Flight Crew Spatial Orientation Training Evaluation Method

Abstract

The present invention discloses a method for evaluating the spatial orientation training of flight personnel, belonging to the technical field of flight training. The method includes the following steps: S1: Collect training data; S2: Calculate basic evaluation indicators; S3: Aggregate the basic evaluation indicators into main evaluation indicators; S4: Construct a basic importance matrix, calculate the basic weight vector and the main evaluation indicators; S5: Aggregate the main evaluation indicators into comprehensive evaluation indicators, construct a main importance matrix, calculate the main weight vector and the comprehensive evaluation indicators; S6: Output evaluation results according to the basic evaluation indicators, the main evaluation indicators and the comprehensive evaluation indicators respectively. The present invention overcomes the problem that the existing evaluation methods are difficult to comprehensively and accurately analyze data. Analyze the training data of flight personnel from multiple levels, solve the problem of lack of standards in the evaluation process, can more accurately evaluate the flight attitude control ability of flight personnel, enable flight personnel to clarify their own advantages and disadvantages, and facilitate targeted improvement.

Description

Technical Field

[0001] The present invention relates to the technical field of flight training, and particularly to an evaluation method for the spatial orientation training of flight personnel. Background Art

[0002] In flight training, the spatial orientation ability is crucial for flight personnel. The spatial orientation ability directly affects the operation accuracy of flight personnel during flight and their ability to respond to emergencies, thus relating to flight safety. However, there are many problems in the current evaluation system for the spatial orientation training of flight personnel.

[0003] Existing evaluation methods often have difficulty in comprehensively and accurately analyzing various data of flight personnel during training. There is no standard evaluation criterion in the evaluation process, resulting in inaccurate evaluation of the flight attitude control ability of flight personnel. This makes it difficult for flight personnel to clarify their own advantages and disadvantages in attitude control after training and unable to make targeted improvements. When facing abnormal attitude situations, flight personnel cannot obtain effective guidance through training evaluation and are difficult to improve their ability to handle emergencies. Existing evaluation means are difficult to judge the degree of change in aircraft attitude parameters and the matching degree between the action execution time and the time determined by the actual action task. This leads to flight personnel not being able to obtain accurate feedback when performing flight actions and being difficult to improve the accuracy and standardization of actions.

[0004] When evaluating the effect of the spatial orientation training of flight personnel, existing technologies often only consider the precise comparison of detailed data or only consider the overall evaluation of training results. When only considering the precise comparison of detailed data, it is restricted to the change process of specific parameters, and evaluating the training effect with the change process of these specific parameters is one-sided and may not conform to the control granularity of aircraft parameters, resulting in difficulty in accurately adjusting the training plan according to the evaluation results. When only considering the overall evaluation of training results, methods such as deep neural networks and machine learning models are often used. However, this method is based on the abstraction and mapping of data features, and the decision-making process of complex models and the evaluation results output are difficult to understand and explain. This makes it difficult for flight training personnel and experts in related fields to judge the reliability and rationality of the results when facing the analysis results of these complex models, and they are also unable to obtain valuable information from the results to take targeted measures to improve flight training and the operation skills of flight personnel.

[0005] Due to the existence of these problems, the existing evaluation methods for the spatial orientation training of flight personnel cannot provide comprehensive and accurate training feedback for flight personnel, are difficult to meet the ever-increasing requirements of flight training, and hinder the effective improvement of the spatial orientation ability of flight personnel. Summary of the Invention

[0006] The present invention overcomes the problem that existing evaluation methods are difficult to comprehensively and accurately analyze data. It comprehensively and accurately analyzes the training data of flight personnel from multiple levels, and by setting standard parameters, solves the problem of lack of standards in the evaluation process, can more accurately evaluate the flight attitude control ability of flight personnel, enables flight personnel to clearly understand their own advantages and disadvantages, and is convenient for targeted improvement.

[0007] To achieve the above object, the present invention adopts the following scheme:

[0008] A method for evaluating the spatial orientation training of flight personnel, comprising the following steps:

[0009] S1: Collect the training data of the spatial orientation training of flight personnel, and the types of the training data include flight attitude data in each flight stage, abnormal situation data when dealing with attitude abnormal events, and flight action data when performing flight actions;

[0010] S2: Calculate the matching degree between the training data and each preset standard parameter as the basic evaluation index;

[0011] S3: Aggregate the basic evaluation indexes according to the types of the training data into main evaluation indexes, and the main evaluation indexes include flight attitude control accuracy, attitude abnormal handling ability, and flight action accuracy;

[0012] S4: Construct a basic importance matrix according to the importance degree of each basic evaluation index, calculate a basic weight vector according to the basic importance matrix, and calculate the main evaluation index according to the basic weight vector and the basic evaluation index;

[0013] S5: Aggregate the main evaluation indexes into a comprehensive evaluation index, construct a main importance matrix according to the importance degree of each main evaluation index, calculate a main weight vector according to the main importance matrix, and calculate the comprehensive evaluation index according to the main weight vector and the main evaluation index;

[0014] S6: Output a parameter-level evaluation result of the spatial orientation training according to the basic evaluation index, output a task-level evaluation result of the spatial orientation training according to the main evaluation index, and output a comprehensive evaluation result of the spatial orientation training according to the comprehensive evaluation index.

[0015] Preferably, the flight attitude data includes the pitch angle, roll angle, and yaw angle of the aircraft, and the basic evaluation indexes used for aggregating to form the flight attitude control accuracy include pitch angle accuracy, roll angle accuracy, and yaw angle accuracy;

[0016] The calculation method of the pitch angle accuracy is:

[0017] Calculate the average pitch angle And the pitch angle standard value A0 Average difference ;

[0018] For the i-th pitch angle A i Pitch angle and tolerance range [a, b], if A i < a, then calculate the range difference where a < A 0 < b; if the pitch angle A i > b, then calculate the range difference If the range difference is greater than the tolerance threshold, it is determined that the i-th pitch angle A i exceeds the limit;

[0019] The method for outputting the corresponding parameter-level evaluation result: output the average difference of the pitch angle as ΔA; for multiple consecutive pitch angles of sampling points, output the average value of the range differences of these pitch angles and the corresponding sampling time interval;

[0020] The standard value of the roll angle is 0, and the upper and lower limits of the tolerance range are the positive and negative values of the tolerance threshold. The corresponding parameter-level evaluation results output are the average value, maximum value, and minimum value of the roll angle;

[0021] The standard value of the yaw angle is 0, and the upper and lower limits of the tolerance range are the positive and negative values of the tolerance threshold. The corresponding parameter-level evaluation results output are the average value, maximum value, and minimum value of the deflection angle.

[0022] Preferably, the abnormal situation data includes the occurrence time of the attitude abnormal event, the steps and time when the flight crew takes operations, the time when the attitude returns to normal, and the attitude adjustment degree. The basic evaluation indicators for aggregating to form the attitude abnormal handling ability include abnormal recognition time, operation time, operation correct rate, and adjustment in-place degree;

[0023] The abnormal recognition time is the time difference from the occurrence of the attitude abnormal event to the flight crew starting to take operations. The operation time is the time difference from the flight crew starting to take operations to the aircraft attitude returning to normal. The operation accuracy rate is the coincidence degree between the operation steps taken by the flight crew and the standard processing steps of the attitude abnormal event.

[0024] Preferably, the flight action data includes the action steps executed by the flight crew, the action execution time, and the aircraft attitude parameters during the action execution. The basic evaluation indicators for aggregating to form the flight action accuracy include action parameter matching degree, action coherence, and action time accuracy;

[0025] The matching degree of the action parameters is the similarity between the change process of the aircraft attitude parameters and the corresponding preset parameter change process. The action coherence is the smoothness of the change curve of the aircraft attitude parameters. The action time accuracy is the matching degree between the action execution time and the action start time, duration, and end time determined according to the actual action task.

[0026] Preferably, the specific steps of step S4 are as follows:

[0027] Construct a basic importance matrix:

[0028]

[0029] Among them, represents the ratio of the importance degree of the i-th basic evaluation index to the j-th basic evaluation index;

[0030] The method for calculating the basic weight vector is: solve the maximum eigenvalue and the corresponding eigenvector of the basic importance matrix P, and perform normalization processing on the eigenvector to obtain the basic weight vector;

[0031] Calculate the matching degree between the basic evaluation index and the preset standard parameter to obtain the matching degree vector, and the dot product of the basic weight vector and the matching degree vector is the corresponding main evaluation index.

[0032] Preferably, the specific steps of step S5 are as follows:

[0033] Construct a main importance matrix:

[0034]

[0035] Among them, represents the ratio of the importance degree of the i-th main evaluation index to the j-th main evaluation index;

[0036] The method for calculating the main weight vector is: solve the maximum eigenvalue and the corresponding eigenvector of the main importance matrix Q, and perform normalization processing on the eigenvector to obtain the main weight vector;

[0037] Calculate the ratio of the main evaluation index to the preset best evaluation value to obtain the matching degree vector, and the dot product of the main weight vector and the matching degree vector is the corresponding comprehensive evaluation index.

[0038] Preferably, step S6 further includes:

[0039] Use the following dynamic evaluation mechanism to evaluate the progress speed of the pilot:

[0040] Score each major evaluation indicator separately, and calculate the comprehensive evaluation score according to the weight of each major evaluation indicator; establish a data chart with the number of training times as the horizontal axis, and the flight attitude control accuracy score, the aircraft attitude abnormality handling effect score, the flight action accuracy score and the comprehensive evaluation score as the vertical axis, to form a score curve that changes with the number of training times, calculate the slope of the score curve and judge the progress speed of the flight personnel according to the size of the slope;

[0041] Set a slope threshold. If the slope of the comprehensive evaluation score curve exceeds the slope threshold, keep the training content of spatial orientation training unchanged and increase the training volume; otherwise, increase the proportion of training content corresponding to the main evaluation indicator with the largest slope of the score curve, and change the training content corresponding to the main evaluation indicator with the smallest slope of the score curve.

[0042] Preferably, the step S1 further comprises the following data preprocessing steps:

[0043] Perform noise filtering on the collected flight attitude data and use sliding window averaging or Kalman filtering algorithm to eliminate sensor noise;

[0044] Standardize the operation steps in the abnormal situation data, map the flight crew's operation actions to the preset standard operation instruction library, and eliminate invalid operation records;

[0045] The action execution time in the flight action data is timestamped and divided into the start phase, the duration phase and the end phase according to the preset action phase division rules, and the time error of each phase is calculated respectively.

[0046] Preferably, step S6 further includes a real-time feedback and training optimization mechanism:

[0047] During the training process, parameter-level evaluation results are displayed in real time. When the basic evaluation index of flight attitude control accuracy exceeds the limit, the flight crew is notified through visual or auditory alarms in the simulator cockpit.

[0048] When the operation time of the posture abnormality handling capability exceeds the preset threshold, the demonstration animation of the standard processing steps is automatically triggered;

[0049] During the execution of flight actions, the preset standard action trajectory is superimposed through augmented reality equipment based on the real-time calculation results of the action parameter matching degree to guide the flight crew to adjust the operation.

[0050] Preferably, the collection of training data in step S1 also includes behavioral data combined with the virtual reality environment:

[0051] For flight simulation training, the head movement trajectory, eye gaze direction, and limb operation actions of flight personnel are obtained through a VR headset and a motion capture device to generate spatial orientation behavior data;

[0052] Fuse the spatial orientation behavior data with flight attitude data to construct a pilot-aircraft coordinated motion model in three-dimensional space;

[0053] Calculate the spatial cognitive deviation of flight personnel according to the coordinated motion model. The deviation includes visual-vestibular coordination error and spatial azimuth misjudgment angle, and add the deviation as an additional basic evaluation index to the aggregation calculation of flight attitude control accuracy.

[0054] The present invention has at least the following beneficial effects: (1) By collecting various types of data, the spatial orientation ability of flight personnel can be comprehensively evaluated, avoiding the one-sidedness of single-data evaluation; (2) By calculating the matching degree between training data and preset standard parameters, the performance of flight personnel can be accurately quantified, providing more accurate evaluation results; (3) By calculating the weight matrix and weight vector, the weights of various indicators can be dynamically adjusted to ensure that the evaluation results meet the actual training needs; (4) By outputting parameter-level, task-level, and comprehensive evaluation results, the performance of flight personnel can be evaluated from different levels, helping flight personnel clarify their own advantages and disadvantages; (5) The evaluation process is based on clear mathematical calculations and weight assignments, and the evaluation results are easy to understand and interpret, facilitating flight training personnel and experts to take targeted improvement measures according to the evaluation results; (6) Preprocess the training data to improve data quality; provide real-time feedback and training optimization mechanisms during training, such as alarm prompts, demonstration animations, augmented reality guidance, etc., and also combine the VR environment to obtain behavior data, construct a coordinated motion model to calculate spatial cognitive deviation and use it as an additional evaluation index, enabling flight personnel to timely understand their own operation problems and make adjustments, comprehensively improving the training effect and spatial orientation ability of flight personnel. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] Figure 1 It is a principle flowchart of a method for evaluating the spatial orientation training of flight personnel provided by the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0056] The following further describes the present invention in detail with reference to the drawings, so that those skilled in the art can implement it according to the description in the specification.

[0057] As Figure 1 shown, the method for evaluating the spatial orientation training of flight personnel provided by the present invention includes the following steps:

[0058] S1: Collect the training data for the spatial orientation training of flight personnel. The types of the training data include the flight attitude data in each flight phase, the abnormal situation data when dealing with attitude abnormal events, and the flight action data when performing flight actions.

[0059] The flight attitude data can be obtained by means of devices such as a high-precision inertial measurement unit (IMU). The IMU can monitor in real time the acceleration, angular velocity, etc. of flight personnel in each flight phase (such as takeoff, cruise, landing, etc.), so as to accurately restore the flight attitude. The abnormal situation data when dealing with attitude abnormal events is generally collected through the fault injection system in the flight simulator. This system can simulate various attitude abnormal scenarios, such as the aircraft suddenly tilting, stalling, etc., and record the operation reactions and response times of flight personnel when facing these abnormalities. The flight action data when performing flight actions can use the flight data recorder on the aircraft to record the action amplitudes, times, etc. of flight personnel operating devices such as the joystick and throttle.

[0060] S2: Calculate the matching degree between the training data and each preset standard parameter as the basic evaluation index.

[0061] The preset standard parameters can be obtained from a large amount of flight training data and experience summaries obtained from previous flight training at the training base. For example, for the flight attitude data, the standard parameters can include the standard attitude angle range, angular velocity change range, etc. in a specific flight phase. By calculating the deviation degree between the actual flight attitude data and these standard parameters, the matching degree is obtained. Common calculation methods include the Euclidean distance method, cosine similarity, etc. Using this as the basic evaluation index, the performance of flight personnel at each data point is quantitatively evaluated.

[0062] S3: Aggregate the basic evaluation index into the main evaluation index according to the type of the training data. The main evaluation index includes flight attitude control accuracy, attitude abnormal handling ability, and flight action accuracy.

[0063] Aggregate the basic evaluation index according to the type of the training data. The main evaluation index of flight attitude control accuracy is obtained by aggregating and calculating the basic evaluation indexes corresponding to the flight attitude data in multiple flight phases through weighted average or other statistical methods, reflecting the overall control level of flight personnel over the flight attitude. The attitude abnormal handling ability is the comprehensive basic evaluation index corresponding to the abnormal situation data when dealing with attitude abnormal events, considering the reaction speed, decision-making correctness, etc. of flight personnel when facing sudden abnormalities. The flight action accuracy is to analyze and integrate the basic evaluation index corresponding to the flight action data when performing flight actions, and judge the accuracy of flight personnel performing flight actions through the actual gap from the expected parameters.

[0064] S4: Construct a basic importance matrix based on the importance degree of each basic evaluation index, calculate a basic weight vector according to the basic importance matrix, and calculate the main evaluation index based on the basic weight vector and the basic evaluation index.

[0065] Construct a basic importance matrix based on the importance degree of each basic evaluation index. The judgment of the importance degree can be carried out by means of expert scoring. Senior flight instructors, aviation engineers, etc. score the importance of different types of basic evaluation indexes according to actual flight experience and theoretical knowledge. It can also be directionally optimized in combination with reference documents such as operation standards of flight training and special mission requirements. Calculate a basic weight vector according to the basic importance matrix, and then combine the basic weight vector and the basic evaluation index to calculate the main evaluation index through mathematical operations such as weighted summation, so that the main evaluation index can more reasonably reflect the abilities of flight personnel in different aspects.

[0066] S5: Aggregate the main evaluation indexes into a comprehensive evaluation index, construct a main importance matrix according to the importance degree of each main evaluation index, calculate a main weight vector according to the main importance matrix, and calculate the comprehensive evaluation index based on the main weight vector and the main evaluation index.

[0067] Construct a main importance matrix according to the importance degree of each main evaluation index, and calculate a main weight vector from a more general data analysis perspective. Finally, based on the main weight vector and the main evaluation index, calculate the comprehensive evaluation index through corresponding mathematical calculations to quantitatively evaluate the effect of the spatial orientation training of flight personnel as a whole.

[0068] S6: Output a parameter-level evaluation result of the spatial orientation training according to the basic evaluation index, output a task-level evaluation result of the spatial orientation training according to the main evaluation index, and output a comprehensive evaluation result of the spatial orientation training according to the comprehensive evaluation index.

[0069] For the basic evaluation index, analyze the performance of each data point in detail and output the parameter-level evaluation result of the spatial orientation training, providing accurate feedback on the performance of flight personnel in specific parameters and achieving the most detailed and accurate problem positioning. For the main evaluation index, output the task-level evaluation result of the spatial orientation training from the task levels such as flight attitude control, abnormal handling, and flight action execution, helping flight personnel and training institutions understand the ability levels in different tasks. Pilots can conduct targeted training according to their performance in different tasks and continuously adjust their operation methods to ensure stable and efficient execution of corresponding tasks. For the comprehensive evaluation index, give the comprehensive evaluation result of the spatial orientation training, presenting the overall effectiveness of the spatial orientation training of flight personnel in an intuitive way and providing an important basis for subsequent training plan adjustment, flight task assignment, etc.

[0070] In practical applications, first, a flight training data acquisition system is set up, including installing high-precision sensors on the aircraft to collect flight attitude data, such as gyroscopes, accelerometers, etc., to obtain the pitch angle, roll angle, and yaw angle data of the aircraft. These means of obtaining data can be easily achieved using existing technologies. At the same time, an event trigger mechanism is set in the training simulator. When an attitude anomaly event occurs, the abnormal situation data is automatically recorded, including the occurrence time, the operation steps and time of the flight crew, and the time when the attitude returns to normal, etc. When the flight crew performs flight actions, the action monitoring module of the simulator records the action steps, execution time, and aircraft attitude parameters.

[0071] When calculating the basic evaluation indicators, the collected training data is compared with the pre-set standard parameters. These standard parameters can be determined according to flight training outlines, flight safety standards, and expert experience, etc. For example, for flight attitude data, the standard values and tolerance ranges of the pitch angle, roll angle, and yaw angle are set.

[0072] When constructing the basic importance matrix, the importance degree ratios of each basic evaluation indicator are determined through pairwise comparison, so as to construct the basic importance matrix. After calculating the basic weight vector, it is aggregated with the matching degree vector of the basic evaluation indicators to obtain the corresponding main evaluation indicators, and the main evaluation indicators are made more objective and accurate through weight addition.

[0073] Similarly, for the construction of the main importance matrix, the importance degree ratios of each main evaluation indicator are determined through historical training experience, training standards, and task requirements, and then the main weight vector and comprehensive evaluation indicators are calculated.

[0074] Finally, according to the basic evaluation indicators, main evaluation indicators, and comprehensive evaluation indicators, a corresponding evaluation report generation system is developed, and the parameter-level, task-level, and comprehensive evaluation results are output according to requirements. The comprehensive evaluation result can quickly show the overall level of the pilot in each training. For flight training with abnormal comprehensive evaluation results, the short board of the flight crew in different tasks can be quickly analyzed through the task-level evaluation results, and further through the parameter-level analysis, the operation defects of the flight crew when performing tasks can be known, so as to achieve targeted analysis at different levels and adjustment of flight training plans.

[0075] By collecting various types of data, the spatial orientation ability of flight personnel can be comprehensively evaluated, avoiding the one-sidedness of single-data evaluation; by calculating the matching degree between training data and preset standard parameters, the performance of flight personnel can be accurately quantified, providing more accurate evaluation results; by calculating the weight matrix and weight vector, the weights of various indicators can be dynamically adjusted to ensure that the evaluation results meet the actual training requirements; by outputting parameter-level, task-level, and comprehensive evaluation results, the performance of flight personnel can be evaluated from different levels, helping flight personnel to clarify their own advantages and disadvantages; the evaluation process is based on clear mathematical calculations and weight assignments, and the evaluation results are easy to understand and interpret, facilitating flight training personnel and experts to take targeted improvement measures according to the evaluation results.

[0076] In another technical solution, the flight attitude data includes the pitch angle, roll angle, and yaw angle of the aircraft, and the basic evaluation indicators for aggregating to form the flight attitude control accuracy include pitch angle accuracy, roll angle accuracy, and yaw angle accuracy;

[0077] The calculation method of the pitch angle accuracy is as follows:

[0078] Calculate the average pitch angle And the average difference from the pitch angle standard value A 0 ; ;

[0079] For the pitch angle A of the i-th pitch angle i And the tolerance range [a, b], if A i < a, then calculate the range difference , where a < A 0 < b; if the pitch angle A i > b, then calculate the range difference , if the range difference Is greater than the tolerance threshold, it is determined that the i-th pitch angle A i Is out of limit;

[0080] The method of outputting the corresponding parameter-level evaluation result: output the average difference of the pitch angle as ΔA; for multiple consecutive pitch angles of sampling points, output the average value of the range differences of these pitch angles and the corresponding sampling time interval;

[0081] The standard value corresponding to the roll angle is 0, and the upper and lower limits of the tolerance range are the positive and negative values of the tolerance threshold, and the corresponding parameter-level evaluation results output are the average value, maximum value, and minimum value of the roll angle;

[0082] The standard value corresponding to the yaw angle is 0, and the upper and lower limits of the tolerance range are the positive and negative values of the tolerance threshold, and the corresponding parameter-level evaluation results output are the average value, maximum value, and minimum value of the deflection angle.

[0083] When collecting flight attitude data, the pitch angle, roll angle, and yaw angle are sampled in real time according to the preset sampling frequency of the device. For the calculation of pitch angle accuracy, the data processing software is used to calculate the average difference between the average pitch angle and the standard value. For the pitch angle at each sampling point, it is judged whether it is within the tolerance range. If it exceeds the range, the range difference is calculated.

[0084] When outputting the evaluation results at the parameter level, through a dedicated display interface, the average difference of the pitch angle, the average value of the range differences of consecutive sampling points, and the sampling time interval are clearly displayed. For the roll angle and yaw angle, their average values, maximum values, and minimum values are also displayed on the interface, facilitating flight personnel and training personnel to view and analyze.

[0085] The abnormal situation data includes the occurrence time of the attitude abnormal event, the steps and time when the flight personnel take actions, the time when the attitude returns to normal, and the degree of attitude adjustment. The basic evaluation indicators used to aggregate and form the attitude abnormal handling ability include abnormal recognition time, operation time, operation correct rate, and adjustment in-place degree;

[0086] The abnormal recognition time is the time difference from the occurrence of the attitude abnormal event to when the flight personnel start to take actions. The operation time is the time difference from when the flight personnel start to take actions to when the aircraft attitude returns to normal. The operation accuracy rate is the coincidence degree between the operation steps taken by the flight personnel and the standard processing steps of the attitude abnormal event.

[0087] In terms of monitoring attitude abnormal events, using the event monitoring system of the simulator, when an aircraft attitude abnormality is detected, the time recording module is immediately started to record the abnormal occurrence time. When the flight personnel start to operate, the operation steps and time are recorded. When the attitude returns to normal, the recovery time is recorded.

[0088] For the calculation of the operation correct rate, the operation steps of the flight personnel are compared with the steps pre-stored in the standard operation steps library, and the coincidence degree is calculated through string matching algorithms, etc. For the adjustment accuracy degree, the gap between the actual parameters and the expected parameters after the analyst's operation makes the aircraft stable again is calculated, including the adjustment amount and overshoot amount. When evaluating the attitude abnormal handling ability, indicators such as abnormal recognition time, operation time, operation correct rate, and adjustment in-place degree are combined to form a comprehensive evaluation report, providing a basis for training improvement.

[0089] The flight action data includes the action steps executed by the flight personnel, the action execution time, and the aircraft attitude parameters during the action execution. The basic evaluation indicators used to aggregate and form the flight action accuracy include action parameter matching degree, action coherence, and action time accuracy;

[0090] The matching degree of the action parameters is the similarity between the change process of the aircraft attitude parameters and the corresponding preset parameter change process. The action coherence is the smoothness of the change curve of the aircraft attitude parameters. The action time accuracy is the matching degree between the action execution time and the action start time, duration, and end time determined according to the actual action task.

[0091] When recording flight action data, use the action capture and data recording functions of the flight training simulator to accurately record the action steps executed by the flight crew, the action execution time, and the aircraft attitude parameters during the action execution process.

[0092] For the matching degree of action parameters, a simple attitude parameter change model can be established through a parameter table, compare the actual aircraft attitude parameter change process with the preset standard parameter change process, and use a similarity calculation algorithm (such as the cosine similarity algorithm) to calculate the matching degree. For action coherence, perform a smoothness analysis on the change curve of the aircraft attitude parameters, such as calculating indicators such as the curvature change of the curve to measure coherence. For action time accuracy, compare and calculate the error between the start time, duration, and end time set according to the actual action task and the actual action execution time.

[0093] In another technical solution, the specific steps of step S4 are as follows:

[0094] Construct a basic importance matrix:

[0095]

[0096] Among them, represents the ratio of the importance degree of the i-th basic evaluation index to the j-th basic evaluation index;

[0097] The method for calculating the basic weight vector is: solve the maximum eigenvalue and the corresponding eigenvector of the basic importance matrix P, and perform normalization processing on the eigenvector to obtain the basic weight vector;

[0098] Calculate the matching degree between the basic evaluation index and the preset standard parameter to obtain the matching degree vector, and the dot product of the basic weight vector and the matching degree vector is the corresponding main evaluation index.

[0099] When constructing the basic importance matrix, experts can be asked to score the importance of each basic evaluation index pairwise. For example, if it is considered that the pitch angle accuracy is 1 time more important than the roll angle accuracy, then the corresponding P 12= 2. Organize the expert scoring results into a basic importance matrix. Use a mathematical calculation software (such as MATLAB) to solve the maximum eigenvalue and the corresponding eigenvector of the basic importance matrix, and then normalize the eigenvector to obtain the basic weight vector. When calculating the matching degree between the basic evaluation index and the preset standard parameter, adopt corresponding calculation methods according to the characteristics of different indexes, such as calculating the difference for the angle index as mentioned above. After obtaining the matching degree vector, perform a dot product operation with the basic weight vector to obtain the main evaluation index.

[0100] The specific steps of step S5 are as follows:

[0101] Construct the main importance matrix:

[0102]

[0103] Among them, represents the ratio of the importance degree between the i-th main evaluation index and the j-th main evaluation index;

[0104] The method for calculating the main weight vector is: solve the maximum eigenvalue and the corresponding eigenvector of the main importance matrix Q, and normalize the eigenvector to obtain the main weight vector;

[0105] Calculate the ratio of the main evaluation index to the preset best evaluation value to obtain the matching degree vector, and the dot product of the main weight vector and the matching degree vector is the corresponding comprehensive evaluation index.

[0106] The process of constructing the main importance matrix is similar to that of the basic importance matrix. It also forms the main importance matrix by experts' pairwise comparison and scoring of the importance of the main evaluation indexes (flight attitude control accuracy, attitude anomaly handling ability, and flight action accuracy).

[0107] Use a professional mathematical tool to solve the maximum eigenvalue and eigenvector of the main importance matrix, and obtain the main weight vector after normalization. When calculating the ratio of the main evaluation index to the preset best evaluation value, first determine the best evaluation value of each main evaluation index. For example, for flight attitude control accuracy, the best value can be set that each angle error is 0. After calculating the ratio to obtain the matching degree vector, perform a dot product with the main weight vector to obtain the comprehensive evaluation index.

[0108] In another technical solution, step S6 further includes:

[0109] Use the following dynamic evaluation mechanism to evaluate the progress speed of the pilot:

[0110] Score each major evaluation indicator separately, and calculate the comprehensive evaluation score according to the weight of each major evaluation indicator; establish a data chart with the number of training times as the horizontal axis, and the flight attitude control accuracy score, the aircraft attitude abnormality handling effect score, the flight action accuracy score and the comprehensive evaluation score as the vertical axis, to form a score curve that changes with the number of training times, calculate the slope of the score curve and judge the progress speed of the flight personnel according to the size of the slope;

[0111] Set a slope threshold. If the slope of the comprehensive evaluation score curve exceeds the slope threshold, keep the training content of spatial orientation training unchanged and increase the training volume; otherwise, increase the proportion of training content corresponding to the main evaluation indicator with the largest slope of the score curve, and change the training content corresponding to the main evaluation indicator with the smallest slope of the score curve.

[0112] During the training process, after each training session, each major evaluation indicator is scored according to the pre-set scoring criteria. For example, for flight attitude control accuracy, the score is based on the evaluation results of pitch angle accuracy, roll angle accuracy, and yaw angle accuracy, with a full score of 100. The comprehensive evaluation score is calculated based on the weights of each major evaluation indicator.

[0113] Use data analysis software (such as Excel) to draw a score curve with the number of training times as the horizontal axis and the scores of each item (flight attitude control accuracy score, aircraft attitude abnormality handling effect score, flight action accuracy score, and comprehensive evaluation score) as the vertical axis. The progress rate of the flight personnel can be judged by calculating the slope of the score curve. Set a slope threshold, such as 0.3. If the slope of the comprehensive evaluation score curve exceeds the threshold, arrange more training tasks of the same type; if it does not exceed the threshold, analyze the slope of the score curve of each major evaluation indicator, increase the proportion of training content corresponding to the indicator with the largest slope, and reduce the training content corresponding to the indicator with the smallest slope.

[0114] In another technical solution, the step S1 further includes the following data preprocessing steps:

[0115] Perform noise filtering on the collected flight attitude data and use sliding window averaging or Kalman filtering algorithm to eliminate sensor noise;

[0116] Standardize the operation steps in the abnormal situation data, map the flight crew's operation actions to the preset standard operation instruction library, and eliminate invalid operation records;

[0117] The action execution time in the flight action data is timestamped and divided into the start phase, the duration phase and the end phase according to the preset action phase division rules, and the time error of each phase is calculated respectively.

[0118] For the noise filtering of flight attitude data, if the sliding window averaging method is adopted, the window size is set to 5 sampling points, and the data within each window is averaged to obtain the denoised flight attitude data. If the Kalman filter algorithm is used, corresponding filtering parameters are set according to the motion model of the aircraft and the noise characteristics of the sensors, and the collected raw data is filtered.

[0119] For the standardized coding of the operation steps of abnormal situation data, a standard operation instruction library is established, and the operation actions of the flight crew are encoded according to the standards in the instruction library. For example, the operation of "pulling the lever" is encoded as "OP001". During the encoding process, invalid operation records such as repeated operations and misoperations are excluded.

[0120] For the timestamp alignment of flight action data, a high-precision clock synchronization system is used to ensure the accuracy of the recorded execution time of each action. According to the preset action stage division rules, such as dividing the takeoff action into stages such as starting to taxi, accelerating taxiing, lifting the wheels, and leaving the ground, the time errors of each stage are calculated respectively to provide accurate data for subsequent evaluation.

[0121] In another technical solution, the step S6 further includes a real-time feedback and training optimization mechanism:

[0122] During the training process, the parameter-level evaluation results are displayed in real time. When the basic evaluation index of the flight attitude control accuracy exceeds the limit, the flight crew is prompted through visual or auditory warnings in the simulator cockpit;

[0123] When the operation time of the attitude anomaly handling ability exceeds the preset threshold, the demonstration animation of the standard handling steps is automatically triggered;

[0124] During the execution of the flight action, according to the real-time calculation result of the action parameter matching degree, a preset standard action trajectory is superimposed through the augmented reality device to guide the flight crew to adjust the operation.

[0125] In the training simulator cockpit, a visual warning light and a voice warning device are installed. When the basic evaluation index of the flight attitude control accuracy exceeds the limit, such as when the pitch angle exceeds the tolerance range, the warning light flashes and a voice prompt is triggered to inform the flight crew of the attitude anomaly.

[0126] In the simulator software system, the demonstration animation of the standard steps for attitude anomaly handling is pre-stored. When the operation time of the attitude anomaly handling ability exceeds the preset threshold, the demonstration animation is automatically played on the simulator screen, and the flight crew can pause the training to watch the animation and learn the standard handling steps.

[0127] During the execution of flight maneuvers, an augmented reality device, such as smart glasses, is used to superimpose a preset standard action trajectory in the form of a virtual image on the field of vision of the flight crew. By calculating the matching degree of action parameters in real time, when it is found that there is a deviation between the operation of the flight crew and the standard trajectory, the display of the virtual trajectory is adjusted in a timely manner to guide the flight crew to adjust their operations.

[0128] In another technical solution, the acquisition of training data in step S1 further includes behavioral data combined with a virtual reality environment:

[0129] For flight simulation training, the head movement trajectory, eye gaze direction, and limb operation actions of the flight crew are obtained through a VR helmet and an action capture device to generate spatial orientation behavioral data;

[0130] The spatial orientation behavioral data is fused with flight attitude data to construct a pilot-aircraft collaborative motion model in a three-dimensional space;

[0131] The spatial cognitive deviation of the flight crew is calculated according to the collaborative motion model. The deviation includes visual-vestibular coordination error and spatial azimuth misjudgment angle, and the deviation is added as an additional basic evaluation index to the aggregation calculation of flight attitude control accuracy.

[0132] A VR helmet and an action capture device are equipped in the flight training simulator. When the flight crew conducts training, the VR helmet real-time collects head movement trajectory data, and the action capture device collects limb operation action data. These data are integrated through a data fusion algorithm to generate spatial orientation behavioral data.

[0133] Using 3D modeling software and data analysis algorithms, the spatial orientation behavioral data is fused with flight attitude data to construct a pilot-aircraft collaborative motion model. For example, according to the head movement trajectory and eye gaze direction, combined with aircraft attitude data, the relationship between the attention distribution of the flight crew and the change of aircraft attitude is analyzed.

[0134] According to the collaborative motion model, the spatial cognitive deviation of the flight crew is calculated. For the visual-vestibular coordination error, the error value is calculated using a mathematical model by comparing the relationship between head movement and aircraft attitude change. For the spatial azimuth misjudgment angle, according to the operation actions of the flight crew and the actual attitude of the aircraft, the judgment error of the spatial azimuth is judged, and these deviations are used as additional basic evaluation indexes and added to the aggregation calculation of flight attitude control accuracy according to a certain weight.

[0135] Preprocess the training data to improve data quality; provide real-time feedback and training optimization mechanisms during training, such as warning prompts, demonstration animations, augmented reality guidance, etc. Also, combine the VR environment to obtain behavioral data, construct a collaborative motion model to calculate spatial cognitive deviation and use it as an additional evaluation index, enabling flight personnel to promptly understand their own operation problems and make adjustments, comprehensively improving the training effect and spatial orientation ability of flight personnel.

[0136] As a more detailed demonstration example, take one of the spatial orientation trainings of flight personnel as an example. Select 10 flight cadets for a 3-month training evaluation.

[0137] At the initial stage of training, collect flight attitude data, abnormal situation data, and flight action data. After collecting the data, use data preprocessing methods to process the collected data and improve data quality.

[0138] Calculate basic evaluation indicators. For example, calculate that the average difference in pitch angle accuracy of flight cadet A is 0.5 degrees, the average roll angle is 0.2 degrees, and the maximum yaw angle is 0.3 degrees; the average abnormal recognition time is 2 seconds, and the operation correct rate is 80%; the action parameter matching degree is 0.7, and the action coherence score is 8 points (out of 10).

[0139] Construct a basic importance matrix and a main importance matrix in advance, calculate the basic weight vector and the main weight vector, and then obtain the main evaluation indicator and the comprehensive evaluation indicator. For example, the score of flight cadet A for flight attitude control accuracy is 75 points, the score for attitude abnormal situation handling ability is 70 points, the score for flight action accuracy is 72 points, and the comprehensive evaluation score is 72 points.

[0140] During the training process, further use a dynamic evaluation mechanism. Every 5 trainings are completed, evaluate the progress speed of flight cadets. It is found that the slope of the comprehensive evaluation score curve of flight cadet B is 0.2, lower than the set threshold of 0.3. Analyze the score curves of its main evaluation indicators and find that the slope of the score curve for flight attitude control accuracy is the largest, at 0.35, while the slope of the score curve for attitude abnormal situation handling ability is the smallest, at 0.1. Therefore, in subsequent training, increase the proportion of flight attitude control training content from the original 30% to 40%, and reduce the proportion of attitude abnormal situation handling training content from the original 30% to 20%.

[0141] Utilize the real-time feedback and training optimization mechanism. When the flight attitude control accuracy of flight cadet C exceeds the limit, give a visual warning prompt in a timely manner. When the abnormal situation handling operation time of cadet D is too long, automatically play a demonstration animation of the standard handling steps. Further collect spatial orientation behavioral data through VR devices, calculate that the visual-vestibular coordination error of flight cadet E is 0.1, and add it as an additional basic evaluation indicator to the flight attitude control accuracy evaluation.

[0142] Through continuous adjustment and optimization during training, the spatial orientation ability of flight cadets has been significantly improved. The average comprehensive evaluation score can be increased by about 10 points. Compared with the training effect without using this evaluation method, the training has more obvious effectiveness and practicability.

[0143] It should be noted that although the above steps are described in a specific order, it does not mean that the steps must be executed in the above specific order. In fact, some of these steps can be executed concurrently or even in a different order, as long as the required functions can be achieved. The number of devices and the processing scale described here are used to simplify the description of the present invention, and the application, modification, and variation of the present invention are obvious to those skilled in the art.

[0144] Although the embodiments of the present invention have been disclosed as above, it is not limited to the applications listed in the specification and embodiments. It can be fully applied to various fields suitable for the present invention. For those familiar with the field, additional modifications can be easily made. Therefore, without departing from the general concept defined by the claims and the equivalent scope, the present invention is not limited to the specific details and the illustrated examples here.

Claims

1. A method for evaluating flight crew spatial orientation training, characterized in that: The following steps are involved: S1: Collecting training data for flight personnel spatial orientation training, the types of training data include flight attitude data in each flight phase, abnormal situation data when handling abnormal attitude events, and flight action data when performing flight actions; S2: Calculate the matching degree between the training data and various preset standard parameters as a basic evaluation indicator; S3: Aggregating the basic evaluation indicators into main evaluation indicators according to the type of training data, wherein the main evaluation indicators include flight attitude control accuracy, attitude abnormality processing capability and flight action accuracy; S4: constructing a basic importance matrix according to the importance of each basic evaluation indicator, calculating a basic weight vector according to the basic importance matrix, and calculating the main evaluation indicator according to the basic weight vector and the basic evaluation indicator; The specific steps of step S4 are as follows: Constructing the basic importance matrix: in, It represents the ratio of the importance of the i-th basic evaluation index to the j-th basic evaluation index; The basic weight vector is calculated by solving the maximum eigenvalue and the corresponding eigenvector of the basic importance matrix P, and normalizing the eigenvector to obtain the basic weight vector; Calculate the matching degree between the basic evaluation index and the preset standard parameters to obtain the matching degree vector. The dot product of the basic weight vector and the matching degree vector is the corresponding main evaluation index. S5: Aggregate the main evaluation indicators into comprehensive evaluation indicators, construct a main importance matrix according to the importance of each main evaluation indicator, calculate a main weight vector according to the main importance matrix, and calculate the comprehensive evaluation indicator according to the main weight vector and the main evaluation indicators; The specific steps of step S5 are as follows: Construct the main importance matrix: in, It represents the ratio of the importance of the i-th main evaluation indicator to the j-th main evaluation indicator; The main weight vector is calculated by: solving the maximum eigenvalue and the corresponding eigenvector of the main importance matrix Q, and normalizing the eigenvector to obtain the main weight vector; The ratio of the main evaluation index to the preset optimal evaluation value is calculated to obtain the matching degree vector, and the dot product of the main weight vector and the matching degree vector is the corresponding comprehensive evaluation index; S6: Output parameter-level evaluation results of spatial orientation training according to the basic evaluation indicators, output task-level evaluation results of spatial orientation training according to the main evaluation indicators, and output comprehensive evaluation results of spatial orientation training according to the comprehensive evaluation indicators.

2. The flight crew spatial orientation training evaluation method according to claim 1, characterized in that: The flight attitude data includes the pitch angle, roll angle and yaw angle of the aircraft, and the basic evaluation indexes for forming the flight attitude control accuracy include the pitch angle accuracy, the roll angle accuracy and the yaw angle accuracy; The pitch angle accuracy is calculated as follows: Calculate the average pitch angle Average difference from the standard value of pitch angle A0 ; For the i-th pitch angle A i The pitch angle and tolerance range [a, b], if A i <a, then calculate the range difference , where a<A0<b; if the pitch angle A i >b, then calculate the range difference , if the range difference If it is greater than the tolerance threshold, the i-th pitch angle A is determined i Exceeding the limit; The method of outputting the corresponding parameter-level evaluation results is as follows: the average difference of the pitch angle is output as ΔA; for multiple pitch angles with consecutive sampling points, the average value of the range difference of these pitch angles and the corresponding sampling time interval are output; The standard value corresponding to the roll angle is 0, the upper and lower limits of the tolerance range are the positive and negative values ​​of the tolerance threshold, and the corresponding output parameter level evaluation results are the average value, maximum value and minimum value of the roll angle; The standard value corresponding to the yaw angle is 0, the upper and lower limits of the tolerance range are the positive and negative values ​​of the tolerance threshold, and the corresponding output parameter level evaluation results are the average value, maximum value and minimum value of the yaw angle.

3. The flight crew spatial orientation training evaluation method according to claim 1, characterized in that: The abnormal situation data includes the occurrence time of the abnormal posture event, the steps and time taken by the flight crew, the time when the posture returns to normal, and the degree of posture adjustment. The basic evaluation indicators for the posture abnormality handling capability are aggregated to form the abnormality recognition time, operation time, operation accuracy, and adjustment degree. The abnormality recognition time is the time difference between the occurrence of the abnormal posture event and the start of the flight crew to take action. The operation time is the time difference between the start of the flight crew to take action and the return of the aircraft attitude to normal. The operation accuracy is the overlap between the operation steps taken by the flight crew and the standard processing steps for abnormal posture events.

4. The flight crew spatial orientation training evaluation method according to claim 1, characterized in that: The flight action data includes the action steps performed by the flight crew, the action execution time, and the aircraft attitude parameters during the action execution process. The basic evaluation indicators for the accuracy of the flight action are aggregated to form the action parameter matching degree, action coherence, and action time accuracy. The action parameter matching degree is the similarity between the aircraft attitude parameter change process and the corresponding preset parameter change process, the action continuity is the smoothness of the aircraft attitude parameter change curve, and the action time accuracy is the matching degree between the action execution time and the action start time, duration, and end time determined according to the actual action task.

5. The flight crew spatial orientation training evaluation method according to claim 1, characterized in that: The step S6 further comprises: The pilot's rate of progress is assessed using the following dynamic assessment mechanism: Score each major evaluation indicator separately, and calculate the comprehensive evaluation score according to the weight of each major evaluation indicator; establish a data chart with the number of training times as the horizontal axis, and the flight attitude control accuracy score, the aircraft attitude abnormality handling effect score, the flight action accuracy score and the comprehensive evaluation score as the vertical axis, to form a score curve that changes with the number of training times, calculate the slope of the score curve and judge the progress speed of the flight personnel according to the size of the slope; Set a slope threshold. If the slope of the comprehensive evaluation score curve exceeds the slope threshold, keep the training content of spatial orientation training unchanged and increase the training volume; otherwise, increase the proportion of training content corresponding to the main evaluation indicator with the largest slope of the score curve, and change the training content corresponding to the main evaluation indicator with the smallest slope of the score curve.

6. The flight crew spatial orientation training evaluation method according to claim 1, characterized in that: The step S1 also includes the following data preprocessing steps: Perform noise filtering on the collected flight attitude data and use sliding window averaging or Kalman filtering algorithm to eliminate sensor noise; Standardize the operation steps in the abnormal situation data, map the flight crew's operation actions to the preset standard operation instruction library, and eliminate invalid operation records; The action execution time in the flight action data is timestamped and divided into the start phase, the duration phase and the end phase according to the preset action phase division rules, and the time error of each phase is calculated respectively.

7. The flight crew spatial orientation training evaluation method according to claim 1, characterized in that: The step S6 also includes a real-time feedback and training optimization mechanism: During the training process, parameter-level evaluation results are displayed in real time. When the basic evaluation index of flight attitude control accuracy exceeds the limit, the flight crew is notified through visual or auditory alarms in the simulator cockpit. When the operation time of the posture abnormality handling capability exceeds the preset threshold, the demonstration animation of the standard processing steps is automatically triggered; During the execution of flight actions, the preset standard action trajectory is superimposed through augmented reality equipment based on the real-time calculation results of the action parameter matching degree to guide the flight crew to adjust the operation.

8. The flight crew spatial orientation training evaluation method according to claim 1, characterized in that: The collection of training data in step S1 also includes behavioral data combined with the virtual reality environment: For flight simulation training, the flight crew's head movement trajectory, eye gaze direction and limb operation movements are obtained through VR helmets and motion capture equipment to generate spatial directional behavior data; The spatial orientation behavior data is integrated with the flight attitude data to construct a pilot-aircraft cooperative motion model in three-dimensional space; The spatial cognitive deviation of the flight personnel is calculated according to the collaborative motion model, wherein the deviation includes visual-vestibular coordination error and spatial orientation misjudgment angle, and the deviation is added as an additional basic evaluation index into the aggregate calculation of flight attitude control accuracy.

Images