Heat stroke prevention and monitoring system based on awareness cognitive behavior feature recognition

Abstract

The invention discloses a heat stroke prevention and monitoring system based on consciousness cognitive behavior feature recognition, and belongs to the technical field of intelligent medical treatment. The problems that in the prior art, intelligent analysis is lacked, certain subjectivity exists, and the effectiveness of preventing and monitoring the heat stroke is difficult to ensure are solved, the physiological signals and behavior data of the patient are analyzed through the linear discriminant analysis model, and whether the patient suffers from the heat stroke or not is predicted; after it is predicted that the patient suffers from the heat stroke, a weighted average method is adopted to further calculate the risk index of the patient, and then different prevention and monitoring strategies are applied to the current patient according to the risk index; on the basis, whether the patient suffers from the heat stroke and the risk index of the patient can be effectively identified, so that targeted prevention measures can be taken, and further development of the disease is avoided; meanwhile, through intelligent analysis and judgment, the morbidity of heat stroke can be effectively reduced, consumption of medical resources is reduced, and social and economic burdens are reduced.

Description

Technical Field

[0001] The present invention relates to the field of intelligent medical technology, and in particular to a heat stroke prevention and monitoring system based on awareness, cognition and behavioral feature recognition. Background Art

[0002] Heat stroke is the most serious type of heat stroke, caused by imbalance in body temperature regulation caused by exposure to high temperature and / or strenuous physical labor. Its main characteristics are increased core body temperature (>40°C) and abnormalities in the central nervous system. The main types are classic heat stroke and exertional heat stroke. Common causes involve increased heat production, excessive heat acquisition, and heat dissipation disorders. Exertional heat stroke mainly occurs in healthy young people, often caused by heavy physical labor or strenuous exercise; while non-exertional heat stroke mainly occurs in the elderly, weak, and chronically ill patients, caused by passive exposure to hot environments. Based on this, it can be seen that heat stroke is an extremely serious heat stroke condition that can be life-threatening; therefore, prevention and monitoring of heat stroke are crucial.

[0003] However, in the existing technology, the prevention and monitoring of heat stroke is usually done by medical staff after routine examinations, and then judging whether the patient has heat stroke based on the examination results; but this method has certain subjective errors, such as: the same examination results may be analyzed by different medical staff, and the conclusions drawn may be different; this makes the judgment of heat stroke not only lack of intelligent analysis, but also difficult to ensure the effectiveness of heat stroke prevention and monitoring. Therefore, the existing needs are not met, and we propose a heat stroke prevention monitoring system based on the recognition of consciousness cognitive behavioral characteristics. Summary of the invention

[0004] The purpose of the present invention is to provide a heat stroke prevention and monitoring system based on the recognition of consciousness cognitive behavioral characteristics. The patient's physiological signals and behavioral data are analyzed through a linear discriminant analysis model to predict whether the patient suffers from heat stroke. After predicting that the patient suffers from heat stroke, the weighted average method is used to further calculate the risk index of the disease, and then different prevention and monitoring strategies are implemented for the current patient according to the level of the risk index. Through intelligent analysis, it is possible to effectively identify whether the patient suffers from heat stroke and the risk index of the disease, so that targeted preventive measures can be taken to avoid further development of the disease, thereby solving the problems raised in the above-mentioned background technology.

[0005] To achieve the above object, the present invention provides the following technical solutions: The heat stroke prevention and monitoring system based on the recognition of consciousness, cognition and behavioral characteristics includes: The data collection unit is used to collect the physiological signals and behavioral data of patients with heat stroke as the first type of data; and is also used to collect the physiological signals and behavioral data of normal patients as the second type of data; wherein the physiological signal data include: the patient's body surface temperature, brain waves, electrocardiogram and respiratory biofeedback signals; the behavioral data include: movement frequency, reaction speed and attention behavior indicators; The feature extraction unit is used to clean and reduce the dimension of the first-class data and the second-class data, and extract feature information reflecting the patient's physical state and metabolic level from the first-class data and the second-class data respectively; the feature information of patients with heat stroke is used as a feature vector, and the label is set to be positive; and the feature information of patients without heat stroke is used as a training sample, and the label is set to be negative; The feature recognition unit is used to use the feature vector and label as a training set, and the feature information and label of the second-class data as a test set; learn the feature vector and label content through the linear discriminant analysis model, and output the learning result; then analyze the feature information and label content of the second-class data through the learned linear discriminant analysis model, determine whether there are heat stroke features in the feature information of the second-class data, and output the prediction result; compare the prediction result with the actual result of the second-class data to verify the performance of the linear discriminant analysis model; The prevention and monitoring unit is used to predict whether the current patient suffers from heat stroke through a linear discriminant analysis model. If the patient suffers from heat stroke, the risk index of the current patient suffering from heat stroke is further calculated, and different prevention and monitoring strategies are implemented for the current patient according to the risk index.

[0006] Furthermore, the feature extraction unit includes: Data processing module, used for cleaning, denoising and normalizing the first and second category data; The feature extraction module is used to analyze the information content in the first and second category data, and extract the heart rate variability, breathing depth change and skin resistance as feature information reflecting the patient's physical state and metabolic level; The feature analysis module is used to compare the feature information of patients with heat stroke with the feature information of patients without heat stroke, and analyze whether the feature information of patients with heat stroke is different from the feature information of patients without heat stroke under the influence of heat stroke.

[0007] Furthermore, the feature analysis module analyzes whether the feature information of patients with heat stroke is different from that of patients without heat stroke under the influence of heat stroke, specifically, based on heart rate variability, breathing depth change and skin resistance, the feature vector of each patient is calculated respectively, and it is determined whether there is a difference in the feature vector of each patient; Calculation of heart rate variability eigenvector: The voltage at each heartbeat point in the ECG signal is matched to the original heart rate through a linear interpolation formula. For each heartbeat point, the time domain interval DT between two adjacent heartbeat points is calculated; for each heartbeat point, the power spectrum density estimate PSDE corresponding to each time domain interval DT is calculated; and then the frequency corresponding to the maximum value of the power spectrum density estimate PSDE of each heartbeat point is taken as the feature vector; Calculation of breathing depth change feature vector: The voltage at each respiratory cycle point in the respiratory signal is matched to the original respiratory cycle number through a linear interpolation formula. For each respiratory cycle point, the time domain interval DT between two adjacent respiratory cycle points is calculated; for each respiratory cycle point, the power spectrum density estimate PSDE corresponding to each time domain interval DT is calculated; and then the frequency corresponding to the maximum value of the power spectrum density estimate PSDE of each respiratory cycle point is taken as the feature vector; Calculation of skin resistance eigenvector: Each resistance value in the skin impedance signal is corresponded to the original skin impedance through a linear interpolation formula. For each skin impedance, the time domain interval DT between two adjacent skin impedances is calculated; for each skin impedance, the power spectrum density estimate PSDE corresponding to each time domain interval DT is calculated; and then the frequency corresponding to the maximum value of the power spectrum density estimate PSDE of each skin impedance is taken as the feature vector; The judgment basis is: patients with heat stroke have higher values ​​of characteristic vectors; patients without heat stroke have lower values ​​of characteristic vectors. Based on this, it is analyzed whether the characteristic information of patients with heat stroke is different from that of patients without heat stroke under the influence of heat stroke.

[0008] Furthermore, the feature recognition unit includes: The model building module is used to create a neural network model, input the training set into the neural network model, use the LDA linear discriminant analysis algorithm to train the model, and output the training results to obtain a linear discriminant analysis model; The model evaluation module is used to input the test set into the linear discriminant analysis model for testing, use the linear discriminant analysis model to identify whether there is characteristic information of heat stroke in the test set, and verify the performance of the linear discriminant analysis model based on the test results.

[0009] Furthermore, the feature recognition unit further includes: The feature interpretation module is used to analyze the meaning of each feature information after the training set and test set training are completed.

[0010] Furthermore, the preventive monitoring unit includes: The risk analysis module is used to collect the current patient's personal information as an additional factor and calculate the risk index of the current patient suffering from heat stroke using a weighted average method; The early warning module is used to build a risk threshold. If the patient's risk index exceeds the set threshold, it means that the patient is at high risk of heat stroke, and an alarm will be issued in time to remind medical staff to pay attention to the patient's condition.

[0011] Furthermore, the data collection unit includes: The real-time collection module is used to collect the patient's heart rate variability, breathing depth changes, and skin resistance data through health monitoring equipment as basic data for clinical diagnosis or daily health prevention monitoring.

[0012] Furthermore, the data collection unit further includes: Real-time monitoring of the patient's heart rate data value; Comparing the heart rate data value of the current patient with a preset heart rate threshold; When the heart rate data value of the current patient exceeds a preset heart rate threshold, the tidal volume corresponding to each breath of the current patient within a preset number of breaths is monitored; Obtaining a respiratory abnormality coefficient using the tidal volume corresponding to each breath of the current patient within a preset number of breaths; The abnormal breathing coefficient is obtained by the following formula:

[0013] Where R represents the respiratory abnormality coefficient; V p V represents the average tidal volume corresponding to each breath in the preset number of breaths; min represents the minimum tidal volume corresponding to each breath in the preset number of breaths; f represents the respiratory frequency of the current breath; f max Indicates the preset maximum allowed respiratory rate; H m Indicates the heart rate data value corresponding to abnormal heart rate; H y Indicates the preset heart rate threshold; The abnormal breathing coefficient is compared with a preset abnormal breathing coefficient threshold, and the collection frequency of heart rate variability and breathing depth changes is adjusted according to the comparison result.

[0014] Further, the abnormal breathing coefficient is compared with a preset abnormal breathing coefficient threshold, and the acquisition frequency of the heart rate variability and the breathing depth change is adjusted according to the comparison result, including: Comparing the abnormal breathing coefficient with a preset abnormal breathing coefficient threshold to obtain a comparison result; When the comparison result shows that the abnormal breathing coefficient is not lower than the preset abnormal breathing coefficient threshold, there is no need to adjust the collection frequency of heart rate variability and breathing depth change; When the comparison result shows that the abnormal breathing coefficient is lower than the preset abnormal breathing coefficient threshold, the abnormal breathing coefficient is used to adjust the acquisition frequency of the heart rate variability and the breathing depth change to obtain the adjusted acquisition frequency of the heart rate variability and the breathing depth change; The adjusted heart rate variability acquisition frequency is obtained by the following formula:

[0015] Among them, F h F represents the acquisition frequency of adjusted heart rate variability; h0 represents the acquisition frequency of heart rate variability before adjustment; R represents the respiratory abnormality coefficient; R y Indicates the preset breathing abnormality coefficient threshold; V p V represents the average tidal volume corresponding to each breath in the preset number of breaths; min V represents the minimum tidal volume corresponding to each breath in the preset number of breaths; max Indicates the maximum tidal volume corresponding to the tidal volume of each breath in the preset number of breaths; At the same time, the acquisition frequency of the adjusted breathing depth change is obtained by the following formula:

[0016] Among them, F t Indicates the sampling frequency of the adjusted breathing depth change; F t0 Indicates the sampling frequency of breathing depth changes before adjustment; V p V represents the average tidal volume corresponding to each breath in the preset number of breaths; max Indicates the maximum tidal volume corresponding to each breath in the preset number of breaths; H m Indicates the heart rate data value corresponding to abnormal heart rate; H y Indicates the preset heart rate threshold; According to the adjusted collection frequency of heart rate variability and breathing depth changes, the health monitoring equipment is controlled to collect data corresponding to the heart rate variability and breathing depth changes.

[0017] Furthermore, it also includes: The human-computer interaction unit is used to visualize the current patient's physiological signals and behavioral data as well as the prediction results and risk index of the linear discriminant analysis model.

[0018] The present invention analyzes the patient's physiological signals and behavioral data through a linear discriminant analysis model to predict whether the patient suffers from heat stroke; and after predicting that the patient suffers from heat stroke, a weighted average method is used to further calculate the risk index of the disease, and then different prevention and monitoring strategies are implemented for the current patient according to the level of the risk index; based on this, it is possible to effectively identify whether the patient suffers from heat stroke and the risk index of the disease, so that targeted preventive measures can be taken to avoid further development of the disease; at the same time, through intelligent analysis and judgment, the incidence of heat stroke can be effectively reduced, the consumption of medical resources can be reduced, and the social and economic burden can be reduced. BRIEF DESCRIPTION OF THE DRAWINGS

[0019] Figure 1 This is a module composition diagram of the heat stroke prevention and monitoring system based on consciousness cognitive behavior feature recognition of the present invention. DETAILED DESCRIPTION

[0020] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0021] In order to solve the problem that in the existing technology, the prevention and monitoring of heat stroke is usually done by medical staff after routine examination, and then judging whether the patient has heat stroke based on the examination results; however, this method has certain subjective errors, such as: the same examination results may be analyzed by different medical staff, and the conclusions drawn may be different; this makes the judgment of heat stroke not only lack intelligent analysis, but also difficult to ensure the effectiveness of heat stroke prevention and monitoring. Please refer to Figure 1 , this embodiment provides the following technical solutions: The heat stroke prevention and monitoring system based on the recognition of consciousness, cognition and behavioral characteristics includes: The data collection unit is used to collect physiological signals and behavioral data of patients with heat stroke as a first type of data; and is also used to collect physiological signals and behavioral data of normal patients as a second type of data; wherein the physiological signal data includes: patient's body surface temperature, brain wave, electrocardiogram and respiratory biofeedback signal; the data collection unit includes: The real-time collection module is used to collect the patient's heart rate variability, breathing depth changes, skin resistance and other physiological signals and behavioral data through health monitoring equipment, such as electrocardiographs, respiratory sensors or thermometers, as basic data for clinical diagnosis or daily health prevention monitoring. Specifically, an abnormal increase in the patient's body surface temperature is one of the main factors in determining whether the patient has heat stroke. During the process of temperature increase, the brain waves of heat stroke patients will show abnormal changes, such as decreased wakefulness, slower frequency and irregular rhythm. Heat stroke patients are under additional pressure on the heart and lungs, and electrocardiograms may show abnormalities, such as tachycardia, myocardial ischemia and myocardial necrosis. The respiratory signals of heat stroke patients will be disturbed, such as shallow breathing and unstable rhythm. Behavioral data include: movement frequency, reaction speed and attention behavior indicators; heat stroke is a serious central nervous system injury caused by long-term exposure to high temperature and high humidity environment. In the early stage of the disease, the body temperature may rise due to abnormal temperature regulation, resulting in a decrease in movement frequency; and the central nervous system is damaged due to the increase in body temperature, which changes its reaction speed, such as: gradually slowing down the reaction speed; heat stroke patients may also have problems such as inattention and memory loss during the course of the disease, which are all related to the patient's nervous system involvement; intelligent analysis of the above indicators can help medical staff better understand the pathophysiological mechanism of heat stroke and provide a basis for clinical diagnosis and treatment.

[0022] The feature extraction unit is used to clean and reduce the dimension of the first-class data and the second-class data, and extract feature information reflecting the patient's physical state and metabolic level from the first-class data and the second-class data respectively; the feature information of patients with heat stroke is used as a feature vector, and the label is set to be positive; and the feature information of patients without heat stroke is used as a training sample, and the label is set to be negative; the feature extraction unit includes: The data processing module is used to clean, denoise and normalize the first and second category data to facilitate subsequent analysis; specifically, by checking the syntax, semantics and logic of the first and second category data, the data records with obvious errors, duplications, missing or damaged data are removed to ensure the accuracy of the data; after cleaning the data, the data needs to be denoised and normalized to remove the noise and inconsistency hidden in the original data, and eliminate unit differences, scaling effects and outliers in the original data, so as to ensure the validity of the data.

[0023] The feature extraction module is used to analyze the information content in the first and second category data, and extract heart rate variability, breathing depth changes and skin resistance as feature information reflecting the patient's physical condition and metabolic level; specifically, by analyzing the heart rate variability, the patient's sympathetic nerve tension and parasympathetic nerve tension are understood, so as to judge the patient's physical condition; by analyzing the breathing depth, the patient's breathing state is understood, and then his physical condition is evaluated; by analyzing the skin resistance, the patient's metabolic level and physical state are understood; by extracting the above-mentioned effective feature information, a basis is provided for subsequent data analysis and decision-making.

[0024] The feature analysis module is used to compare the feature information of patients with heat stroke with that of patients without heat stroke, and analyze whether the feature information of patients with heat stroke is different from that of patients without heat stroke under the influence of heat stroke; specifically, based on heart rate variability, breathing depth change and skin resistance, the feature vector of each patient is calculated respectively, and it is determined whether there is a difference in the feature vector of each patient; Calculation of heart rate variability eigenvector: The voltage at each heartbeat point in the ECG signal is matched to the original heart rate through a linear interpolation formula. For each heartbeat point, the time domain interval DT between two adjacent heartbeat points is calculated; for each heartbeat point, the power spectrum density estimate PSDE corresponding to each time domain interval DT is calculated; and the frequency corresponding to the maximum value of the power spectrum density estimate PSDE of each heartbeat point is taken as the feature vector; in this embodiment, for example: the ECG signal at the point is interpolated with the ECG signal at the adjacent heartbeat point through a linear interpolation formula to obtain the heart rate corresponding to the voltage at the heartbeat point; then, the time domain interval DT between two adjacent heartbeat points is calculated. If the heart rate of a heartbeat point is f1, the heart rates corresponding to the previous heartbeat point and the next heartbeat point are: f0=f1-dt and f2=f1+dt respectively; Where f0 is the previous heartbeat point; f2 is the next heartbeat point; dt is the time interval between two adjacent heartbeat points; Secondly, the power spectral density estimate PSDE corresponding to each time domain interval DT is calculated by the Fourier transform algorithm. The power spectral density estimate PSDE corresponding to each time domain interval DT is obtained by converting the ECG signal from the time domain to the frequency domain and averaging the power spectral density of each heartbeat point.

[0025] Finally, the frequency corresponding to the maximum value of the power spectral density estimation PSDE of each heartbeat point is taken as the feature vector to better capture the high-frequency components of cardiac activity and thus improve the effect of feature extraction.

[0026] Calculation of breathing depth change feature vector: The voltage at each respiratory cycle point in the respiratory signal is corresponded to the original respiratory cycle number through a linear interpolation formula. For each respiratory cycle point, the time domain interval DT between two adjacent respiratory cycle points is calculated; for each respiratory cycle point, the power spectrum density estimate PSDE corresponding to each time domain interval DT is calculated; and the frequency corresponding to the maximum value of the power spectrum density estimate PSDE of each respiratory cycle point is taken as the feature vector; in this embodiment, for example: the respiratory signal at each respiratory cycle point is interpolated with the respiratory signal at the adjacent respiratory cycle point through a linear interpolation formula to obtain the original respiratory cycle number corresponding to the voltage at the respiratory cycle point; then, the time domain interval DT between two adjacent respiratory cycle points is calculated; if the respiratory cycle number at a respiratory cycle point is T0, the original respiratory cycle numbers corresponding to the previous respiratory cycle point and the next respiratory cycle point are respectively: T0-dt and T0+dt; wherein dt is the time interval between two adjacent respiratory cycle points; Secondly, the power spectral density estimate PSDE corresponding to each time domain interval DT is calculated by the Fourier transform algorithm. The power spectral density estimate PSDE corresponding to each time domain interval DT is obtained by converting the respiratory signal from the time domain to the frequency domain and averaging the power spectral density of each respiratory cycle point.

[0027] Finally, the frequency corresponding to the maximum value of the power spectral density estimation PSDE at each respiratory cycle point is taken as the feature vector to better capture the periodic changes of respiratory activity and thus improve the effect of feature extraction.

[0028] Calculation of skin resistance eigenvector: Each resistance value in the skin impedance signal is corresponded to the original skin impedance by a linear interpolation formula. For each skin impedance, the time domain interval DT between two adjacent skin impedances is calculated; for each skin impedance, the power spectrum density estimate PSDE corresponding to each time domain interval DT is calculated; and the frequency corresponding to the maximum value of the power spectrum density estimate PSDE of each skin impedance is taken as the feature vector; in this embodiment, for example: for each skin impedance, the linear interpolation formula is used to correspond its corresponding resistance value to the original skin impedance, and the original resistance value of each skin impedance can be obtained; then, the time domain interval DT between two adjacent skin impedances is calculated. If the value of a skin impedance is R0, the original skin impedances corresponding to the previous skin impedance and the next skin impedance are R0-dt and R0+dt respectively; wherein dt is the time interval between two adjacent skin impedances; Secondly, the power spectral density estimate PSDE corresponding to each time domain interval DT is calculated by the Fourier transform algorithm. The skin impedance signal is converted from the time domain to the frequency domain and the power spectrum of each skin impedance is averaged to obtain the power spectral density estimate PSDE corresponding to each time domain interval DT.

[0029] Finally, the frequency corresponding to the maximum value of the power spectral density estimation PSDE of each skin impedance is taken as the feature vector to better capture the changes in skin impedance and thus improve the effect of feature extraction.

[0030] The judgment basis is: patients with heat stroke have higher values ​​of their eigenvectors; patients without heat stroke have lower values ​​of their eigenvectors; based on this, it is analyzed whether the characteristic information of patients with heat stroke is different from that of patients without heat stroke under the influence of heat stroke; for example, for skin impedance: patients with heat stroke will have higher values ​​of their eigenvectors than those without heat stroke because heat stroke will cause changes in skin impedance, thereby changing the frequency in the eigenvector; based on this, by comparing the skin impedance eigenvectors of different patients, it is analyzed whether there are differences in the skin impedance of patients under the influence of heat stroke; if there is a significant difference between the eigenvectors of patients with heat stroke and those without heat stroke, it means that heat stroke may have an impact on this feature.

[0031] The feature recognition unit is used to use the feature vector and label as a training set and the feature information and label of the second type of data as a test set; learn the feature vector and label content through a linear discriminant analysis model and output the learning result; then analyze the feature information and label content of the second type of data through the learned linear discriminant analysis model to determine whether there are heat stroke features in the feature information of the second type of data and output the prediction result; compare the prediction result with the actual result of the second type of data to verify the performance of the linear discriminant analysis model; the feature recognition unit includes: The model building module imports the required databases, such as TensorFlow, Numpy, etc., sets model parameters, such as batch size and number of iterations, and loads training set data X and Y; creates a neural network model, defines the model structure, loss function and optimizer; divides the training set into k subsets and inputs them into the neural network model, uses the LDA linear discriminant analysis algorithm to train the model, and outputs the training results to obtain a linear discriminant analysis model; and during the training process, calculates the mean square error and finds the optimal model parameters.

[0032] The model evaluation module imports the trained linear discriminant analysis model and loads the feature data of the test set. The loaded test set is then input into the linear discriminant analysis model for testing. The test set features are converted into a format acceptable to the linear discriminant analysis model. The correlation matrix between the test set data and feature information is obtained by calculating the covariance matrix. The linear discriminant analysis model is then used to identify whether there is feature information of heat stroke in the test set. The performance of the linear discriminant analysis model is verified based on the test results for subsequent analysis and decision-making.

[0033] The feature interpretation module is used to analyze the meaning of each feature information after the training of the training set and the test set is completed; for example, changes in breathing depth may reflect changes in oxygen supply in heat stroke patients; increases in heart rate changes may indicate increased tension in the sympathetic nervous system.

[0034] The prevention monitoring unit is used to apply the verified linear discriminant analysis model to the actual scenario, and predict whether the current patient suffers from heat stroke through the linear discriminant analysis model. If the patient suffers from heat stroke, the risk index of the current patient suffering from heat stroke is further calculated, and different prevention monitoring strategies are implemented for the current patient according to the risk index. The prevention monitoring unit includes: The risk analysis module is used to collect the current patient's personal information as additional factors, such as age, gender, body temperature, and body mass index (BMI), clinical manifestations such as persistent fever, excessive sweating, and impaired consciousness, as well as whether the patient has other chronic diseases, whether the patient has undergone surgery recently, and whether the patient is taking specific medications. The weighted average method is used to calculate the patient's risk index for heat stroke, quantify the patient's various indicators and medical history, assign corresponding weights to them, and use the weighted average method to calculate the risk index to obtain a comprehensive assessment result, which helps medical staff judge the possibility of the patient suffering from heat stroke.

[0035] The early warning module is used to construct a risk threshold. If the patient's risk index exceeds the set threshold, it means that the patient is at high risk of heat stroke, and an alarm must be issued in time to remind medical staff to pay attention to the patient's condition; and the risk prediction results are converted into a heat map, and the darker the color, the higher the risk of disease.

[0036] The human-computer interaction unit is used to visualize the current patient's physiological signals and behavioral data as well as the prediction results and risk index of the linear discriminant analysis model.

[0037] The beneficial effects achieved by the above content: Through the above operations, it is possible to effectively identify whether the patient suffers from heat stroke and the risk index of the disease, so that targeted preventive measures can be taken to avoid further development of the disease; at the same time, through intelligent analysis and judgment, it is possible to effectively reduce the incidence of heat stroke, reduce the consumption of medical resources, and reduce the socioeconomic burden.

[0038] Working principle: The patient's physiological signals and behavioral data are analyzed through a linear discriminant analysis model to predict the possibility of the patient suffering from heat stroke. After predicting heat stroke, the weighted average method is used to further calculate the risk index of the disease, and then different prevention and monitoring strategies are implemented for the current patient based on the risk index.

[0039] Specifically, the data collection unit further includes: Real-time monitoring of the patient's heart rate data value; Comparing the heart rate data value of the current patient with a preset heart rate threshold; When the heart rate data value of the current patient exceeds a preset heart rate threshold, the tidal volume corresponding to each breath of the current patient within a preset number of breaths is monitored; Obtaining a respiratory abnormality coefficient using the tidal volume corresponding to each breath of the current patient within a preset number of breaths; The abnormal breathing coefficient is obtained by the following formula:

[0040] Where R represents the respiratory abnormality coefficient; V p V represents the average tidal volume corresponding to each breath in the preset number of breaths; min represents the minimum tidal volume corresponding to each breath in the preset number of breaths; f represents the respiratory frequency of the current breath; f max Indicates the preset maximum allowed respiratory rate; H m Indicates the heart rate data value corresponding to abnormal heart rate; H y Indicates the preset heart rate threshold; The abnormal breathing coefficient is compared with a preset abnormal breathing coefficient threshold, and the collection frequency of heart rate variability and breathing depth changes is adjusted according to the comparison result.

[0041] The technical effect of the above technical solution is that the system can monitor the patient's heart rate data value in real time and immediately compare it with the preset heart rate threshold. This real-time monitoring capability helps to detect abnormal heart rate of patients in time, provide timely warning information to medical personnel, and take necessary medical measures. When it is detected that the patient's heart rate exceeds the preset threshold, the system will further monitor the tidal volume corresponding to each breath of the patient in the preset number of breaths. Tidal volume is an important indicator for measuring the depth of breathing. By monitoring it, more detailed respiratory status information can be obtained. Using the monitored tidal volume data, the system can calculate the respiratory abnormality coefficient, which comprehensively considers the average value, minimum value, respiratory frequency and degree of heart rate abnormality of the tidal volume, and provides a comprehensive indicator for evaluating the patient's respiratory condition. By comparing the calculated respiratory abnormality coefficient with the preset threshold, the system can dynamically adjust the acquisition frequency of heart rate variability and respiratory depth changes according to the patient's real-time condition. This dynamic adjustment mechanism helps to reduce unnecessary monitoring and save resources when the patient's condition is stable; and increase the monitoring frequency when the patient's condition deteriorates, so as to more accurately capture changes in physiological parameters. This technical solution helps improve the efficiency of medical staff in monitoring the physiological status of patients through real-time monitoring, early warning and detailed evaluation, and timely discovers and handles potential medical problems. At the same time, the mechanism of dynamically adjusting the acquisition frequency also helps to balance the efficiency of medical resource utilization and the needs of patient safety. By presetting different heart rate thresholds and respiratory abnormality coefficient thresholds, the system can be personalized for different patients or different conditions to better meet clinical needs. This personalized setting helps to improve the pertinence and effectiveness of medical treatment.

[0042] In summary, this technical solution achieves accurate monitoring and early warning of the patient's physiological state by real-time monitoring of the patient's heart rate and respiratory parameters and comprehensive evaluation using the respiratory abnormality coefficient. At the same time, by dynamically adjusting the acquisition frequency, the utilization efficiency of medical resources and the safety of patients are improved, providing strong support for personalized medicine.

[0043] Specifically, the abnormal breathing coefficient is compared with a preset abnormal breathing coefficient threshold, and the acquisition frequency of heart rate variability and breathing depth change is adjusted according to the comparison result, including: Comparing the abnormal breathing coefficient with a preset abnormal breathing coefficient threshold to obtain a comparison result; When the comparison result shows that the abnormal breathing coefficient is not lower than the preset abnormal breathing coefficient threshold, there is no need to adjust the collection frequency of heart rate variability and breathing depth change; When the comparison result shows that the abnormal breathing coefficient is lower than the preset abnormal breathing coefficient threshold, the abnormal breathing coefficient is used to adjust the acquisition frequency of the heart rate variability and the breathing depth change to obtain the adjusted acquisition frequency of the heart rate variability and the breathing depth change; The adjusted heart rate variability acquisition frequency is obtained by the following formula:

[0044] Among them, F h F represents the acquisition frequency of adjusted heart rate variability; h0 represents the acquisition frequency of heart rate variability before adjustment; R represents the respiratory abnormality coefficient; R y Indicates the preset breathing abnormality coefficient threshold; V p V represents the average tidal volume corresponding to each breath in the preset number of breaths; min V represents the minimum tidal volume corresponding to each breath in the preset number of breaths; max Indicates the maximum tidal volume corresponding to the tidal volume of each breath in the preset number of breaths; At the same time, the acquisition frequency of the adjusted breathing depth change is obtained by the following formula:

[0045] Among them, F t Indicates the sampling frequency of the adjusted breathing depth change; F t0 Indicates the sampling frequency of breathing depth changes before adjustment; V p V represents the average tidal volume corresponding to each breath in the preset number of breaths; max Indicates the maximum tidal volume corresponding to each breath in the preset number of breaths; H m Indicates the heart rate data value corresponding to abnormal heart rate; H y Indicates the preset heart rate threshold; According to the adjusted collection frequency of heart rate variability and breathing depth changes, the health monitoring equipment is controlled to collect data corresponding to the heart rate variability and breathing depth changes.

[0046] The technical effect of the above technical solution is that by comparing the abnormal breathing coefficient with the preset threshold, the system can intelligently determine whether the acquisition frequency needs to be adjusted. When the abnormal breathing coefficient is high (i.e. not lower than the preset threshold), the system maintains the original acquisition frequency to ensure that sufficient data can be obtained for accurate evaluation when the patient's condition is unstable. When the abnormal breathing coefficient is low, the system dynamically reduces the acquisition frequency according to the coefficient value, thereby reducing unnecessary resource consumption while ensuring the effectiveness of monitoring. The adjustment formula takes into account multiple factors, including the abnormal breathing coefficient, the average, minimum, and maximum values ​​of the tidal volume, and heart rate data, which together determine the adjusted acquisition frequency. This personalized adjustment strategy can better adapt to the physiological characteristics and disease changes of different patients, and improve the pertinence and accuracy of monitoring. By dynamically adjusting the acquisition frequency, the system can increase data collection points during critical periods, thereby more accurately capturing the changing trends of the patient's physiological parameters. At the same time, reducing data collection points during non-critical periods helps reduce data noise and improve the overall quality of monitoring data. Reducing unnecessary monitoring can reduce interference to patients, especially during long-term monitoring, which helps to improve the comfort and acceptance of patients, thereby promoting the smooth progress of monitoring work. Real-time, accurate and personalized monitoring data provides strong support for clinical decision-making. Doctors can use this data to more accurately assess the patient's condition and develop more effective treatment plans, thereby improving treatment outcomes and patient satisfaction. The adjustment formula and threshold settings in this technical solution have certain flexibility and scalability. With the development of medical technology and clinical needs, these parameters can be easily adjusted and optimized to adapt to new monitoring needs and treatment strategies.

[0047] In summary, this technical solution achieves refined management of patients' physiological parameters and resource optimization through the acquisition frequency adjustment strategy based on the respiratory abnormality coefficient, improves monitoring efficiency and accuracy, enhances patients' comfort and acceptance, and promotes clinical decision-making and treatment effects. At the same time, this technical solution also shows good scalability and flexibility, providing strong support for the development of future medical monitoring technology.

[0048] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "including", "having" or any other variations thereof are intended to cover non-exclusive possessing, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0049] Although embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and the scope of the present invention is defined by the appended claims and their equivalents.

Claims

1. A heat stroke prevention and monitoring system based on the recognition of consciousness and behavioral characteristics, characterized by: include: A data collection unit, for collecting physiological signals and behavioral data of patients suffering from heat stroke as a type of data; It is also used to collect physiological signals and behavioral data of normal patients as the second type of data; the physiological signal data includes: patient's body surface temperature, brain waves, electrocardiogram and respiratory biofeedback signals; the behavioral data includes: movement frequency, reaction speed and attention behavior indicators; The feature extraction unit is used to clean and reduce the dimension of the first-class data and the second-class data, and extract feature information reflecting the patient's physical state and metabolic level from the first-class data and the second-class data respectively; the feature information of patients with heat stroke is used as a feature vector, and the label is set to be positive; and the feature information of patients without heat stroke is used as a training sample, and the label is set to be negative; The feature recognition unit is used to use the feature vector and label as a training set, and the feature information and label of the second-class data as a test set; learn the feature vector and label content through the linear discriminant analysis model, and output the learning result; then analyze the feature information and label content of the second-class data through the learned linear discriminant analysis model, determine whether there are heat stroke features in the feature information of the second-class data, and output the prediction result; compare the prediction result with the actual result of the second-class data to verify the performance of the linear discriminant analysis model; The prevention and monitoring unit is used to predict whether the current patient suffers from heat stroke through a linear discriminant analysis model. If the patient suffers from heat stroke, the risk index of the current patient suffering from heat stroke is further calculated, and different prevention and monitoring strategies are implemented for the current patient according to the risk index.

2. The heat stroke prevention and monitoring system based on consciousness cognitive behavior feature recognition according to claim 1 is characterized by: The feature extraction unit comprises: Data processing module, used for cleaning, denoising and normalizing the first and second category data; The feature extraction module is used to analyze the information content in the first and second category data, and extract the heart rate variability, breathing depth change and skin resistance as feature information reflecting the patient's physical state and metabolic level; The feature analysis module is used to compare the feature information of patients with heat stroke with the feature information of patients without heat stroke, and analyze whether the feature information of patients with heat stroke is different from the feature information of patients without heat stroke under the influence of heat stroke.

3. The heat stroke prevention and monitoring system based on consciousness cognitive behavior feature recognition according to claim 2 is characterized by: The feature analysis module analyzes whether the feature information of patients with heat stroke is different from that of patients without heat stroke under the influence of heat stroke, specifically, based on heart rate variability, breathing depth change and skin resistance, calculates the feature vector of each patient respectively, and determines whether there is a difference in the feature vector of each patient; Calculation of heart rate variability eigenvector: The voltage at each heartbeat point in the ECG signal is matched to the original heart rate through a linear interpolation formula, and for each heartbeat point, the time domain interval DT between two adjacent heartbeat points is calculated; For each heartbeat point, calculate the power spectrum density estimate PSDE corresponding to each time domain interval DT; then take the frequency corresponding to the maximum value of the power spectrum density estimate PSDE of each heartbeat point as the feature vector; Calculation of breathing depth change feature vector: The voltage at each respiratory cycle point in the respiratory signal is matched to the original respiratory cycle number through a linear interpolation formula. For each respiratory cycle point, the time domain interval DT between two adjacent respiratory cycle points is calculated; for each respiratory cycle point, the power spectrum density estimate PSDE corresponding to each time domain interval DT is calculated; and then the frequency corresponding to the maximum value of the power spectrum density estimate PSDE of each respiratory cycle point is taken as the feature vector; Calculation of skin resistance eigenvector: Each resistance value in the skin impedance signal is corresponded to the original skin impedance by a linear interpolation formula, and for each skin impedance, the time domain interval DT between two adjacent skin impedances is calculated; For each skin impedance, the power spectrum density estimate PSDE corresponding to each time domain interval DT is calculated; then the frequency corresponding to the maximum value of the power spectrum density estimate PSDE of each skin impedance is taken as the feature vector; The judgment basis is: patients with heat stroke have higher values ​​of characteristic vectors; patients without heat stroke have lower values ​​of characteristic vectors. Based on this, it is analyzed whether the characteristic information of patients with heat stroke is different from that of patients without heat stroke under the influence of heat stroke.

4. The heat stroke prevention and monitoring system based on consciousness cognitive behavior feature recognition according to claim 1 is characterized by: The feature recognition unit comprises: The model building module is used to create a neural network model, input the training set into the neural network model, use the LDA linear discriminant analysis algorithm to train the model, and output the training results to obtain a linear discriminant analysis model; The model evaluation module is used to input the test set into the linear discriminant analysis model for testing, use the linear discriminant analysis model to identify whether there is characteristic information of heat stroke in the test set, and verify the performance of the linear discriminant analysis model based on the test results.

5. The heat stroke prevention and monitoring system based on consciousness cognitive behavior feature recognition according to claim 4 is characterized by: The feature recognition unit further includes: The feature interpretation module is used to analyze the meaning of each feature information after the training set and test set training are completed.

6. The heat stroke prevention and monitoring system based on consciousness cognitive behavior feature recognition according to claim 1 is characterized by: The preventive monitoring unit comprises: The risk analysis module is used to collect the current patient's personal information as an additional factor and calculate the risk index of the current patient suffering from heat stroke using a weighted average method; The early warning module is used to build a risk threshold. If the patient's risk index exceeds the set threshold, it means that the patient is at high risk of heat stroke, and an alarm will be issued in time to remind medical staff to pay attention to the patient's condition.

7. The heat stroke prevention and monitoring system based on consciousness cognitive behavior feature recognition according to claim 1 is characterized by: The data collection unit comprises: The real-time collection module is used to collect the patient's heart rate variability, breathing depth changes, and skin resistance data through health monitoring equipment as basic data for clinical diagnosis or daily health prevention monitoring.

8. The heat stroke prevention and monitoring system based on consciousness cognitive behavior feature recognition according to claim 7 is characterized by: The data collection unit further includes: Real-time monitoring of the patient's heart rate data value; Comparing the heart rate data value of the current patient with a preset heart rate threshold; When the heart rate data value of the current patient exceeds a preset heart rate threshold, the tidal volume corresponding to each breath of the current patient within a preset number of breaths is monitored; Obtaining a respiratory abnormality coefficient using the tidal volume corresponding to each breath of the current patient within a preset number of breaths; The abnormal breathing coefficient is obtained by the following formula: Where R represents the respiratory abnormality coefficient; V p V represents the average tidal volume corresponding to each breath in the preset number of breaths; min represents the minimum tidal volume corresponding to each breath in the preset number of breaths; f represents the respiratory frequency of the current breath; f max Indicates the preset maximum allowed respiratory rate; H m Indicates the heart rate data value corresponding to abnormal heart rate; H y Indicates the preset heart rate threshold; The abnormal breathing coefficient is compared with a preset abnormal breathing coefficient threshold, and the collection frequency of heart rate variability and breathing depth changes is adjusted according to the comparison result.

9. The heat stroke prevention and monitoring system based on consciousness cognitive behavior feature recognition according to claim 8 is characterized by: The abnormal breathing coefficient is compared with a preset abnormal breathing coefficient threshold, and the acquisition frequency of the heart rate variability and the breathing depth change is adjusted according to the comparison result, including: Comparing the abnormal breathing coefficient with a preset abnormal breathing coefficient threshold to obtain a comparison result; When the comparison result shows that the abnormal breathing coefficient is not lower than the preset abnormal breathing coefficient threshold, there is no need to adjust the collection frequency of heart rate variability and breathing depth change; When the comparison result shows that the abnormal breathing coefficient is lower than the preset abnormal breathing coefficient threshold, the abnormal breathing coefficient is used to adjust the acquisition frequency of the heart rate variability and the breathing depth change to obtain the adjusted acquisition frequency of the heart rate variability and the breathing depth change; The adjusted heart rate variability acquisition frequency is obtained by the following formula: Among them, F h F represents the acquisition frequency of adjusted heart rate variability; h0 represents the acquisition frequency of heart rate variability before adjustment; R represents the respiratory abnormality coefficient; R y Indicates the preset breathing abnormality coefficient threshold; V p V represents the average tidal volume corresponding to each breath in the preset number of breaths; min V represents the minimum tidal volume corresponding to each breath in the preset number of breaths; max Indicates the maximum tidal volume corresponding to the tidal volume of each breath in the preset number of breaths; At the same time, the acquisition frequency of the adjusted breathing depth change is obtained by the following formula: Among them, F t Indicates the sampling frequency of the adjusted breathing depth change; F t0 Indicates the sampling frequency of breathing depth changes before adjustment; V p V represents the average tidal volume corresponding to each breath in the preset number of breaths; max Indicates the maximum tidal volume corresponding to each breath in the preset number of breaths; H m Indicates the heart rate data value corresponding to abnormal heart rate; H y Indicates the preset heart rate threshold; According to the adjusted collection frequency of heart rate variability and breathing depth changes, the health monitoring equipment is controlled to collect data corresponding to the heart rate variability and breathing depth changes.

10. The heat stroke prevention and monitoring system based on consciousness cognitive behavior feature recognition according to claim 1 is characterized by: Also includes: The human-computer interaction unit is used to visualize the current patient's physiological signals and behavioral data as well as the prediction results and risk index of the linear discriminant analysis model.

Images