Performance test data analysis method for fire-fighting protective clothing of firefighter

Abstract

The invention provides a performance test data analysis method for fireman fire extinguishing protective clothing, and relates to the technical field of protection monitoring, and the method comprises the steps: carrying out the first classification of the protective clothing according to design characteristics; collecting operation conditions and physiological parameters of a wearer in each use process of the plurality of protective garments of each category, setting a first time window according to the operation conditions, and obtaining correlation coefficients among the physiological parameters; performing first grouping on the physiological parameters according to the correlation coefficient; through machine learning, respectively establishing association models of the operation conditions under the first classifications and the first groups, selecting clustering features according to weights of features in the association models, and identifying groups with similar features through clustering analysis; combining the relation model and a clustering analysis result to obtain an analysis report, and determining improvement and use suggestions of the protective clothing; the performance of the existing protective clothing can be effectively evaluated, and a scientific basis can be provided for the design of the protective clothing in the future.

Description

Technical Field

[0001] The present application relates to the field of protection monitoring technology, and in particular to a performance test data analysis method for firefighters' firefighting protective clothing. Background Art

[0002] Firefighters face extremely high temperatures and harsh environmental conditions when performing firefighting operations, which pose a serious threat to their health and safety. In order to protect firefighters from harmful factors such as high temperature, flames, and smoke, firefighting protective clothing for firefighters came into being. These protective clothing uses special materials and designs to provide adequate protection while minimizing the burden on firefighters during operations.

[0003] However, different protective clothing designs have different characteristics and their performance varies. In order to ensure the safety and efficiency of firefighters, it is particularly important to accurately evaluate the performance of protective clothing. Traditional performance evaluation methods often rely on laboratory tests. Although these tests can provide certain data support, they often cannot fully reflect the performance of protective clothing in actual use.

[0004] In actual use, firefighters' working conditions (such as wearing time, ambient temperature, humidity, wind speed, etc.) and physiological parameters (such as heart rate, skin temperature, respiratory rate, etc.) will change significantly, which poses a severe challenge to the performance of protective clothing. Therefore, a method that can comprehensively consider these factors and conduct a comprehensive evaluation of the performance of protective clothing is particularly important.

[0005] At present, although there are some methods for testing the performance of firefighting protective clothing for firefighters, most of these methods are limited to laboratory environments and cannot accurately reflect the performance of protective clothing in actual use; in addition, these methods often only focus on a single physiological parameter or environmental parameter, while ignoring the interaction and influence between these parameters. Therefore, developing a method that can comprehensively consider multiple factors and conduct a comprehensive and accurate evaluation of the performance of protective clothing is of great significance to improving the safety and work efficiency of firefighters. Summary of the invention

[0006] Based on the above problems, the present application provides a performance test data analysis method for firefighters' fire-fighting protective clothing. Through a systematic data analysis process, from the design classification of protective clothing to specific usage conditions and physiological parameter monitoring, and then to the final performance evaluation and improvement suggestions, a complete closed loop is formed, which can not only effectively evaluate the performance of existing protective clothing, but also provide a scientific basis for future protective clothing design.

[0007] The purpose of this application is achieved by the following technical solutions: On the one hand, the present application provides a method for analyzing performance test data of firefighters' firefighting protective clothing, the method comprising: S1. Classify protective clothing according to its design characteristics; S2. Collecting the working conditions and physiological parameters of the wearer of each category of protective clothing during each use, wherein the working conditions include wearing time and environmental parameters; S3. According to the operating conditions, a first time window is set to obtain a correlation coefficient between each physiological parameter; the physiological parameters are first grouped according to the correlation coefficient; through machine learning, a correlation model between the operating conditions under each first category and each first group is established, and a weight of each feature is obtained through the correlation model; S4. Select clustering features according to the weights of the features in the association model, and identify groups with similar characteristics through cluster analysis; S5. Combine the relationship model and cluster analysis results to obtain an analysis report and determine the improvement and usage suggestions for protective clothing.

[0008] Preferably, the physiological parameters include heart rate, skin temperature and respiratory rate; the environmental parameters include temperature, humidity and wind speed.

[0009] Preferably, S3 includes: According to the operating conditions, a first time window is set, and in the first time window, a correlation coefficient between any physiological parameter and other physiological parameters is obtained; According to the correlation coefficient, the physiological parameters are first grouped.

[0010] Preferably, obtaining the correlation coefficient between any physiological parameter and other physiological parameters in the first time window includes: ; in, is the total correlation coefficient between the ith physiological parameter and the jth physiological parameter, is the correlation coefficient between the ith physiological parameter and the jth physiological parameter in the tth first time window; n1 is the number of time windows in which the ith physiological parameter is positively correlated with the jth physiological parameter; n2 is the number of time windows in which the ith physiological parameter is negatively correlated with the jth physiological parameter; n=n1+n2; , is the weight, and abs() is the absolute value.

[0011] Preferably, the first grouping of the physiological parameters according to the correlation coefficient comprises: If the correlation coefficient of any two physiological parameters is greater than or equal to a preset threshold, the two physiological parameters are considered as a group; If the correlation coefficient between any two physiological parameters is less than a preset threshold, but their correlation coefficients with a third physiological parameter are greater than or equal to the preset threshold, then the two physiological parameters and the third physiological parameter are grouped together. If the correlation coefficient of any physiological parameter with other physiological parameters is less than the preset threshold, then the physiological parameter is taken as a separate group.

[0012] Preferably, according to the operating conditions, the first time window is set to include: ; wherein, is the first time window, is the basic time window; is the change rate of the k-th environmental parameter; is the preset change rate of the k-th environmental parameter; is the adjustment coefficient of the k-th environmental parameter, 0.1 < 2, is the non-linear index of the k-th environmental parameter, 0.5 < 3; is the wearing duration; is a constant, 0.01 < z < 0.1; v is the number of environmental parameters; is the preset minimum value of the window; is the preset maximum value of the window.

[0013] Preferably, the S3 further includes: Obtain the average change rate of each parameter under the first time window; the parameters include environmental parameters and physiological parameters; Input the environmental parameters, physiological parameters, and the average change rate of each parameter into a machine learning model; obtain the association model between the operating conditions under each first classification and each first group.

[0014] Preferably, the S4 includes: Sort the weights of the features in the association model; Obtain the key features according to the sorting result; perform clustering analysis on the key features to obtain groups of characteristics.

[0015] This application also provides an electronic device, which includes a memory and a processor. The memory stores a computer program, and when the processor executes the computer program, the steps of the method described in any item of this application are implemented.

[0016] In a third aspect, this application also provides a computer-readable storage medium, which stores computer instructions. When a computer reads the computer instructions, the computer executes the steps of the method described in any item of this application.

[0017] The beneficial effects of the present invention include: the method takes into account a variety of working conditions and physiological parameters of firefighters in the process of using protective clothing, including wearing time, ambient temperature, humidity, wind speed, as well as heart rate, skin temperature, respiratory rate, etc., so as to more comprehensively and accurately evaluate the performance of protective clothing. By setting the first time window and calculating the correlation coefficient between each physiological parameter, the method can capture the dynamic changes and mutual influence between the parameters, improve the accuracy and precision of data analysis, group the physiological parameters according to the correlation coefficient, and classify a group of closely related parameters so as to minimize the number of modeling while achieving more accurate modeling; through cluster analysis, the method can automatically identify groups with similar characteristics, providing a scientific basis and data support for the improvement of protective clothing. The method performs the first classification according to the design characteristics of the protective clothing, performs performance testing and data analysis on protective clothing of different categories, so that the evaluation results are more targeted and applicable. By analyzing the weights of the features in the association model, the method can identify the key factors affecting the performance of protective clothing, providing a clear direction for subsequent improvement and optimization. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is a schematic diagram of a performance test data analysis method for firefighters' firefighting protective clothing provided in an embodiment of the present application. DETAILED DESCRIPTION

[0019] Below, the present application is further described in conjunction with the accompanying drawings and specific implementation methods. It should be noted that, under the premise of no conflict, the various embodiments or technical features described below can be arbitrarily combined to form a new embodiment.

[0020] See also Figure 1 Some embodiments of the present application provide a method for analyzing performance test data of firefighter firefighting protective clothing, the method comprising: S1. Classify protective clothing according to its design characteristics; S2. Collecting the working conditions and physiological parameters of the wearer of each category of protective clothing during each use, wherein the working conditions include wearing time and environmental parameters; S3. According to the operating conditions, a first time window is set to obtain a correlation coefficient between each physiological parameter; the physiological parameters are first grouped according to the correlation coefficient; through machine learning, a correlation model between the operating conditions under each first category and each first group is established, and a weight of each feature is obtained through the correlation model; S4. Select clustering features according to the weights of the features in the association model, and identify groups with similar characteristics through cluster analysis; S5. Combine the relationship model and cluster analysis results to obtain an analysis report and determine the improvement and usage suggestions for protective clothing.

[0021] In some embodiments, the physiological parameters include heart rate, skin temperature, and respiratory rate; and the environmental parameters include temperature, humidity, and wind speed.

[0022] The working principle and effect of the above technical solution are: protective clothing is divided into different categories according to its design characteristics (such as material type, structural design, manufacturing process, etc.); the purpose is to ensure that subsequent analysis can be compared between protective clothing of the same type, reduce the interference caused by design differences, and make the analysis results more accurate and comparable.

[0023] Collect the wearer's working conditions and physiological parameters during each use of multiple protective clothing of each category; Record the wearer's operating conditions (such as wearing time, temperature, humidity, wind speed and other environmental parameters) and physiological parameters (such as heart rate, skin temperature, respiratory rate, etc.) during each use. Integrate data from different sources, including sensor data, wearer reports, etc., to obtain a comprehensive data set; the purpose is to provide a detailed data basis for subsequent analysis to ensure that the analysis results can reflect the actual usage scenarios and the wearer's actual reactions; According to the operating conditions, a first time window is set to obtain a correlation coefficient between each physiological parameter; the physiological parameters are first grouped according to the correlation coefficient; through machine learning, a correlation model between the operating conditions under each first category and each first group is established, and a weight of each feature is obtained through the correlation model; and used for feature selection in subsequent steps; According to the weights of the features in the association model, clustering features are selected, and groups with similar characteristics are identified through cluster analysis; based on the feature weights obtained in S3, the most important features are selected as the basis for cluster analysis; clustering algorithms (such as K-means, hierarchical clustering, etc.) are used to cluster the environmental performance of protective clothing and its wearers according to the selected features. Groups with similar characteristics are identified, which show similar performance characteristics and wearer physiological response patterns in specific environments. The purpose is to discover different types of protective clothing and their performance patterns in various environments, and provide specific directions for subsequent improvements.

[0024] S5: Combine the relationship model and cluster analysis results to obtain an analysis report and determine improvement suggestions for protective clothing.

[0025] Combined with the relationship model established in S3 and the cluster analysis results in S4, a comprehensive integrated analysis is conducted.

[0026] Improvement Suggestions Based on the analysis results, improvement suggestions are put forward according to the specific characteristics of each group, such as optimizing material selection, enhancing heat dissipation performance, improving breathability, etc.

[0027] Personalized configuration: Provide personalized protective clothing configuration suggestions based on the individual differences of different firefighters to ensure optimal safety and comfort.

[0028] Training and operation guides: Develop more scientific and reasonable training plans and operation guides based on group characteristics to help firefighters better cope with various working environments.

[0029] Through a systematic data analysis process, a complete closed loop is formed from the design classification of protective clothing to the specific use conditions and physiological parameter monitoring, and finally to the performance evaluation and improvement suggestions. This method can not only effectively evaluate the performance of existing protective clothing, but also provide a scientific basis for the design of future protective clothing, thereby improving the safety and work efficiency of firefighters.

[0030] In some embodiments, the S3 includes: According to the operating conditions, a first time window is set, and in the first time window, a correlation coefficient between any physiological parameter and other physiological parameters is obtained; According to the correlation coefficient, the physiological parameters are first grouped.

[0031] In some embodiments, obtaining the correlation coefficient between any physiological parameter and other physiological parameters in the first time window includes: ; in, is the total correlation coefficient between the ith physiological parameter and the jth physiological parameter, is the correlation coefficient of the ith physiological parameter to the jth physiological parameter in the tth first time window; n1 is the number of time windows in which the ith physiological parameter is positively correlated with the jth physiological parameter; n2 is the number of time windows in which the ith physiological parameter is negatively correlated with the jth physiological parameter; n=n1+n2; , is the weight, and abs() is the absolute value.

[0032] The working principle and effect of the above technical solution are as follows: first, a first time window is set according to working conditions (such as wearing time, ambient temperature, humidity, wind speed, etc.); the first time window is a time period for capturing and analyzing dynamic changes in physiological parameters; by setting a reasonable time window, it can be ensured that the physiological parameter data collected within a specific time period are representative and can reflect the physiological responses of firefighters under specific working conditions.

[0033] In the first time window, calculate the correlation coefficient of any physiological parameter with other physiological parameters; the correlation coefficient is used to measure the degree of correlation between two physiological parameters. Take a weighted average of the positive and negative correlation time windows respectively and add them together. Adjust the influence of positive and negative correlations by weight to reflect their different importance. Normalize the positive and negative correlation time windows to ensure that no matter how many positive or negative correlation time windows there are, their influence on the final correlation coefficient is relatively fair; ensure that the final correlation coefficient is positive, so that the direction of the correlation can be ignored and only its strength can be considered; retain the directional information of positive and negative correlations in the algorithm, and reflect the difference in importance of the two by setting different weights. If positive correlations are considered more important than negative correlations (for example, a positive correlation between heart rate and skin temperature may be more meaningful than a negative correlation), then w1 can be given a larger value.

[0034] According to the correlation coefficient, the physiological parameters are first grouped.

[0035] Based on the physiological parameter grouping, the association model is established using machine learning technology. The association model is used to describe the association between the working conditions and each physiological parameter grouping. By inputting environmental parameters, physiological parameters and the average change rate of each parameter into the machine learning model, the association model between the working conditions under each first category and each first grouping can be obtained; these models can reflect the dynamic change law and mutual correlation of physiological parameters under different working conditions.

[0036] In summary, by setting the first time window, calculating the correlation coefficient, grouping physiological parameters and establishing the correlation model, an in-depth analysis of the performance test data of firefighters' firefighting protective clothing is achieved. This process helps to reveal the correlation and dynamic change rules between physiological parameters, and provides a scientific basis for subsequent analysis and improvement.

[0037] In some embodiments, the first grouping of the physiological parameters according to the correlation coefficient includes: If the correlation coefficient of any two physiological parameters is greater than or equal to a preset threshold, the two physiological parameters are considered as a group; If the correlation coefficients of any two physiological parameters are less than a preset threshold, but the correlation coefficients of any two physiological parameters with a third physiological parameter are greater than or equal to the preset threshold, the two physiological parameters and the third physiological parameter are grouped together; If the correlation coefficient between any physiological parameter and other physiological parameters is less than the preset threshold, the physiological parameter is regarded as a group separately.

[0038] The working principle of the above technical solution is: if the correlation coefficient of any two physiological parameters is greater than or equal to the preset threshold, the two physiological parameters are grouped together; this helps to find out those parameters that directly affect each other. If the correlation coefficient of any two physiological parameters is less than the preset threshold, but the correlation coefficient of both of them with the third physiological parameter is greater than or equal to the preset threshold, the three physiological parameters are grouped together; in this case, the third parameter may act as a bridge, connecting the other two parameters, indicating that there may be indirect correlations or common influencing factors. If the correlation coefficient of any physiological parameter with all other physiological parameters is less than the preset threshold, the physiological parameter is grouped separately. This rule ensures that every parameter is taken into account, even if it has no obvious correlation.

[0039] The effect of the above technical solution is: by setting a preset threshold to determine whether two physiological parameters are closely related, the correlation pattern between different physiological parameters can be captured more accurately, ensuring that only those parameters with statistically significant correlation are classified into the same group, thereby avoiding misjudgment and unnecessary complexity.

[0040] In addition to direct correlation, indirect correlation is also considered. If the correlation coefficient of any two physiological parameters is less than the preset threshold, but they are both strongly correlated with the third physiological parameter, the three parameters are grouped together. This helps to identify potential common influencing factors or mediating variables and provide a more comprehensive physiological parameter relationship map.

[0041] Parameters whose correlation coefficients with all other physiological parameters are less than a preset threshold are grouped together to ensure that every parameter is considered even if it has no obvious correlation. This is very useful for discovering important parameters that may vary independently under certain conditions.

[0042] By rationally grouping physiological parameters, the subsequent data analysis process can be simplified. The parameters within each group usually show similar change trends or interaction patterns, making subsequent modeling and prediction more efficient and accurate. For example, in a machine learning model, each group is input as a whole feature, which can reduce the feature dimension and improve the training efficiency and interpretability of the model.

[0043] The grouped results are more interpretable and easier for field experts to understand and verify. By clarifying which parameters are closely related, which parameters are indirectly related, and which parameters are relatively isolated, we can better understand the behavior patterns of different physiological parameters under specific environmental conditions, thereby providing a scientific basis for the design and optimization of protective clothing.

[0044] Based on the grouping results, personalized improvement suggestions can be put forward for different groups. For example, for those groups where heart rate and skin temperature are highly correlated in high-temperature environments, optimizing heat dissipation performance can be recommended; while for those groups where breathing frequency is relatively isolated, improving air circulation design can be considered. Such personalized adjustments help improve the overall performance of the protective clothing and the comfort of the wearer.

[0045] This method is not only applicable to static data analysis but also can dynamically adapt to changes in environmental conditions. As new data accumulates and environmental conditions change, by recalculating the correlation coefficients and updating the grouping, it ensures that the analysis results always reflect the latest actual situation. This enables the performance evaluation and improvement of the protective clothing to continuously follow the changes in actual applications.

[0046] In some embodiments, setting the first time window according to the operating conditions includes: ; where, is the first time window, is the base time window; is the change rate of the k-th environmental parameter; is the preset change rate of the k-th environmental parameter; is the adjustment coefficient of the k-th environmental parameter, 0.1 < 2, is the non-linear exponent of the k-th environmental parameter, 0.5 < 3; is the wearing duration; is a constant, 0.01 < z < 0.1; v is the number of environmental parameters; is the preset minimum value of the window; is the preset maximum value of the window.

[0047] The working principle and effect of the above technical solution are: Dynamically adjusting the first time window according to the operating conditions aims to ensure that the time window can both sensitively reflect the impact of environmental changes on the physiological state of the wearer and will not become too large or too small. The ratio of the actual environmental parameter change rate to the preset change rate, which reflects the intensity of the current environmental conditions relative to the standard conditions; Controls the non-linear degree of the influence of the environmental parameter change rate on the time window length. A larger value means that a more剧烈 change rate will cause the time window to shorten faster; Adjusts the importance of each environmental parameter. A larger ) makes the time window gradually shorten but not excessively as the wearing time increases; when the environmental conditions become extreme, the length of the time window will decrease rapidly but will not approach zero indefinitely; this not only improves the flexibility and adaptability of the analysis, but also provides a solid data foundation for subsequent protective clothing performance evaluation and improvement suggestions.

[0048] In some embodiments, S3 further includes: Obtaining the average change rate of each parameter in the first time window; the parameters include environmental parameters and physiological parameters; The environmental parameters, physiological parameters and the average change rate of each parameter are input into the machine learning model; and the association model between the operating conditions under each first category and each first group is obtained.

[0049] The working principle of the above technical solution is as follows: Data collection and processing: In each first time window, the measurement values ​​of all environmental parameters (such as temperature, humidity, wind speed, etc.) and physiological parameters (such as heart rate, skin temperature, respiratory rate, etc.) are collected.

[0050] For each parameter, calculate its average rate of change in each time window : ; is the value of the parameter at the end of the time window; is the value of the parameter at the beginning of the time window; is the length of the time window.

[0051] Environmental parameters, physiological parameters, and the average rate of change of each parameter are input into the machine learning model.

[0052] All collected environmental parameters, physiological parameters and their corresponding average change rates are input into the machine learning model as features. These features may include but are not limited to: Environmental parameters: temperature, humidity, wind speed, etc.

[0053] Physiological parameters: heart rate, skin temperature, respiratory rate, etc.

[0054] Average rate of change: The average rate of change of the above parameters.

[0055] Model training is performed using machine learning algorithms (such as linear regression, decision tree, random forest, support vector machine, neural network, etc.). The training goal is to establish an association model between the operating conditions under each first classification and each first grouping.

[0056] In addition to the original parameters, you can also construct composite features, such as ratios and interaction terms between different parameters, to capture more complex relationships. Standardize or normalize the data to ensure that each feature has a fair impact on the model; handle missing values ​​and outliers to ensure data quality. Select a suitable machine learning algorithm and evaluate the model performance through methods such as cross-validation to ensure that it has good generalization ability.

[0057] Obtain the association model between the working conditions under each first category and each first group. Through the trained machine learning model, for each first category (i.e. protective clothing type) and first group (i.e. physiological parameter group), the corresponding association model between the working conditions and physiological parameters is established. Using the trained model, it is possible to predict what kind of physiological response pattern the wearers of different types of protective clothing may show under specific working conditions. This helps to understand the performance differences of different protective clothing under various environmental conditions. Based on the results of the model, personalized improvement suggestions can be made for different types of protective clothing to optimize their design and function.

[0058] By obtaining the average change rate of each parameter in the first time window and inputting these parameters and their change rates into the machine learning model, we finally established an association model between the working conditions under each first category and each first grouping. This can not only more accurately capture the changing trends of environmental and physiological parameters, but also provide a scientific basis for subsequent protective clothing performance evaluation and improvement suggestions.

[0059] In some embodiments, the S4 includes: Sort the weights of features in the association model; The key features are obtained based on the sorting results; the key features are clustered and analyzed to obtain groups of features.

[0060] The working principle of the above technical solution is: according to the sorting results, the first H features with the highest weights are selected as key features. The selection of H can be flexibly adjusted according to the specific application scenarios and requirements. For example, the top 10% or top 20% of the features can be selected. You can also set a preset weight threshold and define the features with weights higher than the threshold as key features; this method can be more flexible to adapt to different data sets and analysis objectives. By selecting key features, the complexity of subsequent analysis can be reduced, the computational efficiency can be improved, and the analysis results can be more focused on the most important and influential features. By screening key features, the problem of model overfitting can be effectively avoided and the generalization ability and reliability of the model can be improved.

[0061] Select a clustering algorithm (such as K-means, hierarchical clustering, DBSCAN, Gaussian Mixture Models, etc.) to cluster the environmental performance of protective clothing and its wearers according to key features. Through cluster analysis, groups with similar characteristics are identified. These groups show similar performance characteristics and physiological response patterns of wearers under specific conditions. Before clustering, the key features are standardized or normalized to ensure that each feature has a fair impact on the clustering results. The optimal number of clusters is determined by methods such as the Elbow Method, Silhouette Score, and Gap Statistics to ensure the stability and rationality of the clustering results. Use cross-validation or other internal validation methods to evaluate the stability of the clustering results, and combine external validation (such as comparison with known classification information) to ensure the effectiveness and accuracy of the clustering. "Group" refers to a group of protective clothing or wearers that show similar performance characteristics and physiological response patterns under specific environmental conditions; specifically, the group can include the following meanings: 1. Type of protective clothing Design similarity: Protective clothing of the same type (e.g., based on materials, structural design, etc.) may exhibit similar performance characteristics in the same environment. For example, some types of protective clothing may be more effective in keeping the wearer's body temperature stable in high-temperature and high-humidity environments.

[0062] 2. Performance under environmental conditions Environmental adaptability: Different protective clothing may show similar adaptability or limitations under the same environmental conditions (such as high temperature, high humidity, strong wind, etc.). For example, some protective clothing may be more likely to cause the wearer's heart rate to increase in high temperature environments.

[0063] 3. The wearer’s physiological response Physiological parameter patterns: Different protective clothing may cause the wearer to have similar physiological response patterns (such as changes in heart rate, skin temperature, and respiratory rate). For example, under the same working conditions, some protective clothing may cause the wearer to have similar heart rate acceleration or skin temperature increase.

[0064] 4. Working conditions Duration and intensity of work: Certain protective clothing may exhibit consistent performance characteristics under specific duration and intensity of work; for example, certain protective clothing may have similar effects on the wearer's breathing rate during long-term work.

[0065] For example: Group A: The protective clothing type is heavy protective clothing. When used in a high temperature and high humidity environment, the wearer's heart rate increases significantly and the skin temperature is also high.

[0066] Group B: The protective clothing type is light protective clothing. When used in a low-temperature and dry environment, the wearer's breathing rate is relatively stable and the skin temperature does not change much.

[0067] Group C: The protective clothing type is medium protective clothing. When used in a medium temperature and humidity environment, the wearer's physiological parameters are relatively stable, but as the working time increases, the heart rate gradually increases.

[0068] For example, the ratio of heart rate to skin temperature is calculated as a new feature to capture the relative changes between the two.

[0069] Using a combination of K-means and hierarchical clustering, we first quickly find the initial cluster centers through K-means, and then further optimize these centers with hierarchical clustering to ensure the robustness of the clustering results. We use t-SNE to project all samples onto a two-dimensional plane to intuitively display the distribution between different groups; for a cluster group that exhibits higher skin temperature but lower respiratory rate, it is recommended to increase the heat dissipation performance of the protective clothing.

[0070] The present application also provides an electronic device, which includes a memory and a processor, wherein the memory stores a computer program, and when the processor executes the computer program, the steps of any method described in the present application are implemented.

[0071] The present application also provides a computer-readable storage medium, which stores computer instructions. When a computer reads the computer instructions, the computer executes the steps of any method described in the present application.

[0072] This application is explained from the perspectives of purpose of use, effectiveness, progress and novelty, and has met the functional enhancement and usage requirements emphasized by the Patent Law. The above description and drawings of this application are only the preferred embodiments of this application, and are not intended to limit this application. Therefore, all structures, devices, features, etc. that are similar or identical to this application, that is, all equivalent replacements or modifications made in accordance with the scope of the patent application of this application, should fall within the scope of protection of the patent application of this application.

Claims

1. A performance test data analysis method for firefighters' firefighting protective clothing, characterized in that: The method comprises: S1. Classify protective clothing according to its design characteristics; S2. Collecting the working conditions and physiological parameters of the wearer of each category of protective clothing during each use, wherein the working conditions include wearing time and environmental parameters; S3. According to the operating conditions, a first time window is set to obtain a correlation coefficient between each physiological parameter; the physiological parameters are first grouped according to the correlation coefficient; through machine learning, a correlation model between the operating conditions under each first category and each first group is established, and a weight of each feature is obtained through the correlation model; S4. Select clustering features according to the weights of the features in the association model, and identify groups with similar characteristics through cluster analysis; S5. Combine the relationship model and cluster analysis results to obtain an analysis report and determine the improvement and usage suggestions for protective clothing.

2. The method according to claim 1, characterized in that: The physiological parameters include heart rate, skin temperature and respiratory rate; the environmental parameters include temperature, humidity and wind speed.

3. The method according to claim 1, characterized in that The S3 includes: According to the operating conditions, a first time window is set, and in the first time window, a correlation coefficient between any physiological parameter and other physiological parameters is obtained; According to the correlation coefficient, the physiological parameters are first grouped.

4. The method according to claim 3, characterized in that The step of obtaining a correlation coefficient between any physiological parameter and other physiological parameters in the first time window; include: ； in, is the total correlation coefficient between the ith physiological parameter and the jth physiological parameter, is the correlation coefficient between the ith physiological parameter and the jth physiological parameter in the tth first time window; n1 is the number of time windows in which the ith physiological parameter is positively correlated with the jth physiological parameter; n2 is the number of time windows in which the ith physiological parameter is negatively correlated with the jth physiological parameter; n=n1+n2; , is the weight, and abs() is the absolute value.

5. The method according to claim 3, characterized in that: The first grouping of the physiological parameters according to the correlation coefficient includes: If the correlation coefficient of any two physiological parameters is greater than or equal to a preset threshold, the two physiological parameters are considered as a group; If the correlation coefficients of any two physiological parameters are less than a preset threshold, but the correlation coefficients of any two physiological parameters with a third physiological parameter are greater than or equal to the preset threshold, the two physiological parameters and the third physiological parameter are grouped together; If the correlation coefficient between any physiological parameter and other physiological parameters is less than the preset threshold, the physiological parameter is regarded as a group separately.

6. The method according to claim 1, characterized in that According to the operating conditions, setting the first time window includes: ； Among them, is the first time window, is the base time window; is the change rate of the k-th environmental parameter; is the preset change rate of the k-th environmental parameter; is the adjustment coefficient of the k-th environmental parameter, 0.1 < 2, is the non-linear exponent of the k-th environmental parameter, 0.5 < 3; is the wearing duration; is a constant, 0.01 < z < 0.1; v is the number of environmental parameters; is the preset minimum value of the window; is the preset maximum value of the window.

7. The method according to claim 1, characterized in that The S3 further includes: Obtaining the average change rate of each parameter in the first time window; the parameters include environmental parameters and physiological parameters; The environmental parameters, physiological parameters and the average change rate of each parameter are input into the machine learning model; and the association model between the operating conditions under each first category and each first group is obtained.

8. The method according to claim 1, characterized in that The S4 includes: Sort the weights of features in the association model; The key features are obtained based on the sorting results; the key features are clustered and analyzed to obtain groups of features.

9. An electronic device, characterized in that: The electronic device comprises a memory and a processor, the memory stores a computer program, and the processor implements the steps of the method according to any one of claims 1 to 8 when executing the computer program.

10. A computer-readable storage medium, characterized in that: The storage medium stores computer instructions, and when a computer reads the computer instructions, the computer executes the steps of any one of the methods described in claims 1-8.

Images