Data-driven plateau hypoxia damage risk measurement system and method

Abstract

The invention discloses a data-driven plateau hypoxia injury risk measurement system and method, and belongs to the technical field of human body hypoxia injury risk assessment. Comprising an intermittent hypoxia training module, a wearable physiological signal acquisition module, a hypoxia cabin test module, a centralized data storage and learning module, an edge data processing and hypoxia injury risk assessment module and a communication and transmission module. The function of rapidly and dynamically evaluating the hypoxia injury risk of the subject at low cost is achieved, and the hypoxia injury risk of the subject can be evaluated in a safer, faster and low-cost mode under the condition that the subject does not actually enter a plateau environment.

Description

Technical Field

[0001] The present invention relates to the technical field of human hypoxia injury risk assessment, and in particular to a data-driven plateau hypoxia injury risk measurement system and method. Background Art

[0002] my country has vast high-altitude areas that occupy a quarter of its land area. They are important in military, economic and social strategies and are rich in resources. However, the natural environment in this area is extremely harsh and faces a series of climate problems, such as low oxygen and low pressure, dry cold, strong winds and strong ultraviolet rays. The construction and improvement of the Qinghai-Tibet Railway has promoted the vigorous development of tourism in the plateau area, attracting a large number of tourists from low-altitude areas. However, these tourists may encounter a series of dangerous symptoms, including headaches, nausea, coma and even death, due to their inability to adapt to the plateau environment. The health problems of low-altitude residents after entering the plateau are becoming increasingly prominent. In addition, although the permanent residents in high-altitude areas have a certain physical adaptability to the low-oxygen environment, their physiological state has always been maintained at an unsatisfactory level. Compared with people in normal altitude areas, their blood oxygen saturation and heart rate are too low and too fast, respectively, which can easily lead to the occurrence of chronic altitude sickness. Therefore, ensuring and improving the health quality of people in high-altitude areas is a vital task.

[0003] One feasible measure to prevent acute mountain sickness is to test the risk of high altitude hypoxia injury before entering the plateau. Currently available methods mainly include two categories: one uses questionnaires that rely on self-scoring, and the other mainly uses physiological indicators to predict mountain sickness.

[0004] However, the above prediction and evaluation methods all require the subjects to actually enter the plateau, or to undergo long-term testing in a simulated plateau environment. Such methods have long testing time and high cost. Based on this, the present invention proposes a data-driven plateau hypoxia injury risk measurement system and method. Summary of the invention

[0005] The purpose of the present invention is to provide a data-driven plateau hypoxia injury risk measurement system and method, which can realize the function of low-cost, rapid and dynamic assessment of the hypoxia injury risk of subjects.

[0006] To achieve the above object, the present invention provides a data-driven plateau hypoxia injury risk measurement system, comprising:

[0007] Intermittent hypoxia training module, which controls the oxygen content of the air inhaled by the subjects;

[0008] Wearable physiological signal acquisition module, real-time monitoring of subjects' physiological index data related to hypoxia injury risk;

[0009] Hypoxia chamber test module, which generates a hypoxic environment and monitors various indicators of the subject's sleep process;

[0010] Centralized data storage and learning module, which collects physiological index data, sleep index and Lake Louise score, and builds the relationship between physiological data and hypoxic injury risk during intermittent hypoxic training;

[0011] Edge data processing and hypoxia injury risk assessment module to assess the hypoxia injury risk of the subject;

[0012] Communication and transmission module, the bridge for communication between all modules.

[0013] Preferably, the intermittent hypoxia training module includes a hypoxia generator for generating hypoxic gas, a hypoxic storage for storing the hypoxic gas, and a gas mixer for mixing the hypoxic gas and air.

[0014] Preferably, the wearable physiological signal acquisition module uses a wearable device to measure the subject's blood oxygen saturation, heart rate and blood pressure.

[0015] Preferably, the mattress and pillow of the hypoxia chamber test module are provided with pressure sensors for monitoring the number of awakenings, turning over frequency and deep sleep time of the subject.

[0016] The present invention also provides a data-driven high-altitude hypoxia injury risk measurement method, comprising the following steps:

[0017] S1. Generate a hypoxic environment and collect hypoxic environment data;

[0018] S2. A hypoxic environment is input into a hypoxic chamber, and the subject enters the hypoxic chamber for testing, and the subject's continuous physiological data is monitored and recorded;

[0019] S3. Investigate the Lake Louise score of the subjects and evaluate the degree of hypoxic injury of the subjects in combination with continuous physiological data;

[0020] S4, collecting the hypoxic environment data obtained in S1, as well as the Lake Louise score and hypoxic injury degree of the subjects in S3, using a centralized data storage and learning module to learn the hypoxic injury risk virtual measurement data of the subjects through a data-driven method to obtain an estimation model;

[0021] S5. The continuous physiological data collected in S2 and the estimation model obtained in S4 are transmitted to the edge data processing and hypoxia injury risk assessment module, and the hypoxia injury risk assessment is performed on the subject's physiological data through the estimation model.

[0022] Preferably, the process of generating a hypoxic environment in S1 is as follows:

[0023] S11. Use a hypoxic generator to precisely control the oxygen concentration and the ratio of breathing gases to generate hypoxic gas and provide an appropriate intermittent hypoxic environment;

[0024] S12, using a hypoxia storage device to store the hypoxic gas generated by the hypoxic generator;

[0025] S13, sending the hypoxic gas into a gas mixer, and mixing the hypoxic gas and air in a certain proportion and pattern through the gas mixer to obtain a hypoxic environment.

[0026] Preferably, the process of obtaining continuous physiological data in S2 is as follows:

[0027] S21, inputting the hypoxic environment obtained in S13 into a hypoxic chamber, and the subject sleeps in the hypoxic chamber;

[0028] S22. Using wearable devices based on optical sensing technology to non-invasively measure and record the blood oxygen saturation and heart rate of the subjects;

[0029] S23. Use a wearable device to non-invasively measure and record the subject's blood pressure through pulse wave sensing technology;

[0030] S24. Monitor and record the subject's sleep status through the detection equipment in the hypoxia chamber.

[0031] Preferably, the process of investigating the Lake Louise score and the degree of hypoxic injury of the subject in S3 is as follows:

[0032] S31. After the subject wakes up, the Lake Louise score of the subject is obtained through a questionnaire;

[0033] S32. Assess the degree of hypoxic injury in the subject by combining the subject's Lake Louise score with blood oxygen saturation, heart rate, blood pressure and sleep status.

[0034] Preferably, the process of obtaining the estimation model through the centralized data storage and learning module in S4 is as follows:

[0035] S41. The principal component analysis method is used to reduce the dimension of the physiological indicators of hypoxia injury risk. The process is as follows:

[0036] Assume that the K measured data of hypoxic environment are H = {η 1 (k),…,η K (k)}, decentralize this set of data,

[0037]

[0038] Among them, η(k) represents the physiological index of hypoxia injury risk, Represents the measured data after decentralization, express The covariance matrix on the measured data set, The eigenvalues ​​in are sorted from large to small as follows: is the eigenvector of each eigenvalue, P η is the transformation matrix;

[0039] According to the measured data after centralization and the transformation matrix P η , the K hypoxia injury risk state data x after dimension reduction are obtained through the following formula: i (k)

[0040]

[0041] S42. Slow feature analysis is performed by solving the variational optimization method. The variational optimization process is as follows:

[0042]

[0043] In the process of solving variational optimization, some constraints are imposed on the data. The constraints are as follows:

[0044]

[0045] where g SFA (·):γ k (t)→∈ k (t) represents the function variable to be optimized, and the final result is the optimized slow feature signal ∈ k , γ k,i (t) = Γ i,t (k) Indicates n Γ dimensional time series signal, Γ(k) represents the continuous physiological data signal of the subject on the kth day, represents the dimension reduction signal, Represents the continuous physiological data of the subject;

[0046] Construct the physiological characteristic variable function of the subjects during intermittent hypoxia training.

[0047] y(k):=τ(∈ k );

[0048] Where y(k) represents the physiological monitoring signal characteristics in intermittent hypoxia training, τ(·) is the feature extraction function, ∈ k Indicates the optimization of slow feature signal;

[0049] S43. Obtain training data set by monitoring the subject's physiological signals, Lake Louise score and hypoxia injury degree Through the training data set The nonlinear function h(·) is learned and the estimation model is obtained by using Gaussian process regression method.

[0050]

[0051] Where x' is the training data set The data in represents the expected function, κ(x,x') represents the covariance kernel function, represents a Gaussian process.

[0052] Preferably, the process of using edge data processing and hypoxia injury risk assessment module to assess the hypoxia injury risk state in S5 is as follows:

[0053] S51, receive the estimated model obtained in S4 Finally, the evaluation of the hypoxia injury risk state is obtained according to the extended Kalman filter estimator. First, the state prediction and error covariance prediction are performed:

[0054]

[0055] in Indicates the prior prediction state of hypoxic injury risk. When there is no prior information represents the estimated risk of hypoxic injury, The state evolution equation obtained by the prior information of the physiological development process of hypoxia injury in human body is represented by: Represents the Jacobian matrix of the state prediction function, represents the covariance matrix of the state prediction error, u(k-1) represents the input hypoxia signal, express The covariance of , P(k-1) represents the state credibility at time k-1, express The transpose of

[0056] S52, bringing the prediction result obtained in S51 into the Kalman gain calculation,

[0057]

[0058] in is the virtual measurement process error covariance matrix, H(k) represents the virtual measurement process Jacobian matrix, H T (k) represents the transpose of H(k), K(k) represents the ratio of the reliability of the two data sources;

[0059] S53, using the results in S51 and S52 to perform state update and error square update,

[0060]

[0061] in Indicates the hypoxia injury risk status after the update assessment, which is the final output injury risk at that moment. represents the expectation, P(k) represents the covariance at time K, and it is used as the input in the calculation of the error covariance at the next moment, and I represents the identity matrix.

[0062] Therefore, the data-driven plateau hypoxia injury risk measurement system and method using the above structure of the present invention has the following advantages:

[0063] 1. Easy to deploy and highly portable. The physiological signal acquisition modules in the system all use wearable devices, and the data storage and learning are isolated from the hypoxia injury risk assessment. After the training is completed, only the edge data processing module can be used to complete the subject information collection and hypoxia injury risk assessment;

[0064] 2. The hypoxic injury risk simulation measurement process is constructed in a data-driven manner, avoiding the complex mechanism modeling process and better reflecting the relationship between the intermittent hypoxic training physiological signal data of the group subjects and the hypoxic injury risk in the hypoxic environment;

[0065] 3. It can evaluate the hypoxia injury risk of subjects in real time and dynamically, and can evaluate the hypoxia injury risk of subjects in a safer, faster and lower-cost way without the subjects actually entering the plateau environment;

[0066] 4. It can store and utilize a large amount of information to create an individualized physiological database, which can realize long-term physiological health monitoring and forecasting, and provide a basis for active health intervention when necessary.

[0067] The technical solution of the present invention is further described in detail below through the accompanying drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] Figure 1 A workflow diagram between various modules of a data-driven plateau hypoxia injury risk measurement system and method of the present invention;

[0069] Figure 2 A working process diagram of a data-driven plateau hypoxia injury risk measurement system and method of the present invention;

[0070] Figure 3 The oxygen supply control flow chart of a data-driven plateau hypoxia injury risk measurement system and method of the present invention;

[0071] Figure 4 The present invention is a flow chart of a system implementation method of a data-driven plateau hypoxia injury risk measurement system and method. DETAILED DESCRIPTION

[0072] Example

[0073] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0074] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. 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.

[0075] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0076] In the description of the present invention, it should be noted that the terms "upper", "lower", "inside", "outside", etc. indicate directions or positional relationships based on the directions or positional relationships shown in the accompanying drawings, or are directions or positional relationships in which the product of the invention is usually placed when in use. They are only for the convenience of describing the present invention and simplifying the description, and do not indicate or imply that the device or element referred to must have a specific direction, be constructed and operated in a specific direction, and therefore should not be understood as a limitation on the present invention.

[0077] In the description of the present invention, it is also necessary to explain that, unless otherwise clearly specified and limited, the terms "setting", "installation" and "connection" should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral connection; it can be a mechanical connection or an electrical connection; it can be a direct connection, or it can be an indirect connection through an intermediate medium, or it can be the internal communication of two components. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0078] like Figure 1-Figure 4As shown, the present invention is a data-driven plateau hypoxia injury risk measurement system, including an intermittent hypoxia training module, which controls the oxygen content in the air inhaled by the subject; a wearable physiological signal acquisition module, which monitors the physiological index data related to the hypoxia injury risk of the subject in real time; a hypoxia chamber test module, which generates a hypoxic environment and monitors various indicators of the subject's sleep process; a centralized data storage and learning module, which collects physiological index data, as well as sleep indicators and Lake Louise scores, and constructs the relationship between physiological data and hypoxia injury risk in intermittent hypoxia training; an edge data processing and hypoxia injury risk assessment module, which assesses the hypoxia injury risk of the subject; and a communication and transmission module, which is a bridge for communication between all modules.

[0079] The intermittent hypoxic training module includes a hypoxic generator for generating hypoxic gas, a hypoxic storage for storing the hypoxic gas, and a gas mixer for mixing the hypoxic gas and air.

[0080] The wearable physiological signal acquisition module uses wearable devices to measure the subject's blood oxygen saturation, heart rate and blood pressure.

[0081] The mattress and pillow of the hypoxia chamber test module are equipped with pressure sensors for monitoring the number of awakenings, tossing frequency and deep sleep time of the subjects.

[0082] The present invention also provides a data-driven high-altitude hypoxia injury risk measurement method, comprising the following steps:

[0083] S1. Generate a hypoxic environment and collect hypoxic environment data;

[0084] S11. Use a hypoxic generator to precisely control the oxygen concentration and the ratio of breathing gases to generate hypoxic gas and provide an appropriate intermittent hypoxic environment;

[0085] S12, using a hypoxia storage device to store the hypoxic gas generated by the hypoxic generator;

[0086] S13, sending the hypoxic gas into a gas mixer, and mixing the hypoxic gas and air in a certain proportion and pattern through the gas mixer to obtain a hypoxic environment.

[0087] S2. A hypoxic environment is input into a hypoxic chamber, and the subject enters the hypoxic chamber for testing, and the subject's continuous physiological data is monitored and recorded;

[0088] S21, inputting the hypoxic environment obtained in S13 into a hypoxic chamber, and the subject sleeps in the hypoxic chamber;

[0089] S22. Using wearable devices based on optical sensing technology to non-invasively measure and record the blood oxygen saturation and heart rate of the subjects;

[0090] S23. Use a wearable device to non-invasively measure and record the subject's blood pressure through pulse wave sensing technology;

[0091] S24. Monitor and record the subject's sleep status through the detection equipment in the hypoxia chamber.

[0092] S3. Investigate the Lake Louise score of the subjects and evaluate the degree of hypoxic injury of the subjects in combination with continuous physiological data;

[0093] S31. After the subject wakes up, the Lake Louise score of the subject is obtained through a questionnaire;

[0094] S32. Assess the degree of hypoxic injury in the subject by combining the subject's Lake Louise score with blood oxygen saturation, heart rate, blood pressure and sleep status.

[0095] S4. Collect the hypoxic environment data obtained in S1, as well as the Lake Louise score and hypoxic injury degree of the subjects in S3, and use the centralized data storage and learning module to learn the hypoxic injury risk virtual measurement data of the subjects through a data-driven method to obtain an estimation model. The process is as follows:

[0096] S41. The principal component analysis method is used to reduce the dimension of the physiological indicators of hypoxia injury risk. The process is as follows:

[0097] Assume that the K measured data of hypoxic environment are H = {η 1 (k),…,η K (k)}, decentralize this set of data,

[0098]

[0099]

[0100] Among them, η(k) represents the physiological index of hypoxia injury risk, Represents the measured data after decentralization, express The covariance matrix on the measured data set, The eigenvalues ​​in are sorted from large to small as follows: is the eigenvector of each eigenvalue, P η is the transformation matrix;

[0101] According to the measured data after centralization and the transformation matrix P η , the K hypoxia injury risk state data x after dimension reduction are obtained through the following formula: i (k)

[0102]

[0103] S42. Slow feature analysis is performed by solving the variational optimization method. The variational optimization process is as follows:

[0104]

[0105] In the process of solving variational optimization, some constraints are imposed on the data. The constraints are as follows:

[0106]

[0107] where g SFA (·):γ k (t)→∈ k (t) represents the function variable to be optimized, and the final result is the optimized slow feature signal ∈ k , γ k,i (t) = Γ i,t (k) Indicates n Γ dimensional time series signal, Γ(k) represents the continuous physiological data signal of the subject on the kth day, represents the dimension reduction signal, Represents the continuous physiological data of the subject;

[0108] Construct the physiological characteristic variable function of the subjects during intermittent hypoxia training.

[0109] y(k):=τ(∈ k );

[0110] Where y(k) represents the physiological monitoring signal characteristics in intermittent hypoxia training, τ(·) is the feature extraction function, ∈ k Indicates the optimization of slow feature signal;

[0111] S43. Obtain training data set by monitoring the subject's physiological signals, Lake Louise score and hypoxia injury degree Through the training data set The nonlinear function h(·) is learned and the estimation model is obtained by using Gaussian process regression method.

[0112]

[0113] Where x' is the training data set The data in represents the expected function, κ(x,x') represents the covariance kernel function, represents a Gaussian process.

[0114] S5, transmitting the continuous physiological data collected in S2 and the estimation model obtained in S4 to the edge data processing and hypoxia injury risk assessment module, and performing hypoxia injury risk assessment on the subject's physiological data through the estimation model, the process is as follows:

[0115] S51, receive the estimated model obtained in S4 Finally, the evaluation of the hypoxia injury risk state is obtained according to the extended Kalman filter estimator. First, the state prediction and error covariance prediction are performed:

[0116]

[0117] in Indicates the prior prediction state of hypoxic injury risk. When there is no prior information represents the estimated risk of hypoxic injury, The state evolution equation obtained by the prior information of the physiological development process of hypoxia injury in human body is represented by: Represents the Jacobian matrix of the state prediction function, represents the covariance matrix of the state prediction error, u(k-1) represents the input hypoxia signal, express The covariance of , P(k-1) represents the state credibility at time k-1, express The transpose of

[0118] S52, bringing the prediction result obtained in S51 into the Kalman gain calculation,

[0119]

[0120] in is the virtual measurement process error covariance matrix, H(k) represents the virtual measurement process Jacobian matrix, H T (k) represents the transpose of H(k), K(k) represents the ratio of the reliability of the two data sources;

[0121] S53, using the results in S51 and S52 to perform state update and error square update,

[0122]

[0123] in Indicates the hypoxia injury risk status after the update assessment, which is the final output injury risk at that moment. represents the expectation, P(k) represents the covariance at time K, and it is used as the input in the calculation of the error covariance at the next moment, and I represents the identity matrix.

[0124] Therefore, the present invention adopts a data-driven plateau hypoxia injury risk measurement system and method of the above-mentioned structure, which has the function of low-cost, fast and dynamic assessment of the hypoxia injury risk of subjects. At the same time, wearable devices are used to isolate data storage and learning from hypoxia injury risk assessment, which has the advantages of easy deployment and high portability. Moreover, the hypoxia injury risk simulation process is constructed in a data-driven manner to better reflect the relationship between the intermittent hypoxia training physiological signal data of a group of subjects and the hypoxia injury risk in a hypoxic environment; and the hypoxia injury risk of subjects can be assessed in a safer, faster and lower-cost manner without the subjects actually entering the plateau environment. Finally, a large amount of information can be stored and utilized to create an individualized physiological database, which can realize long-term physiological health monitoring and forecasting, and provide a basis for active health intervention when necessary.

[0125] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that they can still modify or replace the technical solution of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified technical solution to deviate from the spirit and scope of the technical solution of the present invention.

Claims

1. A data-driven plateau hypoxia injury risk measurement system, characterized by: include: Intermittent hypoxia training module, which controls the oxygen content of the air inhaled by the subjects; Wearable physiological signal acquisition module, real-time monitoring of subjects' physiological index data related to hypoxia injury risk; Hypoxia chamber test module, which generates a hypoxic environment and monitors various indicators of the subject's sleep process; Centralized data storage and learning module, which collects physiological index data, sleep index and Lake Louise score, and builds the relationship between physiological data and hypoxic injury risk during intermittent hypoxic training; Edge data processing and hypoxia injury risk assessment module to assess the hypoxia injury risk of the subject; Communication and transmission module, the bridge for communication between all modules.

2. A data-driven high-altitude hypoxia injury risk measurement system according to claim 1, characterized in that: The intermittent hypoxic training module includes a hypoxic generator for generating hypoxic gas, a hypoxic storage for storing the hypoxic gas, and a gas mixer for mixing the hypoxic gas and air.

3. A data-driven high-altitude hypoxia injury risk measurement system according to claim 2, characterized in that: The wearable physiological signal acquisition module uses wearable devices to measure the subject's blood oxygen saturation, heart rate and blood pressure.

4. A data-driven high-altitude hypoxia injury risk measurement system according to claim 3, characterized in that: The mattress and pillow of the hypoxia chamber test module are equipped with pressure sensors for monitoring the number of awakenings, tossing frequency and deep sleep time of the subjects.

5. A data-driven method for measuring the risk of plateau hypoxia injury, characterized in that: The following steps are involved: S1. Generate a hypoxic environment and collect hypoxic environment data; S2. A hypoxic environment is input into a hypoxic chamber, and the subject enters the hypoxic chamber for testing, and the subject's continuous physiological data is monitored and recorded; S3. Investigate and record the Lake Louise score of the subjects, and evaluate the degree of hypoxic injury of the subjects in combination with continuous physiological data; S4, collecting the hypoxic environment data obtained in S1, as well as the Lake Louise score and hypoxic injury degree of the subjects in S3, using a centralized data storage and learning module to learn the hypoxic injury risk virtual measurement data of the subjects through a data-driven method to obtain an estimation model; S5. The collected continuous physiological data and the estimation model obtained in S4 are transmitted to the edge data processing and hypoxia injury risk assessment module, and the hypoxia injury risk assessment is performed on the subject's physiological data through the estimation model.

6. A data-driven high-altitude hypoxia injury risk measurement method according to claim 5, characterized in that: The process of generating a hypoxic environment in S1 is as follows: S11. Use a hypoxic generator to precisely control the oxygen concentration and the ratio of breathing gases to generate hypoxic gas and provide an appropriate intermittent hypoxic environment; S12, using a hypoxia storage device to store the hypoxic gas generated by the hypoxic generator; S13, sending the hypoxic gas into a gas mixer, and mixing the hypoxic gas and air in a certain proportion and pattern through the gas mixer to obtain a hypoxic environment.

7. A data-driven high-altitude hypoxia injury risk measurement method according to claim 6, characterized in that: The process of obtaining continuous physiological data in S2 is as follows: S21, inputting the hypoxic environment obtained in S13 into a hypoxic chamber, and the subject sleeps in the hypoxic chamber; S22. Using wearable devices based on optical sensing technology to non-invasively measure and record the blood oxygen saturation and heart rate of the subjects; S23. Use a wearable device to non-invasively measure and record the subject's blood pressure through pulse wave sensing technology; S24. Monitor and record the subject's sleep status through the detection equipment in the hypoxia chamber.

8. A data-driven high-altitude hypoxia injury risk measurement method according to claim 7, characterized in that: The process of investigating the Lake Louise score and hypoxic injury degree of the subjects in S3 is as follows: S31. After the subject wakes up, the Lake Louise score of the subject is obtained through a questionnaire; S32. Assess the degree of hypoxic injury in the subject by combining the subject's Lake Louise score with blood oxygen saturation, heart rate, blood pressure and sleep status.

9. A data-driven high-altitude hypoxia injury risk measurement method according to claim 8, characterized in that: The process of obtaining the estimated model through the centralized data storage and learning module in S4 is as follows: S41. The dimension of the physiological index η(k) of hypoxia injury risk is reduced by principal component analysis. The process is as follows: Assume that the K measured data of hypoxic environment are H = {η 1 (k),…,η K (k)}, decentralize this set of data, Among them, η(k) represents the physiological index of hypoxia injury risk, Represents the measured data after decentralization, express The covariance matrix on the measured data set, The eigenvalues ​​in are sorted from large to small as follows: is the eigenvector of each eigenvalue, P η is the transformation matrix; According to the measured data after centralization and the transformation matrix P η , the K hypoxia injury risk state data x after dimension reduction are obtained through the following formula: i (k) S42. Slow feature analysis is performed by solving the variational optimization method. The variational optimization process is as follows: In the process of solving variational optimization, some constraints are imposed on the data. The constraints are as follows: where g SFA (·):γ k (t)→∈ k (t) represents the function variable to be optimized, and the final result is the optimized slow feature signal ∈ k , γ k，i (t) = Γ i，t (k) Indicates n Γ dimensional time series signal, Γ(k) represents the continuous physiological data signal of the subject on the kth day, represents the dimension reduction signal, Represents the continuous physiological data of the subject; Construct the physiological characteristic variable function of the subjects during intermittent hypoxia training. y(k):=τ(∈ k ); Where y(k) represents the physiological monitoring signal characteristics in intermittent hypoxia training, τ(·) is the feature extraction function, ∈ k Indicates the optimization of slow feature signal; S43. Obtain training data set by monitoring the subject's physiological signals, Lake Louise score and hypoxia injury degree Through the training data set The nonlinear function h(·) is learned and the estimation model is obtained by using Gaussian process regression method. Where x′ is the training data set The data in represents the expected function, K(x, x′) represents the covariance kernel function, represents a Gaussian process.

10. A data-driven high-altitude hypoxia injury risk measurement method according to claim 9, characterized in that: The process of using edge data processing and hypoxia injury risk assessment module in S5 to assess hypoxia injury risk status is as follows: S51, receive the estimated model obtained in S4 Finally, the evaluation of the hypoxia injury risk state is obtained according to the extended Kalman filter estimator. First, the state prediction and error covariance prediction are performed: in Indicates the prior prediction state of hypoxic injury risk. When there is no prior information represents the estimated risk of hypoxic injury, The state evolution equation obtained by the prior information of the physiological development process of hypoxia injury in human body is represented by: Represents the Jacobian matrix of the state prediction function, represents the covariance matrix of the state prediction error, u(k-1) represents the input hypoxia signal, express The covariance of , P(k-1) represents the state credibility at time k-1, express The transpose of S52, bringing the prediction result obtained in S51 into the Kalman gain calculation, in is the virtual measurement process error covariance matrix, H(k) represents the virtual measurement process Jacobian matrix, H T (k) represents the transpose of H(k), K(k) represents the ratio of the reliability of the two data sources; S53, using the results in S51 and S52 to perform state update and error square update, in Indicates the hypoxia injury risk status after the update assessment, which is the final output injury risk at that moment. represents the expectation, P(k) represents the covariance at time K, and it is used as the input in the calculation of the error covariance at the next moment, and I represents the identity matrix.

Images