Individualized plateau hypoxia preview closed-loop dry prediction algorithm

Abstract

The invention discloses an individualized plateau hypoxia preview closed-loop dry prediction algorithm, and relates to the technical field of acute altitude sickness intervention, and the algorithm comprises the following steps: S1, obtaining continuous data of oxygen supply concentration and oxyhemoglobin saturation during an IHT period, and dividing a data set into a training set, a verification set and a test set; s2, on the basis of the training set, a group prediction model of the blood oxygen saturation is constructed through a Kupman operator, model updating and prediction are conducted through online iteration, a blood oxygen saturation prediction value is output, and on the basis of the verification set, model individualization is conducted through the group prediction model; s3, predicting the oxyhemoglobin saturation based on the individualized model, constructing a lower-layer optimization cost function taking the low oxygen concentration as a variable, and carrying out security constraint self-adaptive upper-layer optimization design; s4, the training result is evaluated, and an evaluation result is fed back to the step S3; a dynamic evaluation index is calculated, and the hypoxia stimulation intensity of the IHT is adjusted by controlling an optimization strategy, so that the rapidity and effectiveness of the plateau adaptive capacity are improved.

Description

Technical Field

[0001] The invention relates to the technical field of acute altitude sickness intervention, and in particular to an individualized closed-loop intervention algorithm for high altitude hypoxia preparatory service. Background Art

[0003] Acute mountain sickness is one of the plateau-specific diseases. The risk of illness can be reduced through intermittent hypoxia training (IHT), a preparatory method. IHT is a training method that uses periodic hypoxia stimulation to cause a series of complex adaptive changes in the body's tissues. It can effectively improve the body's ability to adapt to the plateau environment. During the training, key continuous physiological indicators can be detected and recorded in real time through wearable devices, and used to evaluate the degree of training and hypoxia adaptation. Among them, blood oxygen saturation is recognized as the most direct and effective physiological indicator, followed by heart rate. A large number of medical studies use these two indicators and their variant indicators as the criteria for judging the ability to adapt to the plateau environment.

[0004] However, the effectiveness of IHT depends on the training protocol. Within the individual's tolerance range, a higher intensity and longer duration of hypoxic stimulation can more effectively improve adaptability. If the set intensity exceeds the tolerable range, it will be counterproductive and even cause irreversible damage. Existing research on altitude sickness based on IHT is limited to the evaluation of hypoxia tolerance and its ability to improve adaptability. It has not yet considered how to design an optimized decision-making algorithm to dynamically adjust the IHT training strategy to achieve rapid adaptation to the hypoxic environment and improve the ability to climb the plateau.

[0005] Specifically, during an IHT, the hypoxic strategy adopted is fixed, so the training is essentially based on an open-loop structure. There is no feedback link to effectively adjust the strategy based on the evaluation results. Considering the specificity between and within individuals, the use of the same fixed oxygen delivery protocol may not be universal for all individuals.

[0006] Therefore, an individualized closed-loop intervention algorithm for high-altitude hypoxia preparatory service is provided to solve the above problems. Summary of the invention

[0007] The purpose of the present invention is to provide an individualized closed-loop intervention algorithm for high altitude hypoxia preparatory training, which can accelerate the training cycle of individuals performing IHT in plain areas, improve hypoxia tolerance and reduce the risk of acute mountain sickness.

[0008] To achieve the above object, the present invention provides an individualized high altitude hypoxia preparatory service closed-loop intervention algorithm, comprising the following steps:

[0009] S1: During IHT, the subjects were given periodic oxygen supply, and physiological signals were collected in real time to obtain continuous data of oxygen supply concentration and blood oxygen saturation during IHT. The continuous data of blood oxygen saturation were divided into training set, validation set and test set according to different subjects.

[0010] S2: Based on the training set, the Koopman operator is used to learn the changing pattern of blood oxygen saturation under intermittent hypoxia, and a group prediction model of blood oxygen saturation is constructed. The model is updated and predicted through online iteration, and the predicted value of blood oxygen saturation is output. Based on the validation set, the group prediction model is used for model individualization;

[0011] S3: Use the individualized model to predict blood oxygen saturation, build a lower-level optimization cost function with low oxygen concentration as a variable through the dual-layer optimization module of supply and low oxygen concentration, and perform upper-level optimization design with safety constraints and self-adaptation based on historical learning performance;

[0012] S4: The training results are evaluated through the training effect evaluation module, and the evaluation results are fed back to the low oxygen concentration double-layer optimization module.

[0013] Preferably, in step S2, based on the training set, the Koopman operator is used to learn the changing law of blood oxygen saturation under intermittent hypoxia, and a group prediction model of blood oxygen saturation is constructed, which specifically includes the following steps:

[0014] Step 1: Obtain a nonlinear dynamic characteristic equation of blood oxygen saturation changing with oxygen concentration. The nonlinear dynamic characteristic equation is expressed as:

[0015] x k+1 =f(x k ,u k )

[0016] in u k represents the oxygen supply concentration at sampling time k, x k represents the blood oxygen saturation at sampling time k;

[0017] Step 2: Obtain the augmented state equation of the nonlinear dynamic characteristic equation. The augmented state equation is expressed as:

[0018] X k =[x k ,u k ] ·

[0019] The augmented state equation is expanded to obtain the extended equation, which is expressed as:

[0020] X k+1 =[f(x k ,u k ),Su k ]·

[0021] Among them Su k =u k+1 , S represents the left shift operator;

[0022] Step 3: The Koopman operator is obtained by extending the dynamic mode decomposition method, and the nonlinear dynamic characteristic equation in step 1 is converted into a linear dynamic characteristic equation by the Koopman operator.

[0023] Preferably, step 3 specifically includes the following steps:

[0024] Step 1: Get the state augmented matrix of blood oxygen saturation X and State augmented matrix of blood oxygen saturation X and Respectively expressed as:

[0025] X =[X 1 ,X 2 ,…,X M-1 ] ·

[0026]

[0027] Where M represents the total number of samples for constructing the population prediction model;

[0028] Step 2: Augment the blood oxygen saturation state matrix X and Perform state improvement respectively and obtain the state matrix Φ( X )and The state matrix Φ( X )and Respectively expressed as:

[0029] Φ( X )=[φ(X 1 ),φ(X 2 ),…,φ(X M-1 )] ·

[0030]

[0031] Step 3: Get the group Koopman operator Group Koopman operator It is expressed as:

[0032]

[0033] Step 4: Get the linear form dynamic characteristic equation, which is expressed as:

[0034]

[0035] Preferably, in step S2, the model is updated and predicted through online iteration, the blood oxygen saturation prediction value is output, and the model is individualized based on the validation set using the group prediction model, which specifically includes the following steps:

[0036] S21: perform daily model updates on day d online;

[0037] S22: estimated value based on the Koopman operator on day d-1 Iterate to get the estimated value of the Koopman operator on day d The iterative formula is expressed as:

[0038]

[0039] in and X d-1 Both represent the augmented state matrix obtained from the data of day d-1, γ d-1 represents the correction vector for day d-1, γ d-1 It is expressed as:

[0040] γ d-1 =Φ( X d-1 )Γ d

[0041] where Γ d represents the covariance matrix of the dth day, Γ d It is expressed as:

[0042] Γ d =Γ d-1 -Γ d-1 Φ( X d-1 )(Φ( X d-1 )Γ d-1 Φ( X d-1 )+I) -1 Φ( X d-1 ) · Γ d-1 .

[0043] Preferably, in step S22, when d is 1, X 0 , γ 0 and Γ 0 is the corresponding matrix obtained based on population data, Γ 0 It is expressed as:

[0044] Γ 0 =(Φ( X )Φ( X ) · ) -1 .

[0045] Preferably, in step S3, a lower optimization cost function with low oxygen concentration as a variable is constructed through a low oxygen concentration double-layer optimization module, and the lower optimization cost function is expressed as:

[0046]

[0047] X k =[x k ,u d ] T

[0048]

[0049] Where S y (u d ) indicates that during a hypoxic period, the supply of low oxygen concentration u d The area under the curve of the predicted blood oxygen saturation under the action of m It indicates the minimum value of blood oxygen saturation corresponding to the predicted hypoxic segment. ρ and μ are hyperparameters. k Indicates the blood oxygen saturation value corresponding to the predicted hypoxic segment. Indicates the upper threshold of low oxygen concentration, u(ΔSI d-1 ) indicates the lower threshold of low oxygen concentration.

[0050] Preferably, in step S3, an upper-layer optimization design of safety constraint adaptation is performed according to historical learning performance, and the upper-layer optimization design is specifically as follows:

[0051] Design 1: If ΔSI d-1 <0,||ΔSI d-1 ||>S TH , and ||u d-1 - u d-1 ||<U TH ,but It is expressed as:

[0052]

[0053] where α 0 , S TH and U TH are all hyperparameters, u d-1 It represents the lower threshold of oxygen supply concentration safety range on day d-1. u dIt represents the lower threshold of the safe range of oxygen supply concentration on day d;

[0054] Design 2: If ΔSI d-1 >0,||ΔSI d-1 ||>S TH ,and but It is expressed as:

[0055]

[0056] where α 1 is a hyperparameter, It represents the lower threshold of oxygen supply concentration safety range on day d-1. Indicates the lower threshold of the safe range of oxygen supply concentration on day d.

[0057] Preferably, the hyperparameters are determined by validation set data using a grid search method.

[0058] Preferably, step S4 specifically includes the following steps:

[0059] S41: Calculate the training evaluation index SI, which is expressed as:

[0060]

[0061] in It represents the average area of ​​blood oxygen saturation in all hypoxic segments of IHT in one day. Indicates the average hypoxic concentration;

[0062] S42: comparing the training evaluation index of the first day with the training evaluation index of the last day to evaluate the overall training result;

[0063] S43: Changes in the evaluation index ΔSI based on the previous day's training d-1 Evaluate the training effect and calculate the change ΔSI of the previous day’s training evaluation index d-1 Feedback is provided to the low oxygen concentration double-layer optimization module.

[0064] Preferably, in step S43, the change ΔSI of the training evaluation index of the previous day d-1 It is expressed as:

[0065] ΔSI d-1 =SI d-1 -SI d-2 .

[0066] Therefore, the present invention adopts the above-mentioned individualized closed-loop intervention algorithm for high altitude hypoxia preparatory service, which has the following beneficial effects:

[0067] (1) The individualized prediction model learning module of the present invention utilizes multiple hypoxic cycles in one day of IHT to perform Koopman operator modeling, which can fully mine data information and find key features for decision optimization;

[0068] (2) The constraint part of the dual-layer optimization module of the present invention for low oxygen supply concentration takes into account the physiological limitations of the human body and the differences in the adaptation of different individuals to training intensities, and controls the low oxygen supply concentration within a reasonable tolerance range, thereby accelerating the plateau preparatory clothing training while avoiding irreversible damage to individuals due to hypoxia;

[0069] (3) The present invention transforms IHT from an open-loop structure to a closed-loop structure, solving the problem of lack of guidance and direction in the current high-altitude preparatory training through IHT. Taking into account the individual specificity, different subjects will have different reactions when IHT uses the same hypoxic stimulation intensity. Some subjects may have a situation where the training effect becomes worse the more they train. After adopting closed-loop intervention, the algorithm allows real-time feedback based on the different training effects of each individual, intelligently optimizes the next oxygen supply strategy, effectively realizes personalized optimal training, and improves the training effect.

[0070] The method scheme of the present invention is further described in detail below through the drawings and embodiments. BRIEF DESCRIPTION OF THE DRAWINGS

[0071] Figure 1 A flowchart of an individualized closed-loop intervention algorithm for high altitude hypoxia preparatory service of the present invention;

[0072] Figure 2 The correlation result based on the test set of the embodiment of the present invention;

[0073] Figure 3 This is the trend correctness evaluation result based on the test set of the embodiment of the present invention. DETAILED DESCRIPTION

[0074] The method scheme of the present invention is further described below through drawings and embodiments.

[0075] Unless otherwise defined, method terms or scientific terms used in the present invention shall have the common meanings understood by one of ordinary skill in the art to which the present invention belongs.

[0076] The words "include" or "comprises" and the like used in the present invention mean that the elements before the word include the elements listed after the word, and do not exclude the possibility of also including other elements. The orientation or position relationship indicated by the terms "inside", "outside", "upper", "lower", etc. is based on the orientation or position relationship shown in the drawings, which is only for the convenience of describing the present invention and simplifying the description, and does not indicate or imply that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation. Therefore, it cannot be understood as a limitation of the present invention. When the absolute position of the described object changes, the relative position relationship may also change accordingly. In the present invention, unless otherwise clearly specified and limited, the terms "attachment" and the like should be understood in a broad sense, for example, it can be a fixed connection, a detachable connection, or an integral body; it can be directly connected, or indirectly connected through an intermediate medium, and it can be the internal connection of two elements or the interaction relationship between two elements. For ordinary technicians in this field, the specific meanings of the above terms in the present invention can be understood according to specific circumstances.

[0077] Example

[0078] This example uses IHT data from 13 subjects, of which 6 are used as training sets to determine the population model, 3 are used for model individualization based on the population model and as validation sets to determine the hyperparameters of the optimization problem, and the other 4 subjects are used as test sets to verify the effectiveness of the method.

[0079] like Figure 1 As shown, the present invention provides an individualized high-altitude hypoxia preparatory closed-loop intervention algorithm, comprising the following steps:

[0080] S1: During the high altitude preparatory period, the training subjects performed IHT for one hour every day for a total of 7 days. During the one-day IHT, periodic oxygen supply was performed, and the intake concentration was a square wave waveform. Hypoxia lasted for 5 minutes and hyperoxia lasted for 3 minutes. The hypoxia concentration u was the variable to be optimized. During the IHT period, the training subjects wore a blood oximeter. The oxygen supply concentration and blood oxygen saturation during the IHT period were detected through the group and individual prediction model learning module. The online data was acquired and recorded in real time, and the data was divided into training set, validation set and test set;

[0081] S2: Based on the training set, the Koopman operator is used to learn the changing pattern of blood oxygen saturation under intermittent hypoxia, and the data-driven method of the Koopman operator is used to build a group prediction model for blood oxygen saturation; the model is updated and predicted through online iteration, and the blood oxygen saturation prediction value is output. Based on the validation set, the group prediction model is used for model individualization;

[0082] In step S2, based on the training set, the Koopman operator is used to learn the changing law of blood oxygen saturation under intermittent hypoxia, and the group prediction model of blood oxygen saturation is constructed by using the data-driven method of the Koopman operator, which specifically includes the following steps:

[0083] Step 1: Obtain a nonlinear dynamic characteristic equation of blood oxygen saturation changing with oxygen concentration. The nonlinear dynamic characteristic equation is expressed as:

[0084] x k+1 =f(x k ,u k )

[0085] in u k represents the oxygen supply concentration at sampling time k, x k represents the blood oxygen saturation at sampling time k;

[0086] Step 2: Obtain the augmented state equation of the nonlinear dynamic characteristic equation. The augmented state equation is expressed as:

[0087] X k =[x k ,u k ] ·

[0088] The augmented state equation is expanded to obtain the extended equation, which is expressed as:

[0089] X k+1 =[f(x k ,u k ),Su k ] ·

[0090] Among them Su k =u k+1 , S represents the left shift operator;

[0091] Step 3: The Koopman operator is obtained by extending the dynamic mode decomposition method, and the nonlinear dynamic characteristic equation in step 1 is converted into a linear dynamic characteristic equation by the Koopman operator;

[0092] Step 3 specifically includes the following steps:

[0093] Step 1: Get the state augmented matrix of blood oxygen saturation X and State augmented matrix of blood oxygen saturation X and Respectively expressed as:

[0094] X =[X 1 ,X 2 ,…,XM-1 ] ·

[0095]

[0096] Where M represents the total number of samples for constructing the population prediction model;

[0097] Step 2: Augment the blood oxygen saturation state matrix X and Perform state improvement respectively and obtain the state matrix Φ( X )and Get the state matrix Φ( X )and Φ( X )and Respectively expressed as:

[0098] Φ( X )=[φ(X 1 ),φ(X 2 ),…,φ(X M-1 )] ·

[0099]

[0100] Step 3: Get the group Koopman operator Group Koopman operator Can characterize average dynamic characteristics, group Koopman operator It is expressed as:

[0101]

[0102] Step 4: Get the linear form dynamic characteristic equation, which is expressed as:

[0103]

[0104] In step S2, considering the variation of individual oxygen saturation, the model is individualized based on the group model through different individual data, and the daily model update is performed online on the dth day. The Koopman operator estimate value on the d-1th day is used to calculate the average daily model update rate. Iterate to get the estimated value of the Koopman operator on day d The iterative formula is expressed as:

[0105]

[0106] in and X d-1 Both represent the augmented state matrix obtained from the data of day d-1, γ d-1represents the correction vector for day d-1, γ d-1 It is expressed as:

[0107] γ d-1 =Φ( X d-1 )Γ d

[0108] where Γ d represents the covariance matrix of the dth day, Γ d It is expressed as:

[0109] Γ d =Γ d-1 -Γ d-1 Φ( X d-1 )(Φ( X d-1 )Γ d-1 Φ( X d-1 )+I) -1 Φ( X d-1 ) · Γ d-1 ;

[0110] When d is 1, X 0 , γ 0 and Γ 0 is the corresponding matrix obtained based on population data, Γ 0 It is expressed as:

[0111] Γ 0 =(Φ( X )Φ( X ) · ) -1 .

[0112] S3: Based on the predicted value of blood oxygen saturation, a safety constraint adaptive two-layer optimization framework with hypoxic concentration as the variable is constructed. Through the established individualized model, we can predict the blood oxygen saturation level under different hypoxic concentration conditions, and design the daily hypoxic concentration optimization problem based on this result. Through the hypoxic concentration two-layer optimization module, a lower-level optimization cost function with hypoxic concentration as the variable is constructed, and the upper-level optimization design of safety constraint adaptive is performed according to the historical learning performance;

[0113] In step S3, a lower optimization cost function with low oxygen concentration as a variable is constructed through the low oxygen concentration double-layer optimization module. The lower optimization cost function is expressed as:

[0114]

[0115] Xk =[x k ,u d ] T

[0116]

[0117] Where S y (u d ) indicates that during a hypoxic period, the supply of low oxygen concentration u d The area under the curve of the predicted blood oxygen saturation under the action of m It indicates the minimum value of blood oxygen saturation corresponding to the predicted hypoxic segment. ρ and μ are hyperparameters. k Indicates the blood oxygen saturation value corresponding to the predicted hypoxic segment. Indicates the upper threshold of low oxygen concentration, u(ΔSI d-1 ) represents the lower threshold of hypoxic concentration, and these two thresholds are related to the changes of the training evaluation index of the previous day and are determined by the upper optimization design;

[0118] The ratio is set to keep the human body's blood oxygen saturation level as high as possible while ensuring sufficient hypoxic stimulation to stimulate adaptive changes.

[0119] In step S3, the upper-level optimization design of safety constraint adaptation is performed according to the historical learning performance. The upper-level optimization design is specifically as follows:

[0120] Design 1: If ΔSI d-1 <0,||ΔSI d-1 ||>S TH , and ||u d-1 - u d-1 ||<U TH ,but It is expressed as:

[0121]

[0122] where α 0 , S TH and U TH are all hyperparameters, u d-1 It represents the lower threshold of oxygen supply concentration safety range on day d-1. u d It represents the lower threshold of the safe range of oxygen supply concentration on day d;

[0123] Design 2: If ΔSI d-1 >0,||ΔSI d-1 ||>S TH ,and but It is expressed as:

[0124]

[0125] where α 1 is a hyperparameter, It represents the lower threshold of oxygen supply concentration safety range on day d-1. It represents the lower threshold of the safe range of oxygen supply concentration on day d;

[0126] The hyperparameters are determined through the validation set data using the grid search method.

[0127] The basic optimization principle followed by the upper optimization design of this embodiment is consistent with the clinical hypoxia concentration control criteria, that is, when the adaptation performance declines, the hypoxia concentration should be increased to prevent further damage to the body, so the upper limit of the constraint is increased. On the contrary, when the adaptation performance is improved, it means that the body can get better adaptation under the current hypoxia concentration, so we will lower the lower limit of the constraint to test the feasibility of lower oxygen concentration and improve the rapidity of adaptation.

[0128] like Figure 2 and Figure 3 As shown, the decision direction of the method of this embodiment conforms to the basic rule of strategy adjustment, that is, when the performance decreases, the optimization result tends to increase the low oxygen concentration. This result verifies the feasibility of this embodiment.

[0129] S4: Evaluate the training results through the training effect evaluation module to evaluate whether the training effect reaches the required level, and feed back the evaluation results to the low oxygen concentration double-layer optimization module;

[0130] Step S4 specifically includes the following steps:

[0131] S41: After each training session, the data recorded by IHT can be used to calculate the training evaluation index SI, which is used to measure the individual's ability to maintain a high blood oxygen saturation at the lowest possible oxygen concentration. The training evaluation index SI is expressed as:

[0132]

[0133] in It represents the average area of ​​blood oxygen saturation in all hypoxic segments of IHT in one day. Indicates the average hypoxic concentration;

[0134] S42: comparing the training evaluation index of the first day with the training evaluation index of the last day to evaluate the overall training result;

[0135] S43: Changes in the evaluation index ΔSI based on the previous day's training d-1 Evaluate the training effect and calculate the change ΔSI of the previous day’s training evaluation index d-1Feedback is provided to the low oxygen concentration double-layer optimization module to guide the optimization of safety constraints;

[0136] In step S43, the change ΔSI of the training evaluation index of the previous day is calculated. d-1 It is expressed as:

[0137] ΔSI d-1 =SI d-1 -SI d-2 .

[0138] Therefore, the present invention adopts the above-mentioned individualized plateau hypoxia preparatory closed-loop intervention algorithm, which can use IHT data to calculate dynamic evaluation indicators, and use this as a direction to adjust the hypoxic stimulation intensity of IHT through control optimization strategies to improve the rapidity and effectiveness of plateau adaptation ability, accelerate the individual's IHT training cycle in plain areas, improve hypoxia tolerance and reduce the risk of acute mountain sickness.

[0139] Finally, it should be noted that the above embodiments are only used to illustrate the method scheme of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, ordinary method personnel in the field should understand that they can still modify or replace the method scheme of the present invention with equivalents, and these modifications or equivalent replacements cannot cause the modified method scheme to deviate from the spirit and scope of the method scheme of the present invention.

Claims

1. An individualized closed-loop intervention algorithm for high altitude hypoxia preparatory service, characterized in that: The following steps are involved: S1: During IHT, the subjects were given periodic oxygen supply, and physiological signals were collected in real time to obtain continuous data of oxygen supply concentration and blood oxygen saturation during IHT. The continuous data of blood oxygen saturation were divided into training set, validation set and test set according to different subjects. S2: Based on the training set, the Koopman operator is used to learn the changing pattern of blood oxygen saturation under intermittent hypoxia, and a group prediction model of blood oxygen saturation is constructed. The model is updated and predicted through online iteration, and the predicted value of blood oxygen saturation is output. Based on the validation set, the group prediction model is used for model individualization; S3: Use the individualized model to predict blood oxygen saturation, build a lower-level optimization cost function with low oxygen concentration as a variable through the dual-layer optimization module of supply and low oxygen concentration, and perform upper-level optimization design with safety constraints and self-adaptation based on historical learning performance; S4: The training results are evaluated through the training effect evaluation module, and the evaluation results are fed back to the low oxygen concentration double-layer optimization module.

2. According to claim 1, an individualized closed-loop intervention algorithm for high altitude hypoxia preparatory clothing is characterized in that: In step S2, based on the training set, the Koopman operator is used to learn the changing law of blood oxygen saturation under intermittent hypoxia, and a group prediction model of blood oxygen saturation is constructed, which specifically includes the following steps: Step 1: Obtain a nonlinear dynamic characteristic equation of blood oxygen saturation changing with oxygen concentration. The nonlinear dynamic characteristic equation is expressed as: x k+1 =f(x k ,u k ) where u k represents the oxygen concentration at sampling time k, x k represents the blood oxygen saturation at sampling time k; Step 2: Obtain the augmented state equation of the nonlinear dynamic characteristic equation. The augmented state equation is expressed as: X k =[x k ,u k ] · The augmented state equation is expanded to obtain the extended equation, which is expressed as: X k+1 =[f(x k ,u k ),Su k ] · Among them Su k =u k+1 , S represents the left shift operator; Step 3: The Koopman operator is obtained by extending the dynamic mode decomposition method, and the nonlinear dynamic characteristic equation in step 1 is converted into a linear dynamic characteristic equation by the Koopman operator.

3. According to claim 2, an individualized closed-loop intervention algorithm for high altitude hypoxia preparatory clothing is characterized in that: Step 3 specifically includes the following steps: Step 1: Get the state augmented matrix of blood oxygen saturation X and State augmented matrix of blood oxygen saturation X and Respectively expressed as: X =[X1,X2,…,X M-1 ]· Where M represents the total number of samples for constructing the population prediction model; Step 2: Augment the blood oxygen saturation state matrix X and Perform state improvement respectively and obtain the state matrix Φ( X )and The state matrix Φ( X )and Respectively expressed as: Φ( X )=[φ(X1),φ(X2),…,φ(X M-1 )] · Step 3: Get the group Koopman operator Group Koopman operator It is expressed as: Step 4: Get the linear form dynamic characteristic equation, which is expressed as:

4. The individualized closed-loop intervention algorithm for high altitude hypoxia training according to claim 1 is characterized in that: In step S2, the model is updated and predicted through online iteration, and the blood oxygen saturation prediction value is output. Based on the validation set, the model is individualized using the group prediction model, which specifically includes the following steps: S21: perform daily model updates online for the dth day; S22: estimated value based on the Koopman operator on day d-1 Iterate to get the estimated value of the Koopman operator on day d The iterative formula is expressed as: in and X d-1 Both represent the augmented state matrix obtained from the data of day d-1, γ d-1 represents the correction vector for day d-1, γ d-1 It is expressed as: c d-1 =Φ( X d-1 )C d where Γ d represents the covariance matrix of the dth day, Γ d It is expressed as: C d =C d-1 -C d-1 Φ( X d-1 )(Φ( X d-1 )C d-1 Φ( X d-1 )+I) -1 Φ( X d-1 ) · C d-1 。 5. According to claim 4, an individualized closed-loop intervention algorithm for high altitude hypoxia preparatory clothing is characterized in that: In step S22, when d is 1, X 0 , γ 0 and Γ 0 is the corresponding matrix obtained based on population data, Γ 0 It is expressed as: C 0 =(Φ( X )Φ( X ) · ) -1 。 6. The individualized closed-loop intervention algorithm for high altitude hypoxia preparatory service according to claim 1 is characterized in that: In step S3, a lower optimization cost function with low oxygen concentration as a variable is constructed through the low oxygen concentration double-layer optimization module. The lower optimization cost function is expressed as: X k =[x k ,u d ] T Where S y (u d ) indicates that during a hypoxic period, the supply of low oxygen concentration u d The area under the curve of the predicted blood oxygen saturation under the action of m It indicates the minimum value of blood oxygen saturation corresponding to the predicted hypoxic segment. ρ and μ are hyperparameters. k Indicates the blood oxygen saturation value corresponding to the predicted hypoxic segment. Indicates the upper threshold of low oxygen concentration. u (ΔSI d-1 ) indicates the lower threshold of low oxygen concentration.

7. The individualized closed-loop intervention algorithm for high altitude hypoxia training according to claim 6, characterized in that: In step S3, the upper-level optimization design of safety constraint adaptation is performed according to the historical learning performance. The upper-level optimization design is specifically as follows: Design 1: If ΔSI d-1 <0,||ΔSI d-1 ||>S TH , and ||u d-1 - u d-1 ||<U TH ,but It is expressed as: Among them, α0, S TH and U TH are all hyperparameters, u d-1 It represents the lower threshold of oxygen supply concentration safety range on day d-1. u d It represents the lower threshold of the safe range of oxygen supply concentration on day d; Design 2: If ΔSI d-1 >0,||ΔSI d-1 ||>S TH ,and but It is expressed as: Where α1 is a hyperparameter, It represents the lower threshold of oxygen supply concentration safety range on day d-1. Indicates the lower threshold of the safe range of oxygen supply concentration on day d.

8. The individualized closed-loop intervention algorithm for high altitude hypoxia training according to claim 7, characterized in that: The hyperparameters are determined through the validation set data using the grid search method.

9. The individualized closed-loop intervention algorithm for high altitude hypoxia training according to claim 1, characterized in that: Step S4 specifically includes the following steps: S41: Calculate the training evaluation index SI, which is expressed as: in It represents the average area of ​​blood oxygen saturation in all hypoxic segments of IHT in one day. Indicates the average hypoxic concentration; S42: comparing the training evaluation index of the first day with the training evaluation index of the last day to evaluate the overall training result; S43: Changes in the evaluation index ΔSI based on the previous day's training d-1 Evaluate the training effect and calculate the change ΔSI of the previous day’s training evaluation index d-1 Feedback is provided to the low oxygen concentration double-layer optimization module.

10. The individualized closed-loop intervention algorithm for high altitude hypoxia preparatory service according to claim 9, characterized in that: In step S43, the change ΔSI of the training evaluation index of the previous day is calculated. d-1 It is expressed as: ΔSI d-1 =YES d-1 -YES d-2 。

Images