Voice risk assessment method and system

Abstract

The invention discloses a voice risk assessment method and system. The method comprises the following steps of determining an SOP cabin sound standard shouting data set, establishing a pilot cabin sound standard shouting characteristic index set, performing pilot standard shouting coincidence rate analysis based on the SOP cabin sound, designing a questionnaire to determine the weight of a standard shouting list index, performing risk grade division on the SOP-based standard shouting of the pilot through the index weight, and determining the risk grade of the standard shouting of the pilot based on the SOP cabin sound. Risk factors existing in SOP standard calls are monitored, and corresponding improvement measures are provided. The method can comprehensively and systematically identify and evaluate the risk existing in the operation of the airline company, and provides a basis for formulating cabin sound risk prevention and control measures.

Description

Technical Field

[0001] The present application relates to the field of speech algorithms, and more specifically, to a speech risk assessment method and system. Background Art

[0002] The operator of a vehicle will record voice communication when performing operating procedures. These records are of great significance to the operational safety of the vehicle and help with post-incident investigation and analysis in the event of an accident or abnormal situation.

[0003] During the operation of a vehicle, the Cockpit Voice Recorder (CVR) records the communication between vehicle operators, the communication with traffic controllers, the alarm sound of the automation system, and other audio in the cockpit. The analysis computing power that traditional voice systems can provide is relatively weak. Therefore, it is necessary to propose a voice risk assessment method and system to address the defect of the weak analysis computing power that traditional voice systems can provide. Summary of the invention

[0004] The content of this application is used to introduce concepts in a brief form, which will be described in detail in the detailed implementation section below. The content of this application is not intended to identify the key features or essential features of the technical solution claimed for protection, nor is it intended to limit the scope of the technical solution claimed for protection.

[0005] In order to solve the technical problems mentioned in the above background technology section, some embodiments of the present application provide a voice risk assessment method, including:

[0006] Receive voice data based on the cabin sound SOP standard call data set;

[0007] Using speech data, establish a speech feature index set;

[0008] Determine the conformity rate between the speech feature index set and the cabin sound SOP standard call data set;

[0009] Based on the characteristic factors in the cabin sound SOP standard call data set, a weight set of evaluation indicators is established;

[0010] The cabin voice SOP standard call data set is graded using the weight set of evaluation indicators.

[0011] Call compliance rate and cabin voice SOP standard call data set level;

[0012] Generate risk assessment results for voice data;

[0013] Based on the risk assessment results of the voice data, the risk factors in the voice data are fed back.

[0014] Furthermore, the voice data includes the voice content of the cockpit direct preparation process, the voice content of the taxiing process, the voice content of the take-off process, the voice content of the climbing process, the voice content of the cruising process, the voice content of the descent approach process and the voice content of the landing roll process.

[0015] Furthermore, verification information is received while receiving the voice data; the verification information includes flight records, operation logs, weather information and crew information; and both the voice data and the verification information have timestamps.

[0016] Furthermore, the speech feature index includes speech features, behavior features and background features; the speech features are based on speech data; the behavior features and background features are based on verification information.

[0017] Further, marking all speech features in the speech data;

[0018] Select a voice feature;

[0019] Determine whether the voice feature is consistent with the cabin sound SOP standard call data set;

[0020] If the voice feature matches the cabin sound SOP standard call data set, the value in the statistical table is increased by one, and the process returns to selecting a voice feature until all voice features are selected;

[0021] If the voice feature does not match the cabin sound SOP standard shouting data set, return to the selection of a voice feature until all voice features are selected.

[0022] Further, based on the timbre, the owner in the group information of the voice feature is determined; the group information includes the owner information;

[0023] Choose a Belonger;

[0024] The number of voice features of the attributer in the call statistics table;

[0025] Determine the matching rate of the attributer based on the total amount of the attributer's voice features and the number of the attributer's voice features in the statistical table;

[0026] Return to the step of selecting a beneficiary until all beneficiaries have completed their selections.

[0027] Furthermore, based on the objective weighting method of the superior and inferior solution distance, the arrangement order of the characteristic factors in the cabin sound SOP standard call data set is generated;

[0028] Based on the order of characteristic factors and fuzzy comprehensive evaluation method, the weight of each characteristic factor is calculated;

[0029] Based on the weighted average method, the weight of each characteristic factor is weighted averaged;

[0030] Get the weight set of evaluation indicators.

[0031] Furthermore, the levels of the cabin sound SOP standard call data set include standard, not implemented, deviated, difficult to identify, not involved, and no data.

[0032] Furthermore, the voice data is cleaned;

[0033] Recognize voice data;

[0034] Remove outliers and missing values ​​from speech data.

[0035] In order to solve the technical problems mentioned in the above background technology section, some embodiments of the present application provide a voice risk assessment system, including:

[0036] A processor, configured to execute the voice risk assessment method;

[0037] The data collectors are all connected to the processor for communication.

[0038] The beneficial effect of the present application is that: a voice risk assessment method and system are provided. The method includes the following steps: determining the SOP cabin sound standard call data set, establishing a pilot cabin sound standard call feature index set, conducting a pilot standard call compliance rate analysis based on the SOP cabin sound, designing a questionnaire to determine the weight of the standard call single index, dividing the pilot's standard call based on the SOP risk level by the index weight, monitoring the risk factors in the SOP standard call, and providing corresponding improvement measures. This method can comprehensively and systematically identify and evaluate the risks existing in the operation of an airline, and provide a basis for formulating cabin sound risk prevention and control measures. BRIEF DESCRIPTION OF THE DRAWINGS

[0039] The drawings constituting a part of this application are used to provide a further understanding of this application, so that other features, purposes and advantages of this application become more obvious. The illustrative embodiment drawings and their descriptions of this application are used to explain this application and do not constitute an improper limitation on this application.

[0040] In addition, throughout the drawings, the same or similar reference numerals represent the same or similar elements. It should be understood that the drawings are schematic and that the components and elements are not necessarily drawn to scale.

[0041] In the attached picture:

[0042] Figure 1 A method flow chart of a voice risk assessment method provided in one embodiment of the present application.

[0043] Figure 2 A flowchart of a voice risk assessment method provided in one embodiment of the present application.

[0044] Figure 3 A device connection diagram of a speech risk assessment system provided in one embodiment of the present application.

[0045] Reference numerals:

[0046] 100-processor; 200-data collector. DETAILED DESCRIPTION

[0047] Embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although certain embodiments of the present disclosure are shown in the accompanying drawings, it should be understood that the present disclosure can be implemented in various forms and should not be construed as being limited to the embodiments set forth herein. On the contrary, these embodiments are provided to provide a more thorough and complete understanding of the present disclosure. It should be understood that the drawings and embodiments of the present disclosure are only for exemplary purposes and are not intended to limit the scope of protection of the present disclosure.

[0048] It should also be noted that, for ease of description, only the parts related to the invention are shown in the drawings. In the absence of conflict, the embodiments and features in the embodiments of the present disclosure can be combined with each other.

[0049] It should be noted that the concepts such as "first" and "second" mentioned in the present disclosure are only used to distinguish different devices, modules or units, and are not used to limit the order or interdependence of the functions performed by these devices, modules or units.

[0050] It should be noted that the modifications of "one" and "plurality" mentioned in the present disclosure are illustrative rather than restrictive, and those skilled in the art should understand that unless otherwise clearly indicated in the context, it should be understood as "one or more".

[0051] The present disclosure will be described in detail below with reference to the accompanying drawings and in conjunction with embodiments.

[0052] The present application provides a voice risk assessment method.

[0053] Reference Figure 1 As shown, in one embodiment of the present application, a voice risk assessment method includes:

[0054] S100 receives voice data based on the cabin sound SOP standard call data set.

[0055] S200, using the voice data to establish a voice feature index set.

[0056] S300, determining the conformity rate between the speech feature index set and the cabin sound SOP standard call data set.

[0057] S400, based on the characteristic factors in the cabin sound SOP standard call data set, establish a weight set of evaluation indicators.

[0058] S500, using the weight set of evaluation indicators, the cabin sound SOP standard call data set is graded.

[0059] S600, call compliance rate and level of cabin sound SOP standard call data set.

[0060] S700, generating a risk assessment result of the voice data.

[0061] S800, based on the risk assessment result of the voice data, feedback the risk factor in the voice data.

[0062] Determine the standard call data set based on the cockpit sound SOP: Collect the cockpit sound recording data of the airline pilots, including the standard call content of each flight phase such as direct cockpit preparation, taxiing, take-off climb, cruising, descent approach, landing roll, etc. At the same time, collect the corresponding flight records and operation logs, recording the flight status, weather conditions, crew conditions and other information at the time. It is possible to consider using airborne equipment or ground monitoring systems for automated data collection to improve the efficiency and integrity of data collection.

[0063] Establish a standard call feature indicator set for pilots in the cockpit: refer to relevant domestic and international aviation safety standards and research results, and draw on existing mature indicator systems, such as LOSA (Line Operations Safety Audit) and FOQA (Flight Operational Quality Assurance). Combined with the experience of aviation safety experts, determine the key indicators suitable for SOP standard call behavior analysis, including voice characteristics, behavior characteristics, and background characteristics. The design of indicators should reflect measurability, operability, and comparability to ensure the reliability of the data collection and analysis process.

[0064] Carry out cabin voice compliance analysis based on SOP standard calls:

[0065] Standard call compliance analysis: Compare the marked standard call data with the SOP regulations to determine whether each standard call complies with the standard; calculate the standard call compliance rate for each flight phase and each crew member, that is, the matching degree between the actual standard call and the standard call; analyze the distribution characteristics of the compliance rate and identify areas and personnel with low or abnormal standard call rates.

[0066] Identify the key problem points where the shouting behavior deviates from the SOP standard. Propose targeted improvement measures for key influencing factors to optimize the implementation of SOP; at the same time, the fleet's overall habitual reminder shouting can also be screened out to supplement the deficiencies of SOP.

[0067] Design a questionnaire to determine the weights of the standard call sheet indicators: List the standard call sheet indicators to be evaluated, such as accuracy, timeliness, consistency, etc. For each indicator, set a 5-level or 7-level quantitative scoring standard, such as 1-5 points or 1-7 points. Explain the scoring criteria in detail to ensure that the experts understand the basis for the scoring.

[0068] The pilot SOP standard call data set is graded by indicator weight:

[0069] As mentioned above, the relative weights of various standard call list indicators have been determined through questionnaire surveys and other methods. These weights reflect the relative importance of different indicators in the implementation of SOPs.

[0070] According to the indicator weights, SOP implementation is divided into different levels, such as standard, not implemented, deviated, difficult to identify, not involved, no data, etc. Each level has a corresponding score range and descriptive standards to provide a basis for evaluation.

[0071] For each pilot, scores are given for each indicator based on his or her actual performance in fulfilling the SOP.

[0072] Multiply the score of each indicator by its weight, and then sum them up to get the execution score of the overall SOP standard call.

[0073] Monitor the risk factors in the SOP standard calls and provide corresponding improvement measures: Combined with the aforementioned indicator weight analysis, pay attention to those indicators with higher weights and lower scores, such as coverage, sufficiency, effectiveness and other indicators may have problems. At the same time, analyze the degree of dispersion of the scores of each indicator, and pay attention to the indicators with large differences; for indicators with more prominent problems, further analyze the specific reasons for the low scores, such as insufficient training, equipment failure, unreasonable program design, etc. Collect relevant detailed data, including standard call records, pilot feedback, etc.

[0074] Determining the SOP standard cockpit voice call data set includes collecting the standard cockpit voice call records when the pilot performs various SOP tasks, including cockpit direct preparation, taxiing, take-off climb, cruising, descent approach, landing run and other flight phases; establishing a data set to ensure that different aircraft models, routes, weather conditions, etc. are covered, and the data is labeled and pre-processed to remove noise interference, etc. Statistical analysis is performed on the data to understand the distribution characteristics and outliers of the data, laying the foundation for subsequent analysis. Machine learning and other technologies are used to automatically extract and classify voice data to improve analysis efficiency.

[0075] The establishment of a characteristic index set of standard calls for pilots in the cockpit sound includes referring to relevant theories such as phonetics and aviation psychology to determine key indicators that reflect the quality of standard calls. The main indicators include: Sound Pressure Level (SPL): reflects the loudness of the standard call sound; Spectral Features: reflects the clarity and recognizability of the standard call sound; Speech Intelligibility: reflects the comprehensibility of the content of the standard call. In addition, other indicators are also considered, such as speech speed and pitch. The design of indicators should take into account their measurability and operability to ensure the reliability of the data collection and analysis process. Reference can be made to relevant domestic and foreign aviation safety standards and research results, and the existing indicator system can be used for reference.

[0076] The cabin sound-based SOP standard call compliance analysis includes, but is not limited to, analyzing the SOP standard call data using the above-mentioned standard call characteristic indicators; calculating the compliance rate of each indicator, that is, the proportion of standard calls that meet the predetermined standards; and analyzing the causes of non-compliance with the requirements, such as environmental noise interference, poor voice interaction in the cabin, pilot program execution errors, intentional violations, etc. In addition to calculating the overall compliance rate of each indicator, the compliance rate differences under different crew dispatches, different airports, different flight phases, and different meteorological conditions can be further analyzed. The analysis results can be displayed intuitively in a visual way to facilitate the identification of problems and differences. Combined with the experience of aviation safety experts, the analysis results are deeply interpreted to find the root causes of the problems.

[0077] The questionnaire is designed to determine the weight of the standard call single indicator, including but not limited to inviting experts in the aviation field, such as pilots, aviation safety experts, etc. to participate in the questionnaire survey, requiring experts to evaluate the importance of each standard call indicator to flight safety, determine the weight coefficient of each indicator, and use methods such as fuzzy judgment theory to combine expert opinions to obtain the final indicator weight. The graded scoring method is used to let experts score each indicator, and the proportion of each indicator in the total weight is calculated based on the average score of each indicator. The hierarchical analysis method (AHP) can also be used to let experts compare the relative importance of indicators in pairs.

[0078] The grading of the SOP standard shouting procedure by indicator weights includes but is not limited to determining an indicator weight system, formulating SOP standard shouting execution grade standards, calculating the overall SOP shouting execution score, determining the SOP execution grade, and feedback and application of results.

[0079] As mentioned above, the relative weights of various standard indicators have been determined through questionnaire surveys and other means. These weights reflect the relative importance of different indicators in the implementation of SOPs.

[0080] The SOP standard call implementation level standard is formulated, and according to the indicator weights, the SOP standard call implementation is divided into different levels, such as standard, not implemented, deviation, difficult to identify, not involved, no data, etc. Each level has a corresponding score range and descriptive standards to provide a basis for evaluation.

[0081] The overall SOP standard call execution score is calculated by scoring each indicator for each pilot according to his actual performance in fulfilling the SOP standard call, multiplying the score of each indicator by its weight, and then summing them up to obtain the overall SOP standard call execution score.

[0082] The SOP standard shouting execution level is determined by comparing the pilot's overall score with the aforementioned SOP standard shouting execution level standard, and determining the pilot's SOP standard shouting execution level based on the score range.

[0083] The feedback and application of the results will feed back the evaluation results to the pilots, and formulate targeted training plans, incorporate the SOP standard call execution level into the management links such as pilot training, performance appraisal, and promotion, regularly evaluate the pilots' SOP standard call execution level, and track the improvement effects.

[0084] The monitoring of risk factors in the SOP standard calls and providing corresponding improvement measures include but are not limited to continuous monitoring of SOP standard call data, timely discovery of risk factors, analysis of the root causes of the decline in the quality of standard calls, such as environmental factors, operating procedures, equipment failures, etc., and making targeted optimization suggestions, such as adjusting the SOP process, improving the cabin environment, upgrading communication equipment, etc.

[0085] The identification of risk factors, that is, combining the aforementioned indicator weight analysis, focuses on those indicators with higher weights and lower scores, such as accuracy, timeliness, clarity and other indicators that may have problems. At the same time, the degree of dispersion of the scores of various indicators is analyzed, and attention is paid to indicators with large differences.

[0086] The in-depth analysis of risk factors refers to further analyzing the specific reasons for low scores for indicators with more prominent problems, such as insufficient training, equipment failure, unreasonable procedure design, etc., and collecting relevant detailed data, including standard call records, pilot feedback, etc.

[0087] The formulation of improvement measures mentioned above means proposing targeted improvement measures for different risk factors, such as strengthening relevant training, optimizing equipment and program design, and improving management mechanisms; the formulation of measures must fully consider factors such as cost, time, and feasibility.

[0088] Implementing improvements and tracking the results means implementing improvement measures as planned, such as conducting special training, revising SOPs, etc., continuously monitoring changes in various indicators, evaluating the effects of improvement measures, and adjusting and improving improvement measures in a timely manner based on monitoring results.

[0089] The establishment of a continuous improvement mechanism is to incorporate the monitoring and improvement of SOP standard indicators into the daily safety management system, regularly organize expert reviews, collect feedback, analyze risks, and formulate improvement plans to form a closed-loop PDCA management model and continuously optimize the quality of SOP standard implementation.

[0090] Objective and comprehensive risk assessment, based on the SOP standard call execution evaluation of multi-dimensional indicator weights, can more objectively and systematically identify problems in implementation. Targeted improvement measures, classified improvement measures are proposed for specific risk factors, which are conducive to effectively solving practical problems in the implementation of SOP standard calls. The continuously optimized management mechanism incorporates risk assessment and improvement work into daily management, forms a closed-loop improvement mechanism, and realizes the continuous improvement of the quality of SOP standard call execution. Improve overall flight safety. By improving the quality of SOP standard call execution, the risk of human operational errors can be effectively reduced and the safety of the entire flight system can be improved. The invention method integrates SOP standard call execution evaluation, risk identification, improvement implementation and continuous optimization, providing civil aviation enterprises with a set of systematic SOP standard call execution quality management solutions.

[0091] The voice data includes the voice content of the cockpit direct preparation process, the voice content of the taxiing process, the voice content of the take-off process, the voice content of the climbing process, the voice content of the cruising process, the voice content of the descent approach process and the voice content of the landing roll process.

[0092] Verification information is received while receiving voice data; the verification information includes flight records, operation logs, weather information and crew information; and both the voice data and the verification information have timestamps.

[0093] The speech feature index includes speech features, behavior features and background features; the speech features are based on speech data; the behavior features and background features are based on verification information.

[0094] Label all speech features in the speech data.

[0095] Select a voice feature.

[0096] Determine whether the voice feature is consistent with the cabin sound SOP standard call data set.

[0097] If the voice feature matches the cabin sound SOP standard call data set, the value in the statistical table is increased by one, and the process returns to selecting a voice feature until all voice features are selected.

[0098] If the voice feature does not match the cabin sound SOP standard shouting data set, return to the selection of a voice feature until all voice features are selected.

[0099] Based on the timbre, the owner in the group information of the voice feature is determined; the group information includes the owner information.

[0100] Select a beneficiary.

[0101] The number of voice features of the owner in the call statistics table.

[0102] The matching rate of the attributer is determined based on the total amount of the attributer's voice features and the number of the attributer's voice features in the statistical table.

[0103] Return to the step of selecting a beneficiary until all beneficiaries have completed their selections.

[0104] Based on the objective weighting method of the superior and inferior solution distance, the arrangement order of the characteristic factors in the cabin sound SOP standard call data set is generated.

[0105] Based on the arrangement order of characteristic factors and fuzzy comprehensive evaluation method, the weight of each characteristic factor is calculated.

[0106] Based on the weighted average method, the weight of each characteristic factor is weighted averaged.

[0107] Get the weight set of evaluation indicators.

[0108] The levels of the cabin voice SOP standard call data set include standard, not implemented, deviated, difficult to identify, not involved, and no data.

[0109] Clean the voice data.

[0110] Recognize voice data.

[0111] Remove outliers and missing values ​​from speech data.

[0112] Collect cabin voice monitoring data of SOP standard calls, and further supplement other relevant safety data, such as flight delays, unsafe incidents, aviation accidents, etc.; clean the original data, identify and eliminate outliers, missing values, etc.; perform min-max normalization processing on indicator data of different dimensions, and perform Box-Cox transformation of data as needed; check the distribution characteristics of each indicator data to ensure that the prerequisites for subsequent analysis are met.

[0113] All data sets in the standard cabin voice call list are compiled into more than 200 questionnaires. The survey content is to rank the importance of the standard call list procedures, with scores ranging from 1-10. The questionnaires are distributed to the captains of the airlines and weighted according to the captains' technical levels. The questionnaire results can be used as the basis for the division of indicator weights.

[0114] The weight of each indicator is determined by different methods. This step uses the superiority-inferiority solution distance-objective weighting method (TOPSIS-CRITIC) and the fuzzy comprehensive evaluation method to calculate the indicator weight; finally, the obtained weights are weighted averaged to obtain a comprehensive weight.

[0115] The TOPSIS algorithm is as follows:

[0116] Construct a weighted normalized matrix Z, and combine the normalized matrix with the objective weight W_j according to matrix multiplication, as shown below:

[0117] Z=(r ij ) n×m =(W j x ij ) n×m

[0118] Construct positive and negative ideal solutions respectively:

[0119]

[0120] The distances to the positive and negative ideal solutions are calculated by Euclidean distance:

[0121]

[0122] Calculate the posting progress and sort it, as shown in the formula:

[0123]

[0124] The larger the progress value is, the higher the importance ranking is, and vice versa.

[0125] The objective weighting method (CRITIC) algorithm is as follows:

[0126] The variability of the indicator is expressed in the form of standard deviation, as shown in the formula:

[0127]

[0128] Sj represents the standard deviation of the jth indicator, which indicates the difference in the values ​​within the indicator. The larger the standard deviation, the more information it contains, and the more weight should be assigned.

[0129] The conflict of indicators is expressed by the correlation coefficient, as shown in the formula:

[0130]

[0131] rij represents the correlation coefficient between evaluation indicators i and j. The stronger the correlation between this indicator and other indicators, the less conflict there is between this indicator and other indicators, which means that the more repeated information there is with other indicators, the less weight assigned to this indicator.

[0132] The information calculation formula of the indicator is as shown in the formula:

[0133] C j =S j ×R j

[0134] The larger the Cj is, the greater the role of the jth evaluation index in the entire evaluation index system, and more weight should be assigned to it.

[0135] Objective weight of indicator: The objective weight Wj of the jth indicator is as shown in the formula:

[0136]

[0137] The fuzzy comprehensive evaluation model is as follows:

[0138] Determine the factor set. According to the constructed indicator system, establish the risk assessment factor set U of the cabin sound project.

[0139] Determine the comment set. The comment set of the risk assessment index system of the cabin sound project is recorded as V. Taking into account the probability of risk occurrence and the consequences of impact, the assessment level is divided into V = {V1, V2, V3, V4, V5} = {small, small, medium, large, large}. Calculate the consistency ratio (CR) and perform consistency test, and finally determine the comprehensive weight of the index.

[0140] Determine the weight set λ. According to the AHP method mentioned above, determine the comprehensive weight of each indicator, expressed as θ = {θ1, θ2, θ3, θ4, θ5}.

[0141] Establish a fuzzy evaluation matrix. Construct an expert fuzzy evaluation table and invite relevant experts to make risk evaluation judgments on various indicators.

[0142] Through the recovered risk fuzzy assessment table, the membership of the single factor to each comment set V is calculated. The horizontal evaluation quantity of the single factor U_i is R_i = {ri1, ri2, ..., ri5}, rij represents the frequency distribution of the i-th factor Ui on the j-th comment Vj, and the membership matrix R is constructed according to the membership of the indicators.

[0143] Analyze the comprehensive score of each indicator and evaluate the contribution of each indicator to the final evaluation result. Finally, compare the calculation results of the weights of the above four indicators horizontally to obtain the weighted indicator weight as the comprehensive evaluation result.

[0144] Determine the risk level classification threshold. According to the scores of each indicator, the cluster analysis method (k-means) is used to determine the risk level classification threshold. Five levels can be set: no risk, low risk, medium risk, high risk, and extremely high risk, and the specific indicator value range for each level is given.

[0145] The present application provides a speech risk assessment system.

[0146] Reference Figure 3 As shown, in one embodiment of the present application, a speech risk assessment system includes a processor 100 and a data collector 200 .

[0147] The processor 100 is used to execute the voice risk assessment method.

[0148] The data collectors 200 are all in communication connection with the processor 100 .

[0149] The processor 100 receives the collected voice data of the data collector 200, objectively and comprehensively assesses the risk, and the SOP standard shouting execution assessment based on the weight of multi-dimensional indicators can more objectively and systematically identify the problems existing in the execution. Targeted improvement measures, classified improvement measures are proposed for specific risk factors, which are conducive to effectively solving the practical problems in the execution of SOP standard shouting. The management mechanism of continuous optimization incorporates risk assessment and improvement work into daily management, forms a closed-loop improvement mechanism, and realizes the continuous improvement of the execution quality of SOP standard shouting. Improve overall flight safety, by improving the execution quality of SOP standard shouting, the risk of human operation errors can be effectively reduced, and the safety of the entire flight system can be improved. The invention method integrates SOP standard shouting execution assessment, risk identification, improvement implementation and continuous optimization, and provides a set of systematic SOP standard shouting execution quality management solutions for civil aviation enterprises.

[0150] The above descriptions are only some preferred embodiments of the present disclosure and an explanation of the technical principles used. Those skilled in the art should understand that the scope of the invention involved in the embodiments of the present disclosure is not limited to the technical solutions formed by a specific combination of the above-mentioned technical features, but should also cover other technical solutions formed by any combination of the above-mentioned technical features or their equivalent features without departing from the above-mentioned inventive concept. For example, the above-mentioned features are replaced with the technical features with similar functions disclosed in the embodiments of the present disclosure (but not limited to) and the technical solutions formed.

Claims

1. A voice risk assessment method, characterized in that: include: Receive voice data based on the cabin sound SOP standard call data set; Using speech data, establish a speech feature index set; Determine the conformity rate between the speech feature index set and the cabin sound SOP standard call data set; Based on the characteristic factors in the cabin sound SOP standard call data set, a weight set of evaluation indicators is established; The cabin voice SOP standard call data set is graded using the weight set of evaluation indicators. Call compliance rate and cabin voice SOP standard call data set level; Generate risk assessment results for voice data; Based on the risk assessment results of the voice data, the risk factors in the voice data are fed back.

2. The voice risk assessment method according to claim 1, characterized in that: The receiving of voice data based on the cabin sound SOP standard shouting data set includes: The voice data includes the voice content of the cockpit direct preparation process, the voice content of the taxiing process, the voice content of the take-off process, the voice content of the climbing process, the voice content of the cruising process, the voice content of the descent approach process and the voice content of the landing roll process.

3. The voice risk assessment method according to claim 2, characterized in that: The receiving of voice data based on the cabin sound SOP standard shouting data set also includes: Verification information is received while receiving voice data; the verification information includes flight records, operation logs, weather information and crew information; and both the voice data and the verification information have timestamps.

4. The voice risk assessment method according to claim 3, characterized in that: The method of using speech data to establish a speech feature index set includes: The speech feature index includes speech features, behavior features and background features; the speech features are based on speech data; the behavior features and background features are based on verification information.

5. The voice risk assessment method according to claim 4, characterized in that: The determination of the conformity rate between the speech feature index set and the cabin sound SOP standard shouting data set includes: Label all speech features in the speech data; Select a voice feature; Determine whether the voice feature is consistent with the cabin sound SOP standard call data set; If the voice feature matches the cabin sound SOP standard call data set, the value in the statistical table is increased by one, and the process returns to selecting a voice feature until all voice features are selected; If the voice feature does not match the cabin sound SOP standard shouting data set, return to the selection of a voice feature until all voice features are selected.

6. The voice risk assessment method according to claim 5, characterized in that: The determination of the conformity rate between the voice feature index set and the cabin sound SOP standard shouting data set also includes: Determine the owner in the group information of the voice feature based on the timbre; the group information includes the owner information; Choose a Belonger; The number of voice features of the attributer in the call statistics table; Determine the matching rate of the attributer based on the total amount of the attributer's voice features and the number of the attributer's voice features in the statistical table; Return to the step of selecting a beneficiary until all beneficiaries have completed their selections.

7. The voice risk assessment method according to claim 6, characterized in that: The weight set of evaluation indicators is established based on the characteristic factors in the cabin sound SOP standard call data set, including: Based on the objective weighting method of superior and inferior solution distance, the arrangement order of characteristic factors in the cabin sound SOP standard call data set is generated; Based on the order of characteristic factors and fuzzy comprehensive evaluation method, the weight of each characteristic factor is calculated; Based on the weighted average method, the weight of each characteristic factor is weighted averaged; Get the weight set of evaluation indicators.

8. The voice risk assessment method according to claim 7, characterized in that: The weight set of evaluation indicators is used to classify the cabin sound SOP standard call data set, including: The levels of the cabin voice SOP standard call data set include standard, not implemented, deviated, difficult to identify, not involved, and no data.

9. The voice risk assessment method according to claim 8, characterized in that: The method further comprises: Clean the voice data; Recognize voice data; Remove outliers and missing values ​​from speech data.

10. A voice risk assessment system, characterized in that: include: A processor, configured to execute the voice risk assessment method according to any one of claims 1 to 9; The data collectors are all connected to the processor for communication.

Images