Pilot psychological selection method, device and system and storage medium

Abstract

The invention discloses a pilot psychological selection method, device and system and a storage medium. The method comprises the following steps: collecting psychological index data, psychological state data and physiological state data of a subject; performing normalization processing on the psychological index data, the psychological state data and the physiological state data, and fusing subjective and objective loads to calculate a pilot load state score; missing value processing is carried out on the psychological index data, the psychological state data and the physiological state data; the incidence relation between the various work-related psychological indexes and the subjective and objective states is calculated; and psychological selection is carried out according to the pilot load states, the pilot psychological scores and the corresponding secondary index scores in different environments. By adopting the technical scheme of the invention, low-cost and efficient pilot key psychological element assessment is realized.

Description

Technical Field

[0001] The present invention belongs to the technical field of multidimensional data processing, and in particular relates to a pilot psychological selection method and device, a system, and a storage medium. Background Art

[0002] Discrete cognitive component tests and work sample tests are two mainstream assessment methods for pilot selection. Discrete cognitive component tests are based on multiple independent cognitive abilities (such as spatial orientation, reasoning, and cooperative communication) and predict flight performance by appropriately combining test scores from these abilities. However, this approach overemphasizes single-dimensional testing and overlooks the important role of meta-component factors (such as task strategy and ability complementarity), potentially leading to inadequate predictions of performance on complex tasks.

[0003] In contrast, work sample tests directly assess applicants' performance by simulating actual work scenarios, resulting in relatively high predictive validity. However, they are costly to implement and fail to comprehensively cover all required competencies. The utility of work sample tests is particularly limited given the uncertainty surrounding future flight scenarios and the allocation of human and aircraft functions. Furthermore, this method tends to assess short-term performance and is unable to accurately predict long-term potential. Summary of the Invention

[0004] The technical problem to be solved by the present invention is to provide a pilot psychological selection method and device, system, and storage medium to address the limitations of traditional pilot selection methods, including insufficient accuracy in predicting performance in complex and uncertain work scenarios, difficulty in integrating multi-dimensional psychological data, high evaluation costs, and lack of modeling of the relationship between psychological data and pilot status. By integrating psychological data with mathematical models, low-cost and efficient assessment of key psychological factors of pilots can be achieved, and the changing needs of future flight scenarios can be adapted to provide scientific, accurate, and flexible selection and evaluation solutions.

[0005] To achieve the above object, the present invention adopts the following technical solutions:

[0006] A pilot psychological selection method comprising:

[0007] Step S1, collecting the subject's psychological index data, psychological state data and physiological state data;

[0008] Step S2: normalize the psychological index data, psychological state data, and physiological state data, and integrate subjective and objective loads to calculate the pilot load state score;

[0009] Step S3, performing missing value processing on psychological index data, psychological state data, and physiological state data;

[0010] Step S4: calculating the correlation between multiple work-related psychological indicators and subjective and objective states;

[0011] Step S5: Psychological selection is performed based on the pilot's load status, the pilot's psychological score and its corresponding secondary index score under different environments.

[0012] Preferably, in step S3, the missing values are predicted using a multiple interpolation technique based on Bayesian thinking.

[0013] Preferably, step S4 includes:

[0014] S41. Construct a linear regression model, whose model equation is expressed as:

[0015] Y=Xβ+ε

[0016] Among them, Y is the fusion load status, X is the matrix of normalized psychological index scores, β is the coefficient vector of the independent variable, and ε is the error term;

[0017] S42. Determine the objective function, which is expressed as:

[0018]

[0019] Where λ is a parameter that controls the strength of regularization, and p is the number of independent variables;

[0020] S43. Determine the ridge estimate: By minimizing the objective function with the regularization term, the obtained ridge estimate is expressed as:

[0021]

[0022] S44. Adjust the regularization strength: confirm the appropriate λ through cross-validation;

[0023] S45. Analyze the goodness of fit R of the regression equation based on the coefficient vector of the independent variable. 2 :

[0024]

[0025] Among them, SS res Represents the residual sum of squares, SS tot represents the total sum of squared deviations.

[0026] The present invention also provides a pilot psychological selection device, comprising:

[0027] The acquisition module is used to collect the psychological index data, psychological state data and physiological state data of the subjects;

[0028] The first processing module is used to normalize the psychological index data, psychological state data, and physiological state data, and integrate subjective and objective loads to calculate the pilot load state score;

[0029] The second processing module is used to process missing values of psychological index data, psychological state data, and physiological state data;

[0030] A calculation module is used to calculate the correlation between various work-related psychological indicators and subjective and objective states;

[0031] The selection module is used to conduct psychological selection based on the pilot's load status, pilot psychological scores and their corresponding secondary indicator scores in different environments.

[0032] Preferably, the second processing module predicts missing values based on a multiple interpolation technique based on Bayesian thinking.

[0033] The present invention also provides a pilot psychological selection system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes a pilot psychological selection method when executed by the processor.

[0034] The present invention also provides a storage medium having a computer program stored thereon, and the computer program executes the pilot psychological selection method when running.

[0035] This method effectively addresses the impact of high environmental uncertainty on performance evaluation by integrating multiple individual psychological traits and mental workload. Furthermore, it employs techniques such as normalization, the coefficient of variation method, multiple interpolation, and mathematical modeling to improve data utilization efficiency and assessment accuracy. This provides an important reference for pilot selection in uncertain mission environments, empirically quantifying the importance of psychological indicators, and conducting scientific and efficient individual status assessments. BRIEF DESCRIPTION OF THE DRAWINGS

[0036] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are merely embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on the provided drawings without paying any creative work.

[0037] Figure 1 This is a flow chart of a pilot psychological selection method according to an embodiment of the present invention;

[0038] Figure 2 Schematic diagram of the MICE steps. DETAILED DESCRIPTION

[0039] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0040] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0041] Example 1:

[0042] like Figure 1 As shown, an embodiment of the present invention provides a pilot psychological selection method, comprising:

[0043] Step S1, data collection: collecting psychological index data of the subjects; conducting uncertainty environment simulation experiments on some subjects, collecting psychological state data and physiological state data of the subjects;

[0044] Step S2: Data preprocessing: normalizing the psychological index data, subjective psychological state data, and physiological state data, and integrating subjective and objective loads to calculate the pilot load state score;

[0045] Step S3, missing value processing: Since some subjects did not participate in the simulated work experiment, their status data are missing. The missing values are predicted through multiple interpolation;

[0046] Step S4: calculating the correlation between multiple work-related psychological indicators and subjective and objective states;

[0047] Step S5: Psychological selection is performed based on the pilot's load status, the pilot's psychological score and its corresponding secondary index score under different environments.

[0048] As an implementation of the embodiment of the present invention, step S1 includes:

[0049] S11. Collect the psychological data of the subjects and build a psychological index database of the subjects;

[0050] S12. Select flight mode and pilot role. Select the flight mode (single pilot, single pilot with alternate pilot, single pilot with virtual pilot, single pilot with ground pilot, single pilot with onboard automation, single pilot with onboard and ground automation, single pilot with ground operator, and onboard and ground automation) and the corresponding pilot role.

[0051] S13. Conduct an experiment simulating an uncertain environment to assess the subjective and objective workload status of some subjects. Given the high uncertainty of the work assignment and environment, subjective and objective workload status, rather than work performance, was identified as the key analysis variable. The experiment designed specific scenario-based tasks to collect cognitive workload data from subjects in a simulated work environment, serving as the core indicator for workload assessment. Data collection utilized a combination of subjective and objective measurement methods, based on the correlation between the degree of mental resource utilization and subjectively perceived self-effort. In other words, when individuals perceive a higher level of self-effort, their mental resource utilization also increases, leading to changes in subjective workload and related physiological indicators. The subjective measurement method offers a significant advantage due to its ease of use. Relying on subjective perception, this method causes little disruption to the workflow and can be performed without disrupting the subjects' normal work. Meanwhile, the objective measurement method utilizes non-invasive physiological equipment to collect data, ensuring smooth implementation without disrupting workflow. Specifically, the experiment quantified the subjects' subjective load status through the NASA-TLX scale, and quantified their objective physiological status through indicators such as heart rate variability and pupil radius, evaluating the pilots' load level from multiple perspectives and providing multi-dimensional support for status analysis.

[0052] As an implementation of the embodiment of the present invention, step S2 includes:

[0053] S21. Each indicator is calculated normally according to the scoring method, and all types of data are normalized. Because the evaluation indicators of various psychological and state measurement data have different dimensions and orders of magnitude, in order to ensure the reliability of the data, the data must be normalized before conducting project analysis. Therefore, deviation standardization (min-max standardization) is used to map the final indicators to the interval [0,1]. The calculation methods of positive indicators ("benefit-oriented" indicators, where the larger the indicator value, the better) and negative indicators ("cost-oriented" indicators, where the smaller the indicator value, the better) are different, as shown in the following formula:

[0054] Normalization of positive indicators:

[0055]

[0056] Normalization of reverse indicator:

[0057]

[0058] Where: X represents the original data of each indicator, X max Indicates the maximum value in the original data, X min Indicates the minimum value in the original data.

[0059] S22. Fusion calculation of subjective and objective state data, using the coefficient of variation method to calculate the coefficient of variation ν of each dimension of subjective load and objective physiological evaluation indicators i and weight ω i , and finally the total status score is:

[0060]

[0061] Where: Represents the average value of the project indicator evaluation value; σ i Indicates the standard deviation of the project indicator evaluation value; ν i is the coefficient of variation of each project indicator; y i is the normalized score data of each item of subjective and objective load indicators; Y is the total score of the fusion load state.

[0062] As an implementation of an embodiment of the present invention, the missing value processing method in step S3 includes:

[0063] The subjects who did not participate in the uncertainty environment simulation experiment had a problem of missing load status data, which was supplemented by the data interpolation method. The MICE (Multiple Imputation by Chained Equations) method was selected. MICE is a multiple interpolation technology based on Bayesian thinking, which is used to deal with missing data. Its basic principle is to estimate missing values based on the information of existing data through multiple iterations, and continuously update the estimation model. First, use the mice function to interpolate the data to form 5 data sets. Then use the with function to analyze all data sets. Finally, use the pool function to summarize and output the results. The interpolation method uses PMM (Predictive Mean Matching), which is a model-based data interpolation method. It predicts missing values by establishing a prediction model, and selects the closest mean from the existing observations for matching based on the prediction results, such as Figure 2 shown.

[0064] As an implementation of the embodiment of the present invention, step S4 includes:

[0065] S41. Constructing a linear regression model: Ridge regression is based on the linear regression model, and its model equation can be expressed as:

[0066] Y=Xβ+ε

[0067] Among them, Y is the dependent variable (fusion load state, calculated by S22), X is the matrix containing each independent variable (normalized scores of each psychological indicator), β is the coefficient vector of the independent variable, and ε is the error term. For example, if the psychological ability X is divided into general ability A, special ability B and trait attitude C, in the general ability, the normalized score of each indicator is a推理能力 , a 判断能力 , a 记忆能力 ; In special abilities, the normalized score of each indicator is b 眼手能力 、b 交流能力 、b 空间能力 ; The normalized score of each trait attitude indicator is c 人格特质 、c 风险人格 、c 危险态度 .but This step is mainly to estimate β by minimizing the residual sum of squares, and its objective function is In this way, the influence of each indicator on the load state can be obtained.

[0068] S42. Determine the objective function: In ordinary linear regression, when there is multicollinearity in the independent variables, the model will be unstable and the coefficients will be too large. To solve this problem, ridge regression introduces a regularization term (L2 norm) in the objective function to penalize the coefficients of the independent variables, forming a new objective function:

[0069]

[0070] Where λ is a parameter that controls the regularization strength and p is the number of independent variables. The first part is still the residual sum of squares in linear regression, which maintains the model structure of linear regression. The second part It is a regularization term that prevents the model from overfitting by penalizing the size of the coefficient.

[0071] S43. Determine the ridge estimate: By minimizing the objective function with the regularization term, the obtained ridge estimate can be expressed as:

[0072]

[0073] In ridge regression, the coefficients obtained by minimizing the new objective function are called "ridge estimates." This process is still based on the framework of the linear regression model, but regularization terms are considered during optimization. Therefore, ridge regression introduces constraints on the original linear regression model. By optimizing the new objective function, more robust regression coefficients are obtained, which can better analyze the impact of various indicators on load status.

[0074] S44. Adjust the regularization strength: The choice of parameter λ is crucial to model performance. A larger λ results in stronger regularization, which makes the estimated coefficients closer to zero. Use cross-validation to confirm the appropriate λ.

[0075] S45. Analyze the goodness of fit of the regression equation (R 2 ):

[0076]

[0077] Among them, SS res Represents the residual sum of squares, SS tot represents the total sum of squared deviations.

[0078] In summary, the established linear regression model provides a basic framework for the entire ridge regression process. The introduction of regularization, calculation of ridge estimation, and adjustment of parameters are all improvements and optimizations based on this framework to address the shortcomings of linear regression in specific scenarios (such as multicollinearity and high-dimensional data). This method inherits the advantages of linear regression (the simplicity and interpretability of the linear model) and improves the stability and predictive ability of the model through regularization. It can effectively improve the degree of prediction of psychological indicators for load status, accurately calculate the coefficient vector β of each psychological ability, and understand the specific goodness of fit (R 2 ).

[0079] As an implementation of the embodiment of the present invention, step S5 includes:

[0080] S51, state prediction calculation:

[0081]

[0082] Among them, Y 状 is the load state of the test object, x i are the respective variables (normalized scores of psychological indicators), β i is the coefficient vector of the independent variable, and c is the constant term to be sought.

[0083] S52. Calculation of total psychological assessment score:

[0084]

[0085] Among them, Y 总 is the total score of the test subject, x i are the respective variables (normalized scores of psychological indicators), β i is the coefficient vector of the independent variable, and c is the constant term to be sought.

[0086] S53. Calculation of scores for each secondary psychological indicator:

[0087]

[0088] Among them, Y j is the score of each psychological secondary index of the subjects, x j is the jth indicator, β j is the coefficient vector of the j-th indicator.

[0089] S54. Organize the calculated scores to accurately predict pilots' load states in highly uncertain environments, and further clarify pilots' psychological scores and their corresponding secondary indicator scores in different environments. By analyzing the relationship between pilots' psychological indicators and load states in different environments and comparing the coefficients of psychological assessment indicators, we can identify pilots' key psychological indicators in different environments. Based on the secondary indicator scores, we can comprehensively understand the pilots' psychological strengths and weaknesses and analyze their suitability for different environments. The research results will provide strong support for optimizing pilot selection and training programs, rationally allocating tasks, and improving flight safety.

[0090] In response to the high uncertainty in the development of future flight cockpit layouts, an embodiment of the present invention proposes a pilot psychological selection method for highly uncertain task environments. This method, driven by psychological data, meets the psychological selection needs in highly uncertain task environments by accurately assessing and predicting the pilot's psychological state during the task. This method aims to balance the scientific rigor of discrete cognitive component tests with the authenticity of work sample tests, and to achieve the assessment and screening of pilots' key psychological factors in an efficient and low-cost manner. This method, based on the scientific rigor of discrete cognitive component tests and the authenticity of work sample tests, aims to achieve the precise assessment and screening of pilots' key psychological factors in an efficient and low-cost manner.

[0091] Specifically, the embodiment of the present invention uses the psychological indicators in the discrete cognitive component test as independent variables, combines them with mathematical models, and uses the pilot load state variables that are highly correlated with flight performance as dependent variables to perform predictions and evaluations. During implementation, the method involves the collection, cleaning, integration, and fusion of multidimensional data, including operations such as extraction, matching, alignment, and conversion of categorical data. Compared with the limitations of traditional assessment systems that only evaluate single psychological factors or score multidimensional indicators separately, this method uses multi-source data fusion technology to scientifically integrate multidimensional test results and empirically quantify the importance of each psychological factor of pilots within a standard psychological framework.

[0092] The embodiments of the present invention make up for the shortcomings of traditional evaluation methods, provide a more accurate solution for predicting pilot status in complex and ambiguous work scenarios in the future, and significantly improve the scientificity and practicality of pilot selection and evaluation in uncertain environments.

[0093] Example 2:

[0094] An embodiment of the present invention further provides a pilot psychological selection device, comprising:

[0095] The acquisition module is used to collect the psychological index data, psychological state data and physiological state data of the subjects;

[0096] The first processing module is used to normalize the psychological index data, psychological state data, and physiological state data, and integrate subjective and objective loads to calculate the pilot load state score;

[0097] The second processing module is used to process missing values of psychological index data, psychological state data, and physiological state data;

[0098] A calculation module is used to calculate the correlation between various work-related psychological indicators and subjective and objective states;

[0099] The selection module is used to conduct psychological selection based on the pilot's load status, pilot psychological scores and their corresponding secondary indicator scores in different environments.

[0100] As an implementation of the embodiment of the present invention, the second processing module predicts missing values based on a multiple interpolation technique based on Bayesian thinking.

[0101] Example 3:

[0102] An embodiment of the present invention further provides a pilot psychological selection system, comprising: a memory and a processor, wherein the memory stores a computer program executed by the processor, and the computer program executes a pilot psychological selection method when executed by the processor.

[0103] Example 4:

[0104] An embodiment of the present invention further provides a storage medium having a computer program stored thereon. The computer program executes the pilot psychological selection method when running.

[0105] The embodiments described above are merely descriptions of preferred embodiments of the present invention and are not intended to limit the scope of the present invention. Without departing from the spirit of the present invention, various modifications and improvements made to the technical solutions of the present invention by persons skilled in the art should fall within the scope of protection defined by the claims of the present invention.

Claims

1. A method for psychological selection of pilots, characterized in that: include: Step S1, collecting the subject's psychological index data, psychological state data and physiological state data; Step S2: normalize the psychological index data, psychological state data, and physiological state data, and integrate subjective and objective loads to calculate the pilot load state score; Step S3, performing missing value processing on psychological index data, psychological state data, and physiological state data; Step S4: calculating the correlation between multiple work-related psychological indicators and subjective and objective states; Step S5: Psychological selection is performed based on the pilot's load status, the pilot's psychological score and its corresponding secondary index score under different environments.

2. The pilot psychological selection method according to claim 1, characterized in that: In step S3, the missing values are predicted using multiple interpolation techniques based on Bayesian thinking.

3. The pilot psychological selection method according to claim 2, characterized in that: Step S4 includes: S41. Construct a linear regression model, whose model equation is expressed as: Y=Xβ+ε Among them, Y is the fusion load status, X is the matrix of normalized psychological index scores, β is the coefficient vector of the independent variable, and ε is the error term; S42. Determine the objective function, which is expressed as: Where λ is a parameter that controls the strength of regularization, and p is the number of independent variables; S43. Determine the ridge estimate: By minimizing the objective function with the regularization term, the obtained ridge estimate is expressed as: S44. Adjust the regularization strength: confirm the appropriate λ through cross-validation; S45. Analyze the goodness of fit R of the regression equation based on the coefficient vector of the independent variable. 2 : Among them, SS res Represents the residual sum of squares, SS tot represents the total sum of squared deviations.

4. A pilot psychological selection device, characterized in that: include: The acquisition module is used to collect the psychological index data, psychological state data and physiological state data of the subjects; The first processing module is used to normalize the psychological index data, psychological state data, and physiological state data, and integrate subjective and objective loads to calculate the pilot load state score; The second processing module is used to process missing values of psychological index data, psychological state data, and physiological state data; A calculation module is used to calculate the correlation between various work-related psychological indicators and subjective and objective states; The selection module is used to conduct psychological selection based on the pilot's load status, pilot psychological scores and their corresponding secondary indicator scores in different environments.

5. The pilot psychological selection device according to claim 4, characterized in that: The second processing module predicts missing values based on the multiple interpolation technology of Bayesian thinking.

6. A pilot psychological selection system, characterized by: include: A memory and a processor, wherein the memory stores a computer program executed by the processor, and when the computer program is executed by the processor, the pilot psychological selection method according to any one of claims 1 to 3 is executed.

7. A storage medium, characterized in that: The storage medium stores a computer program, which, when running, executes the pilot psychological selection method according to any one of claims 1 to 3.

Images