Flight training evaluation method, system, and electronic device

Abstract

The application provides a flight training evaluation method, system and electronic equipment, the method comprising: acquiring multi-modal original flight training data in the flight training process; preprocessing the original flight training data to obtain flight training data; inputting the flight training data into an evaluation model, performing feature processing on the flight training data by the evaluation model to obtain a plurality of feature data links, identifying a plurality of abnormal data links with deviations from the plurality of feature data links, evaluating each abnormal data link, and outputting an optimal supplementary training scheme based on the evaluation result; and the evaluation model is obtained by training based on multi-modal flight training data samples and corresponding optimal supplementary training scheme labels. By using the evaluation model, the application identifies and evaluates deviations based on multi-modal data to generate an optimal supplementary training scheme, improves the efficiency, accuracy and interpretability of flight training evaluation, and realizes a training closed loop.

Description

Technical Field

[0001] This invention relates to the field of flight training technology, and in particular to a flight training assessment method, system, and electronic device. Background Technology

[0002] As modern aviation training systems become increasingly complex, flight training and mission execution typically encompass multiple stages, including syllabus development, mission planning, live-fire operations, and debriefing. Currently, training consistency assessment relies primarily on manual comparison, which is inefficient, unable to effectively process cross-modal information, provides coarse-grained deviation identification, fails to accurately pinpoint the root cause of problems, affects the accuracy and interpretability of the assessment, lacks a corrective mechanism, and makes it difficult to achieve a closed-loop training process. Summary of the Invention

[0003] This invention provides a flight training assessment method, system, and electronic device to address the shortcomings of existing flight training assessment methods, such as low efficiency, poor accuracy and interpretability, and lack of correction mechanisms.

[0004] This invention provides a flight training assessment method, comprising: Acquire raw multimodal flight training data during the flight training process; The original multimodal flight training data is preprocessed to obtain multimodal flight training data; The multimodal flight training data is input into the evaluation model, which performs feature processing on the multimodal flight training data to obtain multiple feature data links. Multiple abnormal data links with deviations are identified from the multiple feature data links. Each abnormal data link is evaluated to obtain an evaluation result. Based on the evaluation result, the optimal supplementary training scheme is output. The evaluation model is trained based on multimodal flight training data samples and the optimal supplementary training scheme labels corresponding to the multimodal flight training data samples.

[0005] In some embodiments, the multimodal flight training data includes: training target data, execution data, and cognitive summary data.

[0006] In some embodiments, the evaluation model includes: A semantic alignment layer is used to extract and match features from the multimodal flight training data to obtain multiple feature data links. A deviation identification layer is used to identify deviations in multiple feature data links and obtain identification results, the identification results including multiple abnormal data links; An evaluation layer is used to quantitatively evaluate multiple abnormal data links based on the identification results to obtain evaluation results; The decision-making layer is used to determine the optimal supplementary training scheme based on the evaluation results.

[0007] In some embodiments, the step of performing feature extraction and feature matching on the multimodal flight training data to obtain multiple feature data links includes: Feature extraction is performed on the training target data to obtain multiple training target feature data; Feature extraction is performed on the execution data to obtain multiple execution feature data; Feature extraction is performed on the cognitive summary data to obtain multiple cognitive summary feature data; Matching multiple training target feature data, multiple execution feature data, and multiple cognitive summary feature data yields a matching result. Based on the matching results, multiple feature data links are obtained; the feature data links use the training target feature data, the execution feature data, and the cognitive summary feature data as data nodes.

[0008] In some embodiments, the quantitative evaluation of multiple abnormal data links based on the identification results includes: Based on the identification results and preset thresholds, the severity level of deviation for each abnormal data link is determined; The consistency of the abnormal data nodes in each abnormal data link is scored to obtain the consistency score of the abnormal data nodes. Based on the weight of the abnormal data node and the consistency score of the abnormal data node, the consistency score of each abnormal data link is calculated.

[0009] In some embodiments, determining the optimal supplementary training scheme based on the evaluation results includes: Under preset constraints, multiple candidate supplementary training schemes are generated based on the evaluation results; With the goal of minimizing flight training costs, the optimal supplementary training scheme is determined from multiple candidate supplementary training schemes.

[0010] In some embodiments, the method further includes: The evaluation results are then visualized. The evaluation results and the optimal supplementary training scheme are stored.

[0011] In some embodiments, the evaluation model is trained based on the following steps: Obtain raw flight training data samples in multiple modes; The original multimodal flight training data samples are preprocessed to obtain multimodal flight training data samples. Determine the optimal supplementary training scheme label corresponding to the multimodal flight training data sample; Using the multimodal flight training data samples as training samples and the optimal supplementary training scheme labels as sample labels, an initial evaluation model is trained, and after training, the evaluation model is obtained.

[0012] The present invention also provides a flight training assessment system, comprising: The acquisition unit is used to acquire raw multimodal flight training data during the flight training process; A preprocessing unit is used to preprocess the original multimodal flight training data to obtain multimodal flight training data. An evaluation unit is used to input the multimodal flight training data into an evaluation model, whereby the evaluation model performs feature processing on the multimodal flight training data to obtain multiple feature data links, identifies multiple abnormal data links with deviations from the multiple feature data links, evaluates each of the abnormal data links to obtain an evaluation result, and outputs the optimal supplementary training scheme based on the evaluation result. The evaluation model is trained based on multimodal flight training data samples and the optimal supplementary training scheme labels corresponding to the multimodal flight training data samples.

[0013] The present invention also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the flight training evaluation method as described above.

[0014] The flight training evaluation method, system, and electronic device provided by this invention acquire multimodal raw flight training data during the flight training process; preprocess the raw flight training data to obtain flight training data; input the flight training data into an evaluation model, which performs feature processing on the multimodal flight training data to obtain multiple feature data links; identify multiple abnormal data links with deviations from the multiple feature data links; evaluate each abnormal data link to obtain an evaluation result; and output the optimal supplementary training scheme based on the evaluation result. This improves the efficiency, accuracy, and interpretability of flight training evaluation and realizes a training closed loop. Attached Figure Description

[0015] To more clearly illustrate the technical solutions in this invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of this invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0016] Figure 1 This is a flowchart illustrating the flight training evaluation method provided in this embodiment of the invention.

[0017] Figure 2 This is a flowchart illustrating the training process of the evaluation model provided in this embodiment of the invention.

[0018] Figure 3 This is a schematic diagram of the flight training and evaluation system provided in an embodiment of the present invention.

[0019] Figure 4 This is a schematic diagram of the structure of the electronic device provided in an embodiment of the present invention. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this invention. All other embodiments obtained by those skilled in the art based on the embodiments of this invention without creative effort are within the scope of protection of this invention.

[0021] Figure 1 This is a flowchart illustrating the flight training evaluation method provided in an embodiment of the present invention, as shown below. Figure 1 As shown, the method includes: Step 110: Obtain raw multimodal flight training data during the flight training process.

[0022] Optionally, the raw flight training data for the multimodal modes includes at least: text modal data, numerical modal data, image modal data, and behavioral / cognitive modal data.

[0023] Among them, text modal data are training specifications, plans, and summaries in the form of structured or unstructured text; numerical modal data are time-series parameters and state data obtained from aircraft sensors, airborne recording equipment, or simulators, such as flight parameters and system parameters; video modal data are video and image data recorded by airborne camera equipment, which can intuitively present the flight environment and operation scenarios; and behavioral / cognitive modal data are subjective summaries and evaluations expressed by training participants through natural language or structured reports.

[0024] In some embodiments, multimodal flight training data includes: training objective data, execution data, and cognitive summary data.

[0025] Optionally, the training objective data shall include at least the training outline and training plan; the execution data shall include at least flight parameters, sensor data and cockpit / head-up display image data; and the cognitive summary data shall include at least the trainee summary and instructor comments.

[0026] The training syllabus is a standardized textual requirement for flight training courses, such as maneuver standards and weather conditions; the training plan is a specific task planning text for flight training, such as the order of courses, airspace allocation, and time nodes; the training syllabus and training plan serve as target benchmarks for training compliance and are used for subsequent comparison with execution data.

[0027] Step 120: Preprocess the raw multimodal flight training data to obtain multimodal flight training data.

[0028] Optionally, the raw flight training data may be preprocessed, including data cleaning, standardization, normalization, data filling, and spatiotemporal alignment.

[0029] For example, preprocessing of text modal data includes sentence segmentation, terminology normalization, and semantic segmentation; preprocessing of numerical modal data includes filtering, normalization, and data imputation; time synchronization preprocessing of image modal data; and role labeling of behavioral / cognitive modal data.

[0030] Step 130: Input the multimodal flight training data into the evaluation model. The evaluation model performs feature processing on the multimodal flight training data to obtain multiple feature data links. From the multiple feature data links, identify multiple abnormal data links with deviations. Evaluate each abnormal data link to obtain the evaluation result. Based on the evaluation result, output the optimal supplementary training scheme.

[0031] The evaluation model is trained based on multimodal flight training data samples and the optimal supplementary training scheme labels corresponding to the multimodal flight training data samples.

[0032] It should be noted that the evaluation model possesses cross-modal semantic understanding, spatiotemporal logical reasoning, causal inference, and decision generation capabilities.

[0033] Optionally, the evaluation model aims to minimize training costs and maximize remedial effects, and outputs the optimal supplementary training plan. The optimal supplementary training plan includes, but is not limited to, the subjects to be supplemented, specific training conditions, training methods, key practice points, and the expected duration.

[0034] Optionally, the feature data link includes a target node, an execution node, and a cognitive node; if there is a significant inconsistency among the target node, execution node, and cognitive node in the feature data link, then the feature data link is an abnormal data link.

[0035] In some embodiments, the evaluation model includes: The semantic alignment layer is used to extract and match features from multimodal flight training data to obtain multiple feature data links. The deviation identification layer is used to identify deviations in multiple feature data links and obtain identification results, which include multiple abnormal data links. The evaluation layer is used to quantitatively evaluate multiple abnormal data links based on the identification results, and obtain the evaluation results. The decision-making layer is used to determine the optimal supplementary training plan based on the evaluation results.

[0036] Optionally, the identification results may include at least data such as abnormal data links, abnormal nodes, abnormal types, and abnormal evidence.

[0037] Optionally, the assessment results may include at least the deviation quantification value, root cause inference results, severity level, and relevant multimodal evidence index.

[0038] The semantic alignment layer output is a structured, semantically related feature data link.

[0039] Optionally, the deviation identification layer can detect inconsistencies in the feature data links; the evaluation layer can quantify the severity of the identified anomalies, perform root cause analysis, and assess their impact; the decision layer can analyze the evaluation results of all anomalous data links, cluster anomalous data links with the same or similar root causes, and determine the priority of anomalous data link repair.

[0040] In this embodiment of the invention, multimodal raw flight training data is acquired during flight training; the raw flight training data is preprocessed to obtain flight training data; the flight training data is input into an evaluation model, which performs feature processing on the multimodal flight training data to obtain multiple feature data links; multiple abnormal data links with deviations are identified from the multiple feature data links; each abnormal data link is evaluated to obtain an evaluation result; based on the evaluation result, the optimal supplementary training scheme is output, thereby improving the efficiency, accuracy, and interpretability of flight training evaluation and realizing a training closed loop.

[0041] In some embodiments, feature extraction and feature matching are performed on multimodal flight training data to obtain multiple feature data links, including: Feature extraction is performed on the training target data to obtain multiple training target feature data; Feature extraction is performed on the execution data to obtain multiple execution feature data. Feature extraction is performed on the cognitive summary data to obtain multiple cognitive summary feature data; Matching is performed on multiple training target feature data, multiple execution feature data, and multiple cognitive summary feature data to obtain matching results; Based on the matching results, multiple feature data links are obtained; the feature data links use training target feature data, execution feature data, and cognitive summary feature data as data nodes.

[0042] Optionally, the training target feature data may include at least key actions, parameter standards, conditional constraints, success criteria, and other feature data.

[0043] Optionally, time-series features can be extracted from the numerical data in the execution data, such as calculating the average, variance, extreme values, and trend slope of a specific stage, and detecting key event points; features can be extracted from the image data in the execution data to identify key targets and analyze the pilot's line of sight or control actions; and features such as self-evaluation, key decision points, and difficulties can be extracted from the cognitive summary data using techniques such as sentiment analysis, intent recognition, and key information extraction.

[0044] Optionally, spatiotemporal alignment matching and semantic similarity matching can be performed on multiple training target feature data, multiple execution feature data, and multiple cognitive summary feature data.

[0045] Specifically, by using a unified timestamp, feature data that occur within the same time period are initially associated; by using vector similarity calculation, feature data describing the same event or concept are further searched based on spatiotemporal alignment.

[0046] Optionally, based on the matching results, the three types of feature data nodes that are successfully matched are combined in the order of goal-execution-cognition around specific training and evaluation points to form a complete and traceable feature data link.

[0047] In this embodiment of the invention, matching results are obtained by matching multiple training target feature data, multiple execution feature data, and multiple cognitive summary feature data; based on the matching results, multiple feature data links are obtained, realizing the structured and semantic association of cross-modal data, laying the foundation for refined and hierarchical deviation identification.

[0048] In some embodiments, based on the identification results, a quantitative evaluation of multiple anomalous data links is performed, including: Based on the identification results and preset thresholds, the severity level of deviation for each abnormal data link is determined; The consistency of the abnormal data nodes in each abnormal data link is scored to obtain the consistency score of the abnormal data nodes; Based on the weights and consistency scores of the abnormal data nodes, the consistency score of each abnormal data link is calculated.

[0049] Optionally, the severity level of the deviation can be slight, moderate, or severe.

[0050] Optionally, based on the identification results of abnormal data links and with reference to pre-set grading standards related to flight safety and the importance of training objectives, the severity level of deviation for each abnormal data link is determined, providing a priority ranking basis for subsequent decisions and ensuring that the most serious problems are addressed first.

[0051] Optionally, within each anomalous data link, the consistency of the two data nodes in each data node pair can be determined.

[0052] Optionally, the weights of different data nodes can be dynamically determined based on aviation knowledge and specific training scenarios.

[0053] In some embodiments, determining the optimal supplementary training scheme based on the evaluation results includes: Under preset constraints, multiple candidate supplementary training schemes are generated based on the evaluation results; With the goal of minimizing flight training costs, the optimal supplementary training scheme is determined from multiple candidate supplementary training schemes.

[0054] Optionally, the preset constraints include at least security constraints, resource constraints, time constraints, and logical order constraints.

[0055] Optionally, the candidate supplementary training program may include, but is not limited to: a list of subjects requiring supplementary training, key practice content for each subject, training medium, training environment setup, expected duration of each session, total number of training rounds, and logical arrangement between subjects.

[0056] Optionally, flight training costs may include at least: economic costs, time costs, and efficiency costs.

[0057] In this embodiment of the invention, multiple candidate supplementary training schemes are generated based on the evaluation results under preset constraints; with the goal of minimizing flight training costs, the optimal supplementary training scheme is determined from the multiple candidate supplementary training schemes, thereby achieving the optimal allocation of training resources.

[0058] In some embodiments, the method further includes: The evaluation results are displayed visually. The evaluation results and the optimal supplementary training plan are stored.

[0059] Optionally, the recognition results can be visualized.

[0060] Figure 2 This is a flowchart illustrating the training process of the evaluation model provided in an embodiment of the present invention. Figure 2 As shown, in some embodiments, the evaluation model is trained based on the following steps: Step 210: Obtain raw flight training data samples for multimodal training; Step 220: Preprocess the original multimodal flight training data samples to obtain multimodal flight training data samples; Step 230: Determine the optimal supplementary training scheme labels corresponding to the multimodal flight training data samples; Step 240: Using multimodal flight training data samples as training samples and the optimal supplementary training scheme labels as sample labels, train the initial evaluation model. After training, the evaluation model is obtained.

[0061] Optionally, the initial evaluation model includes: An initial semantic alignment layer is used to extract and match features from multimodal flight training data samples to obtain multiple feature data link samples. The initial deviation identification layer is used to identify deviations in multiple feature data link samples to obtain prediction identification results, which include multiple abnormal data link samples. The initial evaluation layer is used to quantitatively evaluate multiple abnormal data link samples based on the prediction and identification results, and obtain the prediction and evaluation results. The initial decision layer is used to determine the optimal supplementary training scheme based on the prediction evaluation results.

[0062] The flight training evaluation system provided by the present invention is described below. The flight training evaluation system described below can be referred to in correspondence with the flight training evaluation method described above.

[0063] Figure 3 This is a schematic diagram of the flight training and evaluation system provided in an embodiment of the present invention, as shown below. Figure 3 As shown, the flight training evaluation system 300 includes the following: Acquisition unit 310 is used to acquire raw multimodal flight training data during the flight training process; The preprocessing unit 320 is used to preprocess the raw multimodal flight training data to obtain multimodal flight training data. Evaluation unit 330 is used to input multimodal flight training data into evaluation model, which performs feature processing on the multimodal flight training data to obtain multiple feature data links, identifies multiple abnormal data links with deviations from the multiple feature data links, evaluates each abnormal data link, obtains evaluation results, and outputs the optimal supplementary training scheme based on the evaluation results. The evaluation model is trained based on multimodal flight training data samples and the optimal supplementary training scheme labels corresponding to the multimodal flight training data samples.

[0064] Optionally, multimodal flight training data includes: training objective data, execution data, and cognitive summary data.

[0065] Optionally, the evaluation model includes: The semantic alignment layer is used to extract and match features from multimodal flight training data to obtain multiple feature data links. The deviation identification layer is used to identify deviations in multiple feature data links and obtain identification results, which include multiple abnormal data links. The evaluation layer is used to quantitatively evaluate multiple abnormal data links based on the identification results, and obtain the evaluation results. The decision-making layer is used to determine the optimal supplementary training plan based on the evaluation results.

[0066] Optionally, feature extraction and feature matching are performed on the multimodal flight training data to obtain multiple feature data links, including: Feature extraction is performed on the training target data to obtain multiple training target feature data; Feature extraction is performed on the execution data to obtain multiple execution feature data. Feature extraction is performed on the cognitive summary data to obtain multiple cognitive summary feature data; Matching is performed on multiple training target feature data, multiple execution feature data, and multiple cognitive summary feature data to obtain matching results; Based on the matching results, multiple feature data links are obtained; the feature data links use training target feature data, execution feature data, and cognitive summary feature data as data nodes.

[0067] Optionally, based on the identification results, multiple anomalous data links are quantitatively evaluated, including: Based on the identification results and preset thresholds, the severity level of deviation for each abnormal data link is determined; The consistency of the abnormal data nodes in each abnormal data link is scored to obtain the consistency score of the abnormal data nodes; Based on the weights and consistency scores of the abnormal data nodes, the consistency score of each abnormal data link is calculated.

[0068] Optionally, based on the evaluation results, the optimal supplementary training scheme is determined, including: Under preset constraints, multiple candidate supplementary training schemes are generated based on the evaluation results; With the goal of minimizing flight training costs, the optimal supplementary training scheme is determined from multiple candidate supplementary training schemes.

[0069] Optionally, the above system also includes: The display unit is used to visualize the evaluation results; The storage unit is used to store the evaluation results and the optimal supplementary training scheme.

[0070] Optionally, the evaluation model is trained based on the following steps: Obtain raw flight training data samples in multiple modes; The original multimodal flight training data samples are preprocessed to obtain multimodal flight training data samples. Determine the optimal supplementary training scheme labels corresponding to the multimodal flight training data samples; Using multimodal flight training data samples as training samples and the optimal supplementary training scheme labels as sample labels, an initial evaluation model is trained, and an evaluation model is obtained after training.

[0071] Figure 4 This is a schematic diagram of the structure of an electronic device provided in an embodiment of the present invention, such as... Figure 4 As shown, the electronic device may include a processor 410, a communications interface 420, a memory 430, and a communication bus 440. The processor 410, communications interface 420, and memory 430 communicate with each other via the communication bus 440. The processor 410 can call logical instructions in the memory 430 to execute a flight training evaluation method. This method includes: preprocessing the original multimodal flight training data to obtain multimodal flight training data; inputting the multimodal flight training data into an evaluation model, where the evaluation model performs feature processing on the multimodal flight training data to obtain multiple feature data links; identifying multiple abnormal data links with deviations from the multiple feature data links; evaluating each abnormal data link to obtain an evaluation result; and outputting an optimal supplementary training scheme based on the evaluation result. The evaluation model is trained based on multimodal flight training data samples and the optimal supplementary training scheme labels corresponding to the multimodal flight training data samples.

[0072] Furthermore, the logical instructions in the aforementioned memory 430 can be implemented as software functional units and, when sold or used as independent products, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of the present invention. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0073] The system embodiments described above are merely illustrative. The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the modules can be selected to achieve the purpose of this embodiment according to actual needs. Those skilled in the art can understand and implement this without any creative effort.

[0074] Through the above description of the embodiments, those skilled in the art can clearly understand that each embodiment can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solutions, in essence or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, magnetic disk, optical disk, etc., and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods described in the various embodiments or some parts of the embodiments.

[0075] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A flight training assessment method, characterized in that, include: Acquire raw multimodal flight training data during the flight training process; The original multimodal flight training data is preprocessed to obtain multimodal flight training data; The multimodal flight training data is input into the evaluation model, which performs feature processing on the multimodal flight training data to obtain multiple feature data links. Multiple abnormal data links with deviations are identified from the multiple feature data links. Each abnormal data link is evaluated to obtain an evaluation result. Based on the evaluation result, the optimal supplementary training scheme is output. The evaluation model is trained based on multimodal flight training data samples and the optimal supplementary training scheme labels corresponding to the multimodal flight training data samples.

2. The flight training evaluation method according to claim 1, characterized in that, The multimodal flight training data includes: training target data, execution data, and cognitive summary data.

3. The flight training evaluation method according to claim 2, characterized in that, The evaluation model includes: A semantic alignment layer is used to extract and match features from the multimodal flight training data to obtain multiple feature data links. A deviation identification layer is used to identify deviations in multiple feature data links and obtain identification results, the identification results including multiple abnormal data links; An evaluation layer is used to quantitatively evaluate multiple abnormal data links based on the identification results to obtain evaluation results; The decision-making layer is used to determine the optimal supplementary training scheme based on the evaluation results.

4. The flight training evaluation method according to claim 3, characterized in that, The process of extracting and matching features from the multimodal flight training data yields multiple feature data links, including: Feature extraction is performed on the training target data to obtain multiple training target feature data; Feature extraction is performed on the execution data to obtain multiple execution feature data; Feature extraction is performed on the cognitive summary data to obtain multiple cognitive summary feature data; Matching multiple training target feature data, multiple execution feature data, and multiple cognitive summary feature data yields a matching result. Based on the matching results, multiple feature data links are obtained; the feature data links use the training target feature data, the execution feature data, and the cognitive summary feature data as data nodes.

5. The flight training evaluation method according to claim 3, characterized in that, The quantitative evaluation of multiple abnormal data links based on the identification results includes: Based on the identification results and preset thresholds, the severity level of deviation for each abnormal data link is determined; The consistency of the abnormal data nodes in each abnormal data link is scored to obtain the consistency score of the abnormal data nodes. Based on the weight of the abnormal data node and the consistency score of the abnormal data node, the consistency score of each abnormal data link is calculated.

6. The flight training evaluation method according to claim 3, characterized in that, The process of determining the optimal supplementary training scheme based on the evaluation results includes: Under preset constraints, multiple candidate supplementary training schemes are generated based on the evaluation results; With the goal of minimizing flight training costs, the optimal supplementary training scheme is determined from multiple candidate supplementary training schemes.

7. The flight training evaluation method according to claim 1, characterized in that, The method further includes: The evaluation results are then visualized. The evaluation results and the optimal supplementary training scheme are stored.

8. The flight training evaluation method according to claim 1, characterized in that, The evaluation model was trained based on the following steps: Obtain raw flight training data samples in multiple modes; The original multimodal flight training data samples are preprocessed to obtain multimodal flight training data samples. Determine the optimal supplementary training scheme label corresponding to the multimodal flight training data sample; Using the multimodal flight training data samples as training samples and the optimal supplementary training scheme labels as sample labels, an initial evaluation model is trained, and after training, the evaluation model is obtained.

9. A flight training assessment system, characterized in that, include: The acquisition unit is used to acquire raw multimodal flight training data during the flight training process; A preprocessing unit is used to preprocess the original multimodal flight training data to obtain multimodal flight training data. An evaluation unit is used to input the multimodal flight training data into an evaluation model, whereby the evaluation model performs feature processing on the multimodal flight training data to obtain multiple feature data links, identifies multiple abnormal data links with deviations from the multiple feature data links, evaluates each of the abnormal data links to obtain an evaluation result, and outputs the optimal supplementary training scheme based on the evaluation result. The evaluation model is trained based on multimodal flight training data samples and the optimal supplementary training scheme labels corresponding to the multimodal flight training data samples.

10. An electronic device comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the flight training evaluation method as described in any one of claims 1 to 8.

Images