High-altitude physical sign monitoring alarm decision-making method and device

Abstract

The invention provides a high-altitude physical sign monitoring alarm decision-making method and a high-altitude physical sign monitoring alarm decision-making device. The method comprises the following steps: acquiring high-altitude environment parameters, vital sign parameters and historical alarm state data, calculating a training set data mean value and a standard deviation, and carrying out standardization processing on training set and test set feature data; and based on the standardized training set feature data, constructing a linear kernel function support vector machine model, and generating a predicted alarm state of the standardized test set feature data. And determining a support vector machine model meeting a classification performance threshold value by comparing the matching degree between the predicted state and the real alarm state. And inputting real-time monitoring data into the model, and outputting an alarm signal. According to the invention, various parameters are comprehensively considered through the support vector machine, an intelligent alarm decision model is constructed, accurate analysis and processing of monitoring data are realized, and the accuracy and intelligent level of alarm decision are improved.

Description

Technical Field

[0001] This application belongs to the field of high-altitude rescue, and particularly relates to a high-altitude physical sign monitoring, alarming and decision-making method and device. Background Art

[0002] Due to its unique natural conditions, such as low oxygen, low temperature, low pressure, etc., the high-altitude environment has a significant impact on human physiological functions. Long-term or short-term exposure to the high-altitude environment may cause a series of adaptive reactions in the human body, such as hypoxemia, altitude sickness (including symptoms such as headache, nausea, fatigue, insomnia, etc.), and may even develop into more serious altitude diseases, such as high-altitude pulmonary edema and high-altitude cerebral edema. If these conditions are not monitored in a timely manner and effectively treated, they will pose a serious threat to the life and health of personnel. Therefore, it is particularly important to develop an efficient and accurate high-altitude vital sign monitoring system.

[0003] However, at present, due to the interference of extreme environmental factors in the high-altitude environment, traditional vital sign monitoring methods often lead to a decrease in the accuracy of monitoring data and cannot accurately reflect the vital sign status of personnel. At the same time, most of the existing alarming and decision-making methods are based on simple threshold judgments and lack intelligent decision-making capabilities for the characteristics of the high-altitude environment, resulting in low alarm accuracy and even possible false alarms or missed alarms. Summary of the Invention

[0004] The purpose of this application is to overcome the defects in the above-mentioned prior art and provide a high-altitude physical sign monitoring, alarming and decision-making method and device.

[0005] This application provides a high-altitude physical sign monitoring, alarming and decision-making method, including: Obtaining high-altitude environmental parameters, vital sign parameters and historical alarm status data; Calculating the mean and standard deviation of the training set data based on the environmental parameters and vital sign parameters; Performing standardization on the training set and test set feature data according to the mean and standard deviation; Constructing a linear kernel function support vector machine model based on the standardized training set feature data; Generating a predicted alarm status of the standardized test set feature data according to the support vector machine model; Determining a support vector machine model that meets the classification performance threshold according to the matching degree between the predicted alarm status and the real alarm status; Inputting real-time monitoring data into the support vector machine model and outputting an alarm signal.

[0006] Optionally, the high-altitude environmental parameters include: temperature, pressure and oxygen concentration; the vital sign parameters include heart rate, blood pressure and blood oxygen saturation.

[0007] Optionally, based on the standardized training set feature data, a linear kernel support vector machine model is constructed, including: The support vectors are solved by the sequential minimal optimization algorithm, and only the support vector parameters are retained for real-time alarm decision-making.

[0008] Optionally, the classification performance threshold is: the F1 score is greater than 0.7.

[0009] Optionally, the alarm signal includes at least one of an acoustic warning, an optical flash, and an electrical pulse signal.

[0010] This application also provides a high-altitude physical sign monitoring and alarm decision-making device, including: An acquisition module that acquires high-altitude environmental parameters, vital sign parameters, and historical alarm status data; A calculation module that calculates the mean and standard deviation of the training set data based on the environmental parameters and vital sign parameters; A standard module that standardizes the training set and test set feature data according to the mean and standard deviation; A construction module that constructs a linear kernel support vector machine model based on the standardized training set feature data; A training module that generates a predicted alarm status of the standardized test set feature data according to the support vector machine model; A screening module that determines a support vector machine model that meets the classification performance threshold according to the matching degree between the predicted alarm status and the true alarm status; An alarm module that inputs real-time monitoring data into the support vector machine model and outputs an alarm signal.

[0011] Optionally, the high-altitude environmental parameters include: temperature, pressure, and oxygen concentration; the vital sign parameters include heart rate, blood pressure, and blood oxygen saturation.

[0012] Optionally, the construction module constructs a linear kernel support vector machine model based on the standardized training set feature data, including: The support vectors are solved by the sequential minimal optimization algorithm, and only the support vector parameters are retained for real-time alarm decision-making.

[0013] Optionally, the classification performance threshold is: the F1 score is greater than 0.7.

[0014] Optionally, the alarm signal includes at least one of an acoustic warning, an optical flash, and an electrical pulse signal.

[0015] The beneficial effects of this application are: The present application provides a high-altitude physical sign monitoring and alarm decision-making method, including: obtaining high-altitude environmental parameters, vital sign parameters, and historical alarm status data; calculating the mean and standard deviation of the training set data based on the environmental parameters and vital sign parameters; performing standardization on the training set and test set feature data according to the mean and standard deviation; constructing a linear kernel function support vector machine model based on the standardized training set feature data; generating a predicted alarm status of the standardized test set feature data according to the support vector machine model; determining a support vector machine model that meets the classification performance threshold according to the matching degree between the predicted alarm status and the true alarm status; inputting real-time monitoring data into the support vector machine model and outputting an alarm signal. By introducing a support vector machine, the present application comprehensively considers high-altitude environmental parameters and vital sign parameters and constructs an intelligent alarm decision-making model. This model realizes accurate analysis and processing of monitoring data, improves the accuracy and intelligence level of alarm decision-making, effectively reduces false alarms and missed alarms, and provides a more reliable vital sign monitoring service for personnel in high-altitude environments. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 is a schematic diagram of the high-altitude physical sign monitoring and alarm decision-making method of the present application. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0017] Exemplary embodiments of the present disclosure will be described in more detail below with reference to the accompanying drawings. Although the exemplary embodiments of the present disclosure are shown in the drawings, it can be understood that the present disclosure can be implemented in various forms and should not be limited by the embodiments set forth herein. On the contrary, the embodiments are provided so that the present disclosure can be more thoroughly understood and the scope of the present disclosure can be fully conveyed to those skilled in the art.

[0018] The present application provides a high-altitude physical sign monitoring and alarm decision-making method, including: S101. Obtain high-altitude environmental parameters, vital sign parameters, and historical alarm status data; The environmental parameters, including temperature, pressure, and oxygen concentration, and the vital sign parameters, including heart rate, blood pressure, and blood oxygen saturation, are collected in real time through sensors.

[0019] The collected data is organized into a standard data set format.

[0020] Specifically, the collected environmental parameters (temperature, pressure, oxygen concentration), vital sign parameters (heart rate, blood pressure, blood oxygen saturation), and the corresponding alarm status (whether to alarm) data are organized into a data set :

[0021] where is the number of samples in the data set.

[0022] The group of data is:

[0023] Among them, corresponds to the temperature attribute T of the th group of data; corresponds to the pressure attribute P of the th group of data; corresponds to the oxygen concentration attribute of the th group of data ; corresponds to the heart rate attribute of the th group of data ; corresponds to the blood pressure attribute of the th group of data ; corresponds to the blood oxygen saturation attribute of the th group of data ; corresponds to the alarm status attribute A of the th group of data.

[0024] It is represented by unalarmed (0) or alarmed (1) respectively.

[0025] Taking the temperature T, pressure P, oxygen concentration , heart rate , blood pressure , blood oxygen saturation as the feature set , and taking the alarm status A as the classification label corresponding to the feature set . .

[0026] Dividing the feature set into training set features and test set features .

[0027] Dividing the classification label into the classification label corresponding to the training set features and the classification label corresponding to the test set features .

[0028] In this application, setting the quantity to 80% of the total number of dataset samples, that is ; setting the quantity to 20% of the total number of dataset samples, that is .

[0029] S102. Calculate the mean and standard deviation of the training set data based on the environmental parameters and vital sign parameters; Calculate the mean of each attribute in the training set feature data respectively. , .

[0030] The expression is:

[0031] where is the value of the th group of data for the th

[0032] Calculate the standard deviation of each attribute in the training set feature data respectively. .

[0033] The expression is:

[0034] where is the value of the th group of data for the

[0035] S103. Standardize the training set and test set feature data according to the mean and standard deviation; Standardize the training set and test set feature data with the mean and standard deviation of the training set data respectively, ensuring that all attributes in the training set feature data and test set feature data are on the same scale, avoiding some large-range features from dominating the model and accelerating the convergence of the algorithm, thereby improving the performance and stability of the model.

[0036] Calculate the standardized value of the value of each attribute in the training set feature data respectively. .

[0037] The expression is:

[0038] where is the

[0039] th feature value of the th sample in the training set.

[0040] The expression is:

[0041] where, is the j-th eigenvalue of the t-th sample in the test set. The test set and the training set are split from the same dataset, and the mean value of each attribute in their feature data can be regarded as equal.

[0042] S104. Based on the standardized training set feature data, construct a linear kernel support vector machine model; Select a linear kernel to create an SVM model, that is, find an optimal hyperplane in the feature space so that the hyperplane can completely separate the two types of samples of the un-alarm (0) and alarm (1) of the alarm status attribute (A), and the distances (i.e., margins) from the two types of samples to the hyperplane are the largest.

[0043] First, in a two-dimensional scenario, the decision surface can be regarded as a straight line, and its equation can be expressed as:

[0044] where, is the normal vector of the straight line, is the intercept of the straight line.

[0045] For the training set feature data any sample point as a feature vector, the distance from it to the decision surface can be expressed as:

[0046] where, is the normal vector of the norm.

[0047] The goal of SVM is to find a classifier that maximizes the minimum distance from the two types of samples of the un-alarm (0) and alarm (1) of the alarm status attribute (A) to the decision surface, that is, to maximize the classification margin.

[0048] To maximize the classification margin, construct the following objective function:

[0049] where, is the label of the sample point and takes a value of 1 or -1.

[0050] Since the parameters and can be scaled arbitrarily without changing the hyperplane, the objective function is transformed into:

[0051] Set the regularization parameter to control the complexity and generalization ability of the model.

[0052] Train the model: To solve the above constrained optimization problem, introduce the Lagrangian function:

[0053] where are Lagrange multipliers, and .

[0054] Using Lagrangian duality, transform the above optimization problem into a dual problem:

[0055] Solve the above dual problem through the Sequential Minimal Optimization (SMO) algorithm to obtain the optimal α value. According to the optimal α value, construct the decision function:

[0056] where is the sample point to be classified, and b is the intercept calculated through the support vectors.

[0057] Input the standardized training set feature data and the corresponding label data into the SVM classifier. Solve the dual problem of SVM through the SMO algorithm to find the optimal classification hyperplane and the corresponding support vectors.

[0058] S105. Generate the predicted alarm status of the standardized test set feature data according to the support vector machine model; Use the trained SVM model to predict the standardized test set feature data to obtain the predicted classification label .

[0059] S106. Determine the support vector machine model that meets the classification performance threshold according to the matching degree between the predicted alarm status and the true alarm status; According to and calculate the Precision, Recall, and F1-Score metrics of the SVM model.

[0060] The expression of Precision is:

[0061] Among them, TP is the true positive, which is the number of samples that the model correctly predicts as positive; FP is the false positive, which is the number of negative samples that the model wrongly predicts as positive.

[0062] The recall rate expression is:

[0063] Among them, FN is the false negative, which is the number of samples that are actually positive but are wrongly predicted as negative by the model.

[0064] The F1-score calculation formula is:

[0065] The value range of the F1-score is from 0 to 1, and the closer the value is to 1, the higher the prediction accuracy of the model.

[0066] In the present invention, if the F1-score is higher than 0.7, it is considered that the SVM model has a good performance and meets the alarm decision usage scenario.

[0067] S107. Input the real-time monitoring data into the support vector machine model and output an alarm signal.

[0068] In practical applications, the trained SVM model is used to make alarm decisions on new environmental parameter and vital sign parameter data. This includes reading new data, normalizing the new data using the mean and standard deviation calculated from the training set, and then using the SVM model for classification decisions.

[0069] If the classification result is 0, no alarm is issued; if the classification result is 1, an alarm signal is sent to the user through sound, light, electricity, etc.

[0070] This application also provides a high-altitude sign monitoring and alarm decision device, including: An acquisition module that acquires high-altitude environmental parameters, vital sign parameters, and historical alarm status data; A calculation module that calculates the mean and standard deviation of the training set data based on the environmental parameters and vital sign parameters; A standard module that performs standardization on the training set and test set feature data according to the mean and standard deviation; A construction module that constructs a linear kernel function support vector machine model based on the standardized training set feature data; A training module that generates a predicted alarm status for the standardized test set feature data according to the support vector machine model; A screening module that determines a support vector machine model that meets the classification performance threshold according to the matching degree between the predicted alarm status and the real alarm status; An alarm module that inputs the real-time monitoring data into the support vector machine model and outputs an alarm signal.

[0071] Further, the high-altitude environmental parameters include: temperature, pressure, and oxygen concentration; the vital sign parameters include heart rate, blood pressure, and blood oxygen saturation.

[0072] Further, the building block constructs a linear kernel function support vector machine model based on the standardized training set feature data, including: Solving for support vectors through the sequential minimal optimization algorithm and only retaining the support vector parameters for real-time alarm decision-making.

[0073] Further, the classification performance threshold is: the F1 score is greater than 0.7.

[0074] Further, the alarm signal includes at least one of an acoustic warning, an optical flash, and an electrical pulse signal.

[0075] The above description of the embodiments is to enable those of ordinary skill in the art to understand and apply the present invention. It is obvious that those skilled in the art can easily make various modifications to the above embodiments and apply the general principles described herein to other embodiments without creative efforts. Therefore, the present invention is not limited to the above embodiments, and the improvements and modifications made by those skilled in the art to the present invention should be within the protection scope of the present invention according to the disclosure of the present invention.

Claims

1. A high altitude physical sign monitoring alarm decision method, characterized in that: include: Obtain high-altitude environmental parameters, vital signs parameters and historical alarm status data; Based on the environmental parameters and vital sign parameters, calculating the mean and standard deviation of the training set data; Standardize the training set and test set feature data according to the mean and standard deviation; Based on the standardized training set feature data, a linear kernel function support vector machine model is constructed; Generate a predicted alarm state of the standardized test set feature data according to the support vector machine model; Determining a support vector machine model that meets a classification performance threshold according to a matching degree between the predicted alarm state and the actual alarm state; The real-time monitoring data is input into the support vector machine model and an alarm signal is output.

2. A high altitude physical sign monitoring alarm decision method according to claim 1, characterized in that: The high altitude environment parameters include: temperature, pressure and oxygen concentration; the vital sign parameters include heart rate, blood pressure and blood oxygen saturation.

3. A high altitude physical sign monitoring alarm decision method according to claim 1, characterized in that: Based on the standardized training set feature data, a linear kernel function support vector machine model is constructed, including: The support vector is solved by a sequential minimum optimization algorithm, and only the support vector parameters are retained for real-time alarm decision making.

4. The high altitude physical sign monitoring alarm decision method according to claim 1, characterized in that: The classification performance threshold is: F1 score greater than 0.

7.

5. The high altitude physical sign monitoring alarm decision-making method according to claim 1 is characterized in that: The alarm signal includes at least one of an acoustic alarm, an optical flash and an electrical pulse signal.

6. A high altitude physical sign monitoring alarm decision device, characterized in that: include: Acquisition module, to obtain high altitude environmental parameters, vital signs parameters and historical alarm status data; A calculation module, which calculates the mean and standard deviation of the training set data based on the environmental parameters and vital sign parameters; A standard module, performing standardization on the feature data of the training set and the test set according to the mean and the standard deviation; The construction module builds a linear kernel function support vector machine model based on the standardized training set feature data; A training module, generating a predicted alarm state of the standardized test set feature data according to the support vector machine model; A screening module, which determines a support vector machine model that meets a classification performance threshold according to a matching degree between the predicted alarm state and the actual alarm state; The alarm module inputs the real-time monitoring data into the support vector machine model and outputs an alarm signal.

7. The high altitude physical sign monitoring alarm decision-making device according to claim 6, characterized in that: The high altitude environment parameters include: temperature, pressure and oxygen concentration; the vital sign parameters include heart rate, blood pressure and blood oxygen saturation.

8. The high altitude physical sign monitoring alarm decision-making device according to claim 6, characterized in that: The construction module constructs a linear kernel function support vector machine model based on the standardized training set feature data, including: The support vector is solved by a sequential minimum optimization algorithm, and only the support vector parameters are retained for real-time alarm decision making.

9. The high altitude physical sign monitoring alarm decision-making device according to claim 6, characterized in that: The classification performance threshold is: F1 score greater than 0.

7.

10. The high altitude physical sign monitoring alarm decision-making device according to claim 6, characterized in that: The alarm signal includes at least one of an acoustic alarm, an optical flash and an electrical pulse signal.

Images