection information of the historical flight mission is filled into the standard checklist template, or the different parts between the standard checklist template and the generated predicted inspection information of the historical flight mission are replaced, and finally the predicted inspection list of the historical flight mission is generated, which is not limited here.
[0112] S204: Determine whether the error between the predicted checklist of the historical flight mission and the historical checklist of the historical flight mission meets a preset convergence condition.
[0113] The preset convergence condition is preset and changed by a technician or authorized staff and is not limited here.
[0114] It should be noted that, in the specific implementation process of the present application, it is not limited to using the preset convergence conditions to train the model, and the model can also be trained using a preset number of iterations, which is also not limited here.
[0115] Specifically, if it is determined that the error between the predicted checklist of the historical flight mission and the historical checklist of the historical flight mission meets the preset convergence condition, step S205 is executed; if it is determined that the error between the predicted checklist of the historical flight mission and the historical checklist of the historical flight mission does not meet the preset convergence condition, step S206 is executed.
[0116] S205. Use the deep learning model as a checklist generation model.
[0117] S206. Adjust the hyperparameters in the deep learning model using the error between the predicted checklist of the historical flight mission and the historical checklist of the historical flight mission.
[0118] like Figure 4 As shown, a more detailed flow chart of model training provided for another embodiment of the present application is first defined, and a deep learning model suitable for large-scale data processing is designed or selected, such as a model with a Transformer architecture, to utilize its powerful sequence processing capabilities and contextual understanding capabilities. Data preprocessing is then performed so that the data can be processed by the model, and then the preprocessed data set is input into the model, including historical checklist data, flight records, fault cases, safety regulations, etc., to train the model to learn the logical structure of the flight checklist, key check items and their dependencies. Finally, through continuous iteration and hyperparameter adjustment, the model accuracy and generalization ability are improved to ensure that the generated checklist is both compliant with the specification and targeted.
[0119] When data cannot be fed into the model, the data processing strategy is adjusted. After the model training starts, forward propagation and loss calculation are performed, and then backpropagation and weight update are performed to evaluate the model performance and determine whether the model meets the stopping conditions. If not, the parameters are adjusted to continue training. If so, the model optimization is completed, the model is evaluated, and finally the model is applied and monitored while it is being applied.
[0120] It should be noted that forward propagation refers to the flow of data from the input layer to the output layer, generating predictions. Calculating loss refers to evaluating the difference between the prediction and the true label through the loss function. Backward propagation refers to calculating the gradient of the loss function with respect to each weight. Updating weights refers to adjusting the model parameters based on the calculated gradients. Evaluating the model refers to evaluating the performance of the model using the validation set and the test set. Stopping conditions are usually determined based on the number of iterations or when the performance reaches a preset standard.
[0121] Specifically, if the stopping condition is not met, adjust the hyperparameters according to the following:
[0122] (1) Performance of validation set / test set:
[0123] Accuracy / Error Rate: The most direct metric used to measure the performance of a model on unseen data.
[0124] Confusion Matrix: Provides a more detailed view of the model’s performance, including the number of true positives (TP), false positives (FP), true negatives (TN), and false negatives (FN).
[0125] ROC curve and AUC value: For classification problems, the ROC curve shows the relationship between the true positive rate and the false positive rate at different thresholds, and the AUC value quantifies the area under the ROC curve, which is an indicator for evaluating the overall performance of the classifier.
[0126] Loss value: Although the loss is mainly calculated during the training process, the loss value on the validation set can also reflect the generalization ability of the model.
[0127] (2) Learning rate:
[0128] If the loss on the validation set stops decreasing, it may be that the learning rate is too low, resulting in slow convergence; or the learning rate is too high, causing the model to oscillate around the optimal solution.
[0129] You can try using a learning rate decay strategy, such as fixed step size decay, exponential decay, or an adaptive learning rate algorithm (such as Adam).
[0130] like Figure 5 As shown, it is an overall flow chart of a method for generating a checklist based on a large model provided by an embodiment of the present application. Flight data and mission details (flight mission information) are received in real time, and then the flight data and flight mission information are input into the checklist generation model. Through model analysis and the standard checklist template in the model, the checklist for this flight mission (pre-flight checklist, in-flight checklist and post-flight checklist) is output, and the flight data is analyzed in real time during the flight, and the checklist is dynamically adjusted. During the flight, various special events and emergencies (such as emergency situations, equipment failures, and unconventional weather) may occur, and the aircraft system will generate corresponding emergency instructions. In another embodiment of the present application, after receiving the emergency instruction, the real-time acquisition of flight data and flight mission information is immediately executed, and the flight data and flight mission information are input into the checklist generation model, and the checklist steps for this flight mission are output to obtain the emergency checklist. Relevant inspection items are added or emphasized to ensure flight safety.
[0131] Optionally, in another embodiment of the present application, Figure 6 As shown in the figure, it is a specific implementation flow chart of human-computer interaction and verification. First, an easy-to-use interface is developed to allow pilots and maintenance personnel to view and confirm the automatically generated checklists through electronic devices (such as tablets and electronic flight bags). Then, manual review is performed. If the user's review information indicates that no adjustment is required, no processing is performed; if the user's review information indicates that there is an adjustment, the user's feedback information and actual adjustment information are collected; the checklist generation model is optimized based on the user's feedback information and actual adjustment information to obtain the optimized checklist generation model.
[0132] Optionally, in another embodiment of the present application, Figure 7As shown in the figure, the real-time monitoring of checklists and the updating process of standard checklist templates monitor the execution of flight data and checklists in real time; if any problem (abnormality) is found, the problem is detected and abnormal feedback information is generated, and the model is automatically updated regularly to automatically update the standard checklist template in the checklist generation model according to the latest aviation rules, safety standards, and aircraft upgrade information.
[0133] Optionally, in another embodiment of the present application, Figure 8 As shown, risk assessment is integrated into the checklist generation process. Based on historical flight record data, route information, flight data and external intelligence data, the risk assessment results of this flight mission are generated to determine the potential safety risks of the flight mission.
[0134] Among them, historical flight record data includes but is not limited to historical accident reports, mechanical failures, etc., which are not limited here; external intelligence data at least includes geopolitical tensions, terrorist threat levels, air traffic density, severe weather forecasts, etc., which are not limited here. Flight data includes but is not limited to aircraft status (such as engine temperature, fuel volume), etc., which are not limited here; route information includes but is not limited to route conditions (such as wind speed, turbulence), air traffic control instructions, etc., which are not limited here.
[0135] In the specific implementation process of this application, a risk management model can be pre-trained, but not limited to, to generate the risk assessment results of this flight mission based on historical flight record data, route information, flight data and external intelligence data. The risk management model first needs to ensure that there are sufficient data sources, which may include: historical flight record data, route information, flight data, external intelligence data, etc. Then, data processing and integration are performed, including: data cleaning: removing invalid or erroneous data. Data standardization: ensuring that the data format from different sources is consistent. After that, the risk management model is constructed, and when selecting the model: select the appropriate machine learning or statistical model according to the characteristics of the problem. For example, a classification model can be used to predict a specific type of flight risk, or a regression model can be used to estimate the degree of risk. Then, the model is trained using historical data. Finally, the model performance is evaluated using an independent validation set. Finally, a trained risk management model is used for real-time risk assessment: real-time monitoring data and the latest external intelligence are continuously received; the risk assessment results are dynamically adjusted according to the new data, and a risk score is assigned to each flight mission.
[0136] In the specific implementation process of this application, if Figure 8 As shown, the items in the checklist can also be prioritized according to risk level to ensure that high-risk items are checked and confirmed first, thereby improving the efficiency and pertinence of pre-flight preparations.
[0137] Specifically, first, you need to define the risk level standard. This is usually based on historical data, industry best practices, and expert opinions. For example, four risk levels can be defined: low risk (L), medium-low risk (ML), medium-high risk (MH), and high risk (H); each risk level can correspond to different risk standards or impact ranges, and different inspection items can correspond to the checklist. Once the risk level of each inspection is determined, you can use an intelligent sorting algorithm to arrange the items on the checklist. Common sorting algorithms include: direct sorting: directly sorting by risk level; weighted sorting: assigning different weights to different risk levels, and then sorting by weighted scores.
[0138] It is understandable that in the specific implementation process of this application, although the degree of automation is high, the manual review step is retained, allowing professionals to adjust the risk assessment results according to actual conditions, and adjust the checklist according to the risk assessment results to obtain an optimized checklist.
[0139] Optionally, in another embodiment of the present application, natural language processing (NLP) can also be applied to enable the model to understand natural language queries or instructions from pilots and maintenance personnel. For example, by inputting specific requirements through voice or text ("additional checks for thunderstorms"), the model can instantly generate corresponding checklists.
[0140] In the specific implementation process of this application, if Fig. 9 As shown, it also supports multiple terminals, ensuring that the checklist generation system can run across multiple platforms, including desktop, mobile (iOS / Android), and aviation-specific electronic equipment, with a friendly interface and easy operation. The system is also integrated to seamlessly connect with the airline's existing IT systems (such as flight planning systems, maintenance management systems, and flight management systems) to achieve automatic data synchronization, reduce human input errors, and improve the overall operational synergy.
[0141] Optionally, in another embodiment of the present application, after obtaining the flight data and flight mission information in real time, the pilot's personality data is also obtained, wherein the pilot's personality data includes: flight experience, flight records, the pilot's pre-set preference settings and the pilot's feedback information; a user profile of the pilot is constructed based on the pilot's personality data; then, the pilot's user profile, flight data and flight mission information are input into a checklist generation model, and a checklist for this flight mission is output, so that the generated checklist is more in line with the pilot's usage pattern and preferences, and provides them with more personalized checklist suggestions.
[0142] Specifically, Fig.10As shown, the model training method can be used to learn and pre-adapt pilot work habits, and generate personalized checklists based on the knowledge graph (the knowledge graph includes the knowledge graph of aircraft models, system components, common faults and their solutions. The model generates more specific and targeted inspection suggestions based on this knowledge graph to improve the efficiency of problem solving). First, the pilot's behavior data needs to be collected, including but not limited to: Flight experience: the pilot's flight hours, aircraft type, etc. Flight records: the completion of the pre-flight checklist, the order of common inspection items, skipped inspection items, etc. Personalization settings: the pilot's specific preference settings in the pre-flight preparation stage. Feedback information: the pilot's feedback on the checklist, such as modification suggestions for certain items. External environment: special conditions of the flight mission, such as weather, route characteristics, etc. Then, user portraits are constructed to analyze the pilot's behavior patterns, such as which items they usually check first, and whether these behaviors change with time and experience, and through cluster analysis and other means, the preference differences between different pilot groups are identified. After that, the collected data is used to train the model so that it can learn the pilot's behavior patterns and preferences. Finally, the pilot’s user profile, flight data, and flight mission information are input into the checklist generation model, and the checklist for this flight mission is output.
[0143] Of course, in the specific implementation process of this application, security and privacy protection are also considered, through data encryption: ensure that in the entire data processing process, all sensitive information is protected by industry-standard encryption technology to prevent data leakage. And privacy compliance: follow international privacy protection regulations such as GDPR and HIPAA, anonymize information involving personal privacy, and ensure that the use of the model is legal and compliant. Through the above expansion plan, the aviation industry data generation checklist system based on large models can not only improve aviation safety and operational efficiency, but also provide a more humane and intelligent service experience, and contribute to the digital transformation of the aviation industry.
[0144] It can be seen from the above scheme that the present application provides a method for generating a checklist based on a large model, by acquiring flight data and flight mission information in real time; then, the flight data and flight mission information are input into the checklist generation model, and the checklist for this flight mission is output; wherein, the checklist generation model includes a standard checklist template; the checklist generation model is obtained by training a deep learning model with a training sample set; the checklist for this flight mission includes a pre-flight checklist, an in-flight checklist, and a post-flight checklist. The present application automatically generates checklists through a pre-trained model, and can adjust the checklists in real time according to flight data and flight mission information, thereby improving efficiency and accuracy while enhancing the dynamic adaptability of the checklists.
[0145] The embodiment of the present application provides a device for generating a checklist based on a large model, such as Fig.11 As shown, specifically including:
[0146] The first acquisition unit 1101 is used to acquire flight data and flight mission information in real time.
[0147] The flight data and the flight mission information include pilot information, airport information, route information, weather information and aircraft information.
[0148] The first input unit 1102 is used to input the flight data and the flight mission information into the checklist generation model, and output a checklist for the current flight mission.
[0149] Among them, the checklist generation model includes a standard checklist template; the checklist generation model is obtained by training a deep learning model with a training sample set; the training sample set includes at least one training sample, and the training sample includes historical flight data and historical flight mission information as well as historical checklists for historical flight missions; the checklist for this flight mission includes a pre-flight checklist, an in-flight checklist, and a post-flight checklist.
[0150] For the specific working process of the units disclosed in the above embodiments of the present application, please refer to the corresponding method embodiments, such as Figure 1 As shown, no further description is given here.
[0151] Optionally, in another embodiment of the present application, an implementation of a construction unit of a checklist generation model includes:
[0152] The training sample set construction unit is used to construct the training sample set.
[0153] The training sample set includes at least one training sample; the training sample includes historical flight data and historical flight mission information and historical checklists of historical flight missions.
[0154] The second input unit is used to input the historical flight data and the historical flight mission information into the deep learning model, and output the predicted inspection information of the historical flight mission.
[0155] A generating unit is used to generate a predicted checklist for the historical flight mission according to the predicted checklist information of the historical flight mission and the standard checklist template.
[0156] An adjustment unit is used to adjust the hyperparameters in the deep learning model by using the error between the predicted checklist of the historical flight mission and the historical checklist of the historical flight mission, until the error between the predicted checklist of the historical flight mission and the historical checklist of the historical flight mission meets a preset convergence condition, and the deep learning model is used as a checklist generation model.
[0157] For the specific working process of the units disclosed in the above embodiments of the present application, please refer to the corresponding method embodiments, such as Figure 2 As shown, no further description is given here.
[0158] Optionally, in another embodiment of the present application, an implementation of the device for generating a checklist based on a large model further includes:
[0159] The emergency processing unit is used to receive the emergency instruction, immediately execute the real-time acquisition of flight data and flight mission information, input the flight data and the flight mission information into the checklist generation model, output the checklist steps of this flight mission, and obtain the emergency checklist.
[0160] The specific working process of the units disclosed in the above embodiments of the present application can be found in the corresponding method embodiments, which will not be repeated here.
[0161] Optionally, in another embodiment of the present application, an implementation of the device for generating a checklist based on a large model further includes:
[0162] The first receiving unit is used to receive a checklist query instruction and display the checklist for this flight mission.
[0163] The second receiving unit is used to receive the user's review information.
[0164] Among them, if the audit information of the user indicates that no adjustment is required, no processing will be performed.
[0165] The collecting unit is used to collect the user's feedback information and actual adjustment information if the user's review information indicates that there is an adjustment.
[0166] The optimization unit is used to optimize the checklist generation model according to the user's feedback information and the actual adjustment information to obtain an optimized checklist generation model.
[0167] The specific working process of the units disclosed in the above embodiments of the present application can be found in the corresponding method embodiments, which will not be repeated here.
[0168] Optionally, in another embodiment of the present application, an implementation of the device for generating a checklist based on a large model further includes:
[0169] The monitoring unit is used to monitor the flight data and the execution of the checklist in real time.
[0170] The abnormality feedback unit is used to generate abnormality feedback information if the checklist is not executed or the flight data is abnormal.
[0171] The specific working process of the units disclosed in the above embodiments of the present application can be found in the corresponding method embodiments, which will not be repeated here.
[0172] Optionally, in another embodiment of the present application, an implementation of the device for generating a checklist based on a large model further includes:
[0173] The risk assessment unit is used to generate a risk assessment result of the current flight mission based on historical flight record data, the route information, the flight data and external intelligence data.
[0174] The historical flight record data includes at least historical accident reports and mechanical failures; the external intelligence data includes at least controlled traffic density and severe weather forecasts.
[0175] The specific working process of the units disclosed in the above embodiments of the present application can be found in the corresponding method embodiments, which will not be repeated here.
[0176] Optionally, in another embodiment of the present application, an implementation of the device for generating a checklist based on a large model further includes:
[0177] The second acquisition unit is used to acquire the pilot's personality data.
[0178] The pilot's personality data includes: flight experience, flight records, pilot's pre-set preferences and pilot's feedback information.
[0179] A portrait construction unit is used to construct a user portrait of the pilot according to the personality data of the pilot.
[0180] Wherein, the first input unit is used to:
[0181] The pilot's user portrait, the flight data, and the flight mission information are input into a checklist generation model, and a checklist for this flight mission is output.
[0182] The specific working process of the units disclosed in the above embodiments of the present application can be found in the corresponding method embodiments, which will not be repeated here.
[0183] Optionally, in another embodiment of the present application, each check item in the checklist of the flight mission is added with corresponding self-explanatory information.
[0184] It can be seen from the above scheme that the present application provides a checklist generation device based on a large model, which acquires flight data and flight mission information in real time through the first acquisition unit 1101; then, the first input unit 1102 inputs the flight data and flight mission information into the checklist generation model, and outputs the checklist for this flight mission; wherein, the checklist generation model includes a standard checklist template; the checklist generation model is obtained by training a deep learning model with a training sample set; the checklist for this flight mission includes a pre-flight checklist, an in-flight checklist, and a post-flight checklist. The present application automatically generates checklists through a pre-trained model, and can adjust the checklists in real time according to flight data and flight mission information, thereby improving efficiency and accuracy while enhancing the dynamic adaptability of the checklists.
[0185] The functions described above herein may be performed at least in part by one or more hardware logic components. For example, without limitation, exemplary types of hardware logic components that may be used include: field programmable gate arrays (FPGAs), application specific integrated circuits (ASICs), application specific standard products (ASSPs), systems on chips (SOCs), complex programmable logic devices (CPLDs), and the like.
[0186] Another embodiment of the present application provides an electronic device, such as Fig.12 As shown, including:
[0187] One or more processors 1201.
[0188] The storage device 1202 stores one or more programs.
[0189] When the one or more programs are executed by the one or more processors 1201, the one or more processors 1201 implement the method for generating a checklist based on a large model as described in any one of the above embodiments.
[0190] Another embodiment of the present application provides a computer storage medium having a computer program stored thereon, wherein when the computer program is executed by a processor, the method for generating a checklist based on a large model as described in any one of the above embodiments is implemented.
[0191] In the context of the present application, a machine-readable medium may be a tangible medium that may contain or store a program for use by or in conjunction with an instruction execution system, device, or equipment. A machine-readable medium may be a machine-readable signal medium or a machine-readable storage medium. A machine-readable medium may include, but is not limited to, an electronic, magnetic, optical, electromagnetic, infrared, or semiconductor system, device, or equipment, or any suitable combination of the foregoing. A more specific example of a machine-readable storage medium may include an electrical connection based on one or more lines, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the foregoing.
[0192] It should be noted that the computer-readable medium mentioned above in the present application may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present application, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present application, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, in which a computer-readable program code is carried. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. The computer readable signal medium may also be any computer readable medium other than a computer readable storage medium, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any suitable medium, including but not limited to: wires, optical cables, RF (radio frequency), etc., or any suitable combination of the above.
[0193] The computer-readable medium may be included in the electronic device, or may exist independently without being incorporated into the electronic device.
[0194] Another embodiment of the present application provides a computer program product. When the computer program product is executed, it is used to execute any of the above-mentioned methods for generating a checklist based on a large model.
[0195] In particular, according to an embodiment of the present application, the process described above with reference to the flowchart can be implemented as a computer software program. For example, an embodiment of the present application includes a computer program product, which includes a computer program carried on a non-transitory computer-readable medium, and the computer program includes a program code for executing the method shown in the flowchart. In such an embodiment, the computer program can be downloaded and installed from a network via a communication device, or installed from a storage device, or installed from a ROM. When the computer program is executed by a processing device, the above-mentioned functions defined in the method of the embodiment of the present application are executed.
[0196] Although the subject matter has been described in language specific to structural features and / or methodological logical actions, it should be understood that the subject matter defined in the appended claims is not necessarily limited to the specific features or actions described above. On the contrary, the specific features and actions described above are merely example forms of implementing the claims.
[0197] Although several specific implementation details are included in the above discussion, these should not be interpreted as limiting the scope of the present application. Certain features described in the context of a separate embodiment can also be implemented in a single embodiment in combination. On the contrary, the various features described in the context of a single embodiment can also be implemented in multiple embodiments individually or in any suitable sub-combination.
[0198] The above description is only a preferred embodiment of the present application and an explanation of the technical principles used. Those skilled in the art should understand that the scope of application involved in the present application is not limited to the technical solution formed by a specific combination of the above technical features, but should also cover other technical solutions formed by any combination of the above technical features or their equivalent features without departing from the above application concept. For example, the above features are replaced with (but not limited to) technical features with similar functions applied in the present application.
Claims
1. A method for generating a checklist based on a large model, characterized in that: include: Acquire flight data and flight mission information in real time; wherein the flight data and flight mission information include pilot information, airport information, route information, weather information and aircraft information; The flight data and the flight mission information are input into a checklist generation model, and a checklist for this flight mission is output; wherein the checklist generation model includes a standard checklist template; the checklist generation model is obtained by training a deep learning model with a training sample set; the training sample set includes at least one training sample, and the training sample includes historical flight data and historical flight mission information as well as historical checklists for historical flight missions; the checklist for this flight mission includes a pre-flight checklist, an in-flight checklist, and a post-flight checklist.
2. The method for generating a checklist based on a large model according to claim 1, characterized in that: The method for constructing the checklist generation model comprises: Constructing a training sample set; wherein the training sample set includes at least one training sample; the training sample includes historical flight data and historical flight mission information and historical checklists of historical flight missions; Inputting the historical flight data and the historical flight mission information into a deep learning model, and outputting predicted inspection information of the historical flight mission; generating a predicted checklist for the historical flight mission according to the predicted checklist information of the historical flight mission and the standard checklist template; The hyperparameters in the deep learning model are adjusted using the error between the predicted checklist of the historical flight mission and the historical checklist of the historical flight mission until the error between the predicted checklist of the historical flight mission and the historical checklist of the historical flight mission meets a preset convergence condition, and the deep learning model is used as a checklist generation model.
3. The method for generating a checklist based on a large model according to claim 1, characterized in that: After the flight data and the flight mission information are input into the checklist generation model and the checklist for the current flight mission is output, the method further includes: Receive the emergency instruction, immediately execute the real-time acquisition of flight data and flight mission information, input the flight data and the flight mission information into the checklist generation model, output the checklist steps of this flight mission, and obtain the emergency checklist.
4. The method for generating a checklist based on a large model according to claim 1, characterized in that: After the flight data and the flight mission information are input into the checklist generation model and the checklist for the current flight mission is output, the method further includes: Receive checklist query instructions and display the checklist for this flight mission; Receive user's review information; If the user's audit information indicates that no adjustment is required, no action will be taken; If the user's review information indicates that there is an adjustment, collecting the user's feedback information and actual adjustment information; The checklist generation model is optimized according to the user's feedback information and the actual adjustment information to obtain an optimized checklist generation model.
5. The method for generating a checklist based on a large model according to claim 1, characterized in that: After the flight data and the flight mission information are input into the checklist generation model and the checklist for the current flight mission is output, the method further includes: Real-time monitoring of flight data and checklist execution; If the checklist is not followed or the flight data is abnormal, an abnormal feedback message will be generated.
6. The method for generating a checklist based on a large model according to claim 1, characterized in that: Also includes: Regularly and automatically update the standard checklist templates in the checklist generation model based on the latest aviation regulations, safety standards, and aircraft upgrade information.
7. The method for generating a checklist based on a large model according to claim 1, characterized in that: After the real-time acquisition of flight data and flight mission information, the method further includes: A risk assessment result of the flight mission is generated based on the historical flight record data, the route information, the flight data and the external intelligence data; wherein the historical flight record data at least includes historical accident reports and mechanical failures; and the external intelligence data at least includes controlled traffic density and severe weather forecasts.
8. The method for generating a checklist based on a large model according to claim 1, characterized in that: After the real-time acquisition of flight data and flight mission information, the method further includes: Acquiring the pilot's personality data; wherein the pilot's personality data includes: flight experience, flight records, pilot's pre-set preferences, and pilot's feedback information; constructing a user profile of the pilot according to the personality data of the pilot; The step of inputting the flight data and the flight mission information into a checklist generation model and outputting a checklist for the flight mission includes: The pilot's user portrait, the flight data, and the flight mission information are input into a checklist generation model, and a checklist for this flight mission is output.
9. The method for generating a checklist based on a large model according to claim 1, characterized in that: Each check item in the checklist of the flight mission is added with corresponding self-explanatory information.
10. A device for generating a checklist based on a large model, characterized in that: include: A first acquisition unit is used to acquire flight data and flight mission information in real time; wherein the flight data and flight mission information include pilot information, airport information, route information, weather information and aircraft information; A first input unit is used to input the flight data and the flight mission information into a checklist generation model, and output a checklist for this flight mission; wherein the checklist generation model includes a standard checklist template; the checklist generation model is obtained by training a deep learning model with a training sample set; the training sample set includes at least one training sample, and the training sample includes historical flight data and historical flight mission information as well as historical checklists for historical flight missions; the checklist for this flight mission includes a pre-flight checklist, an in-flight checklist, and a post-flight checklist.
Citation Information
Cited By
Aircraft inspection control method, electronic equipment and aircraft
CN120220476A
