Portable long-term health monitoring and self-adaptive oxygen supply system in plateau environment

Abstract

The invention discloses a portable long-term health monitoring and self-adaptive oxygen supply system in a plateau environment, belongs to the technical field of health monitoring, and solves the problems that in the prior art, an oxygen supply system can only supply oxygen, a monitoring system only provides a monitoring function, and the oxygen supply system and the monitoring system cannot cooperate to guarantee user safety in a closed loop mode. The system comprises a blood oxygen monitoring module, a monitoring cloud platform, a self-adaptive oxygen generation control module and a portable oxygen generation host, according to the embodiment of the invention, the blood oxygen monitoring module, the monitoring cloud platform, the self-adaptive oxygen generation control module and the portable oxygen generation host are cooperated, remote parameter monitoring and danger alarm functions are integrated, information such as environmental parameters, human body blood oxygen data, GPS and the like is collected and monitored in real time by using a 4G network, and when an abnormal condition is detected, the portable oxygen generation host is started. And the system can automatically give an alarm prompt to provide necessary help-seeking information, so that the safety of the user in the plateau environment is remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of health monitoring, and particularly relates to a portable long-term health monitoring and adaptive oxygen supply system in a plateau environment. Background Art

[0002] Outdoor activities in a plateau environment are threatened by altitude sickness caused by body hypoxia, leading to a series of diseases such as headache, nausea, insomnia, and fatigue. Altitude sickness is caused by body hypoxia, and the level of blood oxygen in the blood directly reflects the severity of human altitude sickness. When the blood oxygen level is too low, the functions of various systems of the human body may be severely affected, even endangering life. At the same time, timely oxygen supply is beneficial to relieve altitude sickness. By increasing the oxygen concentration in the blood and improving the oxygen supply to various organs of the body, the symptoms of altitude sickness can be alleviated, and the quality of life and action safety in a plateau environment can be improved. Therefore, monitoring the blood oxygen level and supplying oxygen in a timely manner is an important measure to ensure health in a plateau environment.

[0003] In addition, altitude sickness not only affects physical health but may also have a negative impact on mental state, such as causing symptoms such as anxiety, irritability, and inattention. These symptoms will further reduce an individual's action ability and decision-making ability in a plateau environment, increasing the risk of accidents. Especially in the case of physical labor or long-term outdoor activities, the threat of altitude sickness is more significant.

[0004] In the prior art, conventional mechanical oxygen supply valves usually cannot be adaptively adjusted according to the breathing actions and frequencies of users, resulting in oxygen waste and shortened oxygen supply time of high-pressure oxygen cylinders. Traditional oxygen supply systems still continuously supply oxygen during exhalation and when high oxygen flow rates are not required, resulting in low energy efficiency.

[0005] Chinese Patent CN219332868U discloses a high-pressure oxygen adaptive mechanical oxygen supply valve, which includes an air inlet joint, a decompression spring, a valve core, a housing I, a housing II, a boosting spring, a gasket, an orifice plate, a pilot valve plate, an air film, a housing III, a return spring, a pipe joint, etc.; High-pressure oxygen enters the oxygen supply valve through the air inlet joint, and after being decompressed to low-pressure oxygen, it flows into the cavity under the gasket through the orifice plate all the way, lifting the gasket to seal with the housing II, and the other way of low-pressure oxygen flows to the sealed part of the gasket. When inhaling, a negative pressure is generated in the lower cavity of the air film, the air film moves downward, the gas in the cavity under the gasket is discharged, and the gasket is opened under the action of the boosting spring to realize oxygen supply. When exhaling, the air film is closed under the action of the return spring, the pressure is built in the cavity under the gasket, and the gasket is closed to stop oxygen supply; However, the existing oxygen supply valve lacks portability and the ability to adaptively adjust to environmental changes. The existing oxygen supply system can only supply oxygen, and the monitoring system only provides monitoring functions. The oxygen supply system and the monitoring system cannot cooperate in a closed loop to ensure user safety. In view of the above problems, we propose a portable long-term health monitoring and adaptive oxygen supply system in a plateau environment. Summary of the Invention

[0006] The purpose of the present invention is to provide a portable long-term health monitoring and adaptive oxygen supply system in a plateau environment for the deficiencies of the existing technology, and solve the problems that the existing oxygen supply valve lacks portability and the ability to adaptively adjust to environmental changes, the existing oxygen supply system can only supply oxygen, the monitoring system only provides monitoring functions, and the oxygen supply system and the monitoring system cannot cooperate in a closed loop to ensure user safety.

[0007] The present invention is realized as follows. A portable long-term health monitoring and adaptive oxygen supply system in a plateau environment, the portable long-term health monitoring and adaptive oxygen supply system in the plateau environment includes:

[0008] A blood oxygen monitoring module for real-time monitoring and obtaining the user's blood oxygen monitoring data and uploading the blood oxygen monitoring data in real time;

[0009] A monitoring cloud platform, which is used to obtain blood oxygen monitoring data, oxygen generation related data, and environmental monitoring data, analyze and store the blood oxygen monitoring data and oxygen generation related data. The blood oxygen monitoring module establishes a data transmission channel with the monitoring cloud platform through a low-power Bluetooth module, and the sampling frequency ≥ 10Hz;

[0010] Preferably, the monitoring cloud platform includes:

[0011] A data acquisition unit, which is respectively communicatively connected with an adaptive oxygen generation control module, a blood oxygen monitoring module, and a portable oxygen generation host. The data acquisition unit is used to acquire environmental monitoring data and blood oxygen monitoring data;

[0012] A GPS module, which is used to locate the user's position in real time, and the GPS module is communicatively connected to the blood oxygen monitoring module;

[0013] A high-altitude warning unit, which analyzes and monitors the environmental monitoring data and blood oxygen monitoring data based on preset alarm rules, and conducts adaptive warning of high-altitude reactions through a fitting curve of the relationship between blood oxygen value, heart rate value, high-altitude reaction and time, and triggers a high-altitude warning instruction.

[0014] An adaptive oxygen generation control module, which is used to monitor the environmental monitoring data and blood oxygen monitoring data in real time, dynamically adjust the pressurization time and oxygen generation-related data, realize adaptive oxygen supply, and generate a personalized oxygen output mode. The adaptive oxygen generation control module consists of an environmental parameter acquisition unit and a control adjustment unit;

[0015] Among them: The environmental parameter acquisition unit collects environmental monitoring data in real time based on sensors. Among them, the environmental monitoring data includes environmental air pressure and environmental oxygen concentration;

[0016] The control adjustment unit is used to load the environmental monitoring data, calculate the compressor pressurization time based on the environmental monitoring data, and adjust the pressurization time and oxygen generation-related data in combination with the compressor pressurization time and the PID control algorithm.

[0017] A portable oxygen generation main unit, which is used to obtain a personalized oxygen output mode, uses a double-tower pressure swing adsorption technology to separate oxygen from the air, and provides a high-purity oxygen output.

[0018] Preferably, the method of adjusting the pressurization time and oxygen generation-related data in combination with the compressor pressurization time and the PID control algorithm specifically includes:

[0019] Load the environmental monitoring data and identify the environmental air pressure P env and the environmental oxygen concentration C env ;

[0020] Set the target oxygen concentration C t , and the target oxygen concentration is pure oxygen or a preset concentration;

[0021] According to the collected environmental air pressure and oxygen concentration, use formula (1) to calculate the required pressurization time T p ;

[0022]

[0023] Among them, in formula (1), k is an empirical constant, and P s is the standard atmospheric pressure;

[0024] According to the calculated pressurization time T p , adjust the operation time of the compressor to ensure the expected oxygen concentration during the oxygen generation process;

[0025] Through a feedback mechanism, the system monitors the actual concentration C of the output oxygen in real time o , and compares it with the target oxygen concentration C t . If C o does not meet C t , the pressurization time and oxygen generation related data are further adjusted through the PID control algorithm. The feedback mechanism monitors the actual concentration C of the output oxygen by establishing a feedback network o , and feeds back the monitored actual concentration C of the output oxygen o to the PID controller. The PID controller calculates the error between the target oxygen concentration C t and the actual concentration C o , and updates the parameters of the feedback network according to the backpropagation rule to assist the PID control algorithm in adjusting the pressurization time and oxygen generation related data;

[0026]

[0027] Among them, formula (2) is used to adjust the pressurization time and oxygen generation related data, eliminate the static error of the pressurization time and oxygen generation related data, and improve the deadbeat of the compressor. In formula (2), ΔT p represents the output signal of the PID controller, e(t) = C t -C o is the error, and K p , K i , K d are the proportional, integral and differential coefficients respectively.

[0028] Preferably, the adaptive oxygen generation control module further includes:

[0029] An output mode generation module, which generates a personalized oxygen output mode based on the pressurization time and oxygen generation related data;

[0030] The output mode generation module includes:

[0031] A constant oxygen supply unit, which generates a constant oxygen concentration supply mode based on a preset oxygen concentration value;

[0032] An adaptive oxygen supply unit, which generates an adaptive oxygen supply mode based on a time neural network model combined with the actual blood oxygen level;

[0033] Among them, when generating an adaptive oxygen supply mode based on a time neural network model combined with the actual blood oxygen level, the real-time blood oxygen value, historical blood oxygen value and environmental oxygen concentration data are input into the time neural network model, and the output of the time neural network model is the adjusted oxygen concentration parameter C t, according to the output oxygen concentration, the portable oxygen generator host dynamically adjusts the oxygen output concentration. The time neural network model is based on the Transformer model as the main framework, and uses the sharpness-aware minimization of the SAMformer model to optimize and improve the Transformer model. The Transformer model includes two groups of encoders and decoders, and the encoders and decoders respectively contain self-attention mechanisms and feed-forward neural networks.

[0034] Preferably, the training method of the time neural network model specifically includes:

[0035] Using the Transformer model as the initial model of the time neural network model, the Transformer model includes two groups of encoders and decoders. The encoders and decoders respectively contain self-attention mechanisms and feed-forward neural networks. The self-attention mechanism calculates the attention weights of each position through linear transformations of queries, keys, and values, so as to extract global context information;

[0036] Using the sharpness-aware minimization of the SAMformer model to optimize and improve the Transformer model;

[0037] Obtaining temporal training data and test data, using the training data as input, iteratively training the Transformer model through the training data, and outputting a converged time neural network model;

[0038] Loading the test data, using the test data as input, executing the time neural network model, outputting the test result, and judging whether the test result is qualified based on the preset test accuracy. If the test result is qualified, output the trained time neural network model.

[0039] Preferably, the method of using the sharpness-aware minimization of the SAMformer model to optimize and improve the Transformer model includes:

[0040] Defining a linear generation model based on the SAMformer model, and the linear generation model is used to simulate temporal prediction. Among them, the output of the linear generation model is expressed as:

[0041] Y = XW + ε(3)

[0042] Where, X represents the input matrix (sequence length L × sequence dimension D), W represents the model weight parameter, ε represents the loss, and Y is the model output;

[0043] Introducing a simplified Transformer encoder into the Transformer model. The Transformer encoder is composed of an attention mechanism, a residual connection layer, and a linear layer. The residual connection layer and the linear layer are one layer and three layers respectively. The Transformer encoder is expressed as:

[0044] f(X) = [X + A(X)XW V W O W(4)

[0045] Wherein, d m is the model dimension, and A(X) is the attention matrix of the input sequence:

[0046]

[0047] Wherein,

[0048] The Transformer encoder is modified based on formula (4) using the SAMformer model. The invertible instance normalization is applied to the input X, and the SAMformer model is used to optimize the Transformer model, making the Transformer model tend to a flatter local minimum to complete the optimization and improvement.

[0049] Preferably, the portable oxygen generator host includes:

[0050] An oxygen supply pipe for supplying oxygen to the user;

[0051] Medical molecular sieve, installed in the main housing, for filtering and purifying air;

[0052] An air compressor fixedly arranged in the main housing, for compressing air and assisting in oxygen production;

[0053] Molecular adsorption tower A;

[0054] Molecular adsorption tower B. Molecular adsorption tower A and molecular adsorption tower B are respectively fixedly installed in the main housing. Molecular adsorption tower A and molecular adsorption tower B are respectively filled with adsorbents. Molecular adsorption tower A and molecular adsorption tower B cooperate with each other, and are used for alternately performing adsorption and desorption operations to achieve continuous oxygen production;

[0055] A radiator, which is communicated with the air compressor. One end of the radiator away from the air compressor is connected with a solenoid valve, and the solenoid valve is respectively communicated with molecular adsorption tower A and molecular adsorption tower B;

[0056] A gas storage tank, which is respectively communicated with molecular adsorption tower A and molecular adsorption tower B, and one end of the gas storage tank is also communicated with the oxygen supply pipe.

[0057] Compared with the prior art, the embodiments of the present application mainly have the following beneficial effects:

[0058] In the embodiments of the present invention, through the collaborative cooperation of a blood oxygen monitoring module, a monitoring cloud platform, an adaptive oxygen generation control module, and a portable oxygen generation host, and based on an adaptive oxygen supply method for the human blood oxygen level, by real-time monitoring of blood oxygen data, precise adjustment of the oxygen supply amount is achieved. Compared with the traditional oxygen supply method based on respiration detection, it not only improves the oxygen supply accuracy, but also extends the service time and battery life of the system. At the same time, it integrates remote parameter monitoring and danger alarm functions, uses the 4G network to collect and monitor environmental parameters, human blood oxygen data, GPS and other information in real time. When an abnormal situation is detected, the system will automatically send an alarm reminder and provide necessary distress information, significantly enhancing the safety of users in the plateau environment.

[0059] In the embodiments of the present invention, through an intelligent adaptive oxygen generation control algorithm, the portable oxygen generation host can efficiently and stably provide high-purity oxygen in various environments, and is especially suitable for areas with drastic changes in air pressure and oxygen concentration such as plateaus. The adaptive algorithm not only improves the oxygen generation efficiency, but also ensures the safety and comfort of users in different environments.

[0060] The portable oxygen generation host in the embodiments of the present invention is designed with two innovative oxygen output modes, namely the constant oxygen concentration oxygen supply mode and the adaptive oxygen supply mode based on the actual blood oxygen level. These modes are designed to provide efficient and personalized oxygen supply solutions for users in different environments and usage scenarios. The introduction of these two oxygen supply modes enables the portable oxygen generation host of the present invention to provide efficient and safe oxygen supply solutions under different user needs and usage environments. The constant oxygen concentration oxygen supply mode ensures stable oxygen output for a long time and is suitable for long-term outdoor activities and continuous medical treatments; while the adaptive oxygen supply mode can dynamically adjust the oxygen supply according to the actual physiological state of the user and environmental changes, ensuring the best blood oxygen support in various situations. The combination of these functions not only improves the practicality and flexibility of the device, but also provides an important guarantee for the health and safety of users. Brief Description of the Drawings

[0061] Figure 1 It is a schematic structural diagram of a portable long-term health monitoring and adaptive oxygen supply system in a plateau environment provided by the present invention.

[0062] Figure 2 It is a schematic implementation flow diagram of a method for combining the compressor pressurization time, adjusting the pressurization time by the PID control algorithm, and oxygen generation related data provided by the present invention.

[0063] Figure 3 It is a schematic implementation flow diagram of a method for training a time neural network model provided by the present invention.

[0064] Figure 4It is the architecture diagram of the SAMformer model provided by the present invention.

[0065] Figure 5 It is the schematic structural diagram provided by the present invention.

[0066] In the figure: 100 - blood oxygen monitoring module, 200 - monitoring cloud platform, 210 - data acquisition unit, 220 - GPS module, 230 - plateau early warning unit, 300 - adaptive oxygen generation control module, 310 - environmental parameter acquisition unit, 320 - control and adjustment unit, 330 - output mode generation module, 331 - constant oxygen supply unit, 332 - adaptive oxygen supply unit, 400 - portable oxygen generation main unit. Detailed implementation manners

[0067] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this application belongs; the terms used in the specification of this application are only for the purpose of describing specific embodiments and are not intended to limit this application; the terms "including" and "having" and any variations thereof in the specification and claims of this application and the above drawings are intended to cover non-exclusive inclusion. The terms "first", "second", etc. in the specification and claims of this application or the above drawings are used to distinguish different objects and are not used to describe a specific order.

[0068] The existing oxygen supply valves lack portability and the ability to adaptively adjust to environmental changes. The existing oxygen supply systems can only supply oxygen, and the monitoring systems only provide monitoring functions. The oxygen supply systems and the monitoring systems cannot cooperate in a closed loop to ensure the safety of users. In response to the above problems, we have proposed a portable long-term health monitoring and adaptive oxygen supply system in a plateau environment. The system consists of a blood oxygen monitoring module 100, a monitoring cloud platform 200, an adaptive oxygen generation control module 300, and a portable oxygen generation host 400. Briefly, when the system is implemented, first, the blood oxygen monitoring module 100 (which can be a wristband-type or fingertip-type blood oxygen monitoring device) monitors and obtains the user's blood oxygen monitoring data in real time. The blood oxygen monitoring module 100 establishes a data transmission channel with the monitoring cloud platform 200 through a low-power Bluetooth module and uploads the blood oxygen monitoring data in real time. The data acquisition unit 210 in the monitoring cloud platform 200 is communicatively connected to the adaptive oxygen generation control module 300, the blood oxygen monitoring module 100, and the portable oxygen generation host 400 respectively, and real-time locates the user's position through the GPS module 220. The GPS module 220 is communicatively connected to the blood oxygen monitoring module 100. Subsequently, the plateau warning unit 230 in the monitoring cloud platform 200 analyzes and monitors the environmental monitoring data and the blood oxygen monitoring data based on the preset alarm rules, and performs an adaptive warning of altitude sickness through the fitting curve of the relationship between the blood oxygen value, heart rate value, altitude sickness, and time, triggering a plateau warning instruction. Subsequently, the adaptive oxygen generation control module 300 monitors the environmental monitoring data and the blood oxygen monitoring data in real time, dynamically adjusts the pressurization time and the oxygen generation-related data, realizes adaptive oxygen supply, and generates a personalized oxygen output mode. Finally, the portable oxygen generation host 400 obtains the personalized oxygen output mode, uses the dual-tower pressure swing adsorption technology to separate oxygen from the air, and provides high-purity oxygen output. In terms of emergency handling, when receiving a low blood oxygen alarm, the user can also choose to manually adjust the oxygen output to increase the oxygen supply. The application provides the accurate position of the user through GPS positioning, helping rescue personnel quickly find the user and provide emergency medical assistance. The entire system is designed to be highly portable and is suitable for plateau outdoor activities such as mountain climbing and hiking, as well as occasions that require long-term health monitoring. After completing the activity, the user turns off the blood oxygen monitoring device and the oxygen generation host and puts away the devices. All monitoring data is automatically saved to the cloud platform, and the user can access these data at any time for review and analysis. At the same time, the user should check the device status and perform necessary maintenance and charging for the next use. Through the above process, the system provides users with comprehensive health monitoring and adaptive oxygen supply support in plateau outdoor adventures. The portability and real-time data analysis ability of the system ensure that users can obtain timely and accurate health protection in harsh environments, effectively preventing and coping with health problems such as altitude sickness.In the embodiments of the present invention, through the collaborative cooperation of the blood oxygen monitoring module 100, the monitoring cloud platform 200, the adaptive oxygen generation control module 300, and the portable oxygen generation host 400, and based on the adaptive oxygen supply method for the human blood oxygen level, by real-time monitoring of blood oxygen data, precise adjustment of the oxygen supply amount is achieved. Compared with the traditional oxygen supply method based on respiration detection, it not only improves the oxygen supply accuracy, but also extends the service time and battery life of the system. At the same time, it integrates the functions of remote parameter monitoring and danger alarm, uses the 4G network to collect and monitor environmental parameters, human blood oxygen data, GPS and other information in real time. When abnormal situations are detected, the system will automatically send out alarm reminders and provide necessary distress information, significantly enhancing the safety of users in the plateau environment.

[0069] The embodiments of the present invention provide a portable long-term health monitoring and adaptive oxygen supply system in the plateau environment, as Figure 1 shown. The portable long-term health monitoring and adaptive oxygen supply system in the plateau environment specifically includes:

[0070] A blood oxygen monitoring module 100, which is used to monitor and obtain the user's blood oxygen monitoring data in real time and upload the blood oxygen monitoring data in real time;

[0071] The blood oxygen monitoring module 100 can be a wristband-type or fingertip-type blood oxygen monitoring device, and the blood oxygen monitoring device can continuously monitor the user's blood oxygen saturation (SpO2). Through the 4G network, these data are transmitted to the cloud platform in real time for storage and analysis to help identify and prevent health problems such as altitude sickness caused by low blood oxygen. The real-time monitoring and analysis of blood oxygen data are important means to ensure the health of users, especially in special environments such as plateaus.

[0072] A monitoring cloud platform 200, which is communicatively connected to the blood oxygen monitoring module 100. The monitoring cloud platform 200 is used to obtain blood oxygen monitoring data, oxygen generation-related data, and environmental monitoring data, and analyze and store the blood oxygen monitoring data and oxygen generation-related data;

[0073] It should be noted that the blood oxygen monitoring module 100 establishes a data transmission channel with the monitoring cloud platform 200 through a low-power Bluetooth module, and the sampling frequency ≥ 10Hz;

[0074] An adaptive oxygen generation control module 300, which is used to monitor the environmental monitoring data and blood oxygen monitoring data in real time, dynamically adjust the pressurization time and oxygen generation-related data, realize adaptive oxygen supply, and generate a personalized oxygen output mode;

[0075] A portable oxygen generation host 400, which is used to obtain the personalized oxygen output mode, and uses the dual-tower pressure swing adsorption technology to separate oxygen from the air and provide high-purity oxygen output.

[0076] In an embodiment of the present invention, the monitoring cloud platform 200 includes:

[0077] A data acquisition unit 210, which is communicatively connected to the adaptive oxygen generation control module 300, the blood oxygen monitoring module 100, and the portable oxygen generation host 400 respectively. The data acquisition unit 210 is used to acquire environmental monitoring data and blood oxygen monitoring data;

[0078] A GPS module 220, which is used to locate the user's position in real time. The GPS module 220 is communicatively connected to the blood oxygen monitoring module 100;

[0079] A high altitude warning unit 230, which analyzes and monitors the environmental monitoring data and blood oxygen monitoring data based on a preset alarm rule, and performs adaptive warning of high altitude reaction through a fitting curve of the relationship between blood oxygen value, heart rate value, high altitude reaction, and time, triggering a high altitude warning instruction.

[0080] In terms of remote monitoring and alarm, the monitoring cloud platform 200 analyzes the uploaded data to monitor the user's health status and the operating status of the device. If an abnormal situation is detected (such as a sharp drop in blood oxygen level or deterioration of environmental conditions), the monitoring cloud platform 200 will immediately send an alarm to the user through the application, and at the same time notify the preset emergency contact person or medical service provider. The monitoring cloud platform 200 regularly generates the user's health report, providing detailed historical data and analysis results. The user can view and download these reports to understand their own health status.

[0081] In terms of emergency handling, when receiving a low blood oxygen alarm, the user can choose to manually adjust the oxygen output to increase the oxygen supply. The application provides the user's accurate location through GPS positioning to help rescue personnel quickly find the user and provide emergency medical assistance. The entire system is designed to be highly portable, suitable for high altitude outdoor activities such as mountaineering and hiking, as well as occasions that require long-term health monitoring.

[0082] After completing the activity, the user turns off the blood oxygen monitoring device and the oxygen generation host, and puts away the devices. All monitoring data is automatically saved to the cloud platform, and the user can access these data at any time for review and analysis. At the same time, the user should check the device status and perform necessary maintenance and charging for the next use.

[0083] Through the above process, the system provides users with comprehensive health monitoring and adaptive oxygen supply support in high altitude outdoor adventures. The portability and real-time data analysis ability of the system ensure that users can obtain timely and accurate health protection in harsh environments, effectively preventing and coping with health problems such as high altitude reaction.

[0084] It should be noted that the monitoring cloud platform 200 can use the 4G network to achieve real-time data transmission and processing. First of all, through the 4G network, the system can collect various environmental parameters in real time, including temperature, humidity, air pressure, and air quality, etc. These data are crucial for analyzing the impact of high-altitude or other harsh environments on human health. For example, in high-altitude areas, changes in air pressure and oxygen concentration will directly affect the blood oxygen level of the human body.

[0085] In addition, the operating status and output parameters of the portable oxygen generator host 400 are also uploaded to the cloud platform in real time through the 4G network. This includes information such as oxygen flow rate, oxygen supply time, and oxygen concentration. Through these data, the working conditions of the portable oxygen generator host 400 can be remotely monitored to ensure that users obtain sufficient oxygen support in high-altitude environments. Real-time monitoring of oxygen supply data can not only improve the usage efficiency of the equipment but also ensure that users can get timely oxygen supply when needed.

[0086] In this embodiment, the blood oxygen monitoring module 100, the adaptive oxygen generation control module 300, the portable oxygen generator host 400, and the monitoring cloud platform 200 can be connected through Bluetooth, 4G, or DTU communication. The blood oxygen monitoring module 100 can be a wristband-type or fingertip-type blood oxygen monitoring device. The blood oxygen monitoring module 100 also includes environmental parameter sensors, and the environmental parameter sensors include, but are not limited to, air pressure, temperature, and humidity sensors. The environmental parameter sensors are fixed on the equipment carried by the user to monitor the external environmental conditions in real time.

[0087] The monitoring cloud platform 200 also has a built-in GPS module 220, which can locate the user's position in real time. Combining environmental parameters and health data, the GPS information helps to comprehensively understand the specific environment and geographical conditions where the user is located, and provide more accurate health monitoring and early warning services. The geographical location data combined with environmental and health parameters can provide more personalized health management suggestions for users.

[0088] Finally, the monitoring cloud platform 200 uses preset alarm rules to analyze and monitor the data uploaded in real time. When abnormal situations (such as too low blood oxygen level, environmental parameters exceeding the safe range, etc.) are detected, the system will immediately trigger an alarm and send a reminder to the user or medical service provider. This real-time early warning mechanism can significantly improve the emergency response speed and take timely measures to ensure the health and safety of users.

[0089] The wireless data acquisition system of the monitoring cloud platform 200 realizes the all-round monitoring and management of the user's health status by integrating technologies such as 4G network transmission, environmental parameter monitoring, blood oxygen monitoring, oxygen supply data acquisition, and GPS positioning. Through the intelligent analysis and early warning mechanism in the background, potential health risks can be detected and addressed in a timely manner, providing users with safe and reliable health protection.

[0090] In high-altitude outdoor adventures, the portable long-term health monitoring and adaptive oxygen supply system provides comprehensive health monitoring and oxygen supply support. Users first wear a wristband or fingertip blood oxygen monitoring device, ensuring good contact between the device and the skin for accurate measurement of blood oxygen saturation. At the same time, place the portable miniaturized oxygen generator host in a backpack or hang it on the waistband, ensuring that the device's battery is fully charged and the oxygen reserve is sufficient.

[0091] In the embodiments of the present invention, through the coordinated cooperation of the blood oxygen monitoring module 100, the monitoring cloud platform 200, the adaptive oxygen generation control module 300, and the portable oxygen generator host 400, and based on the adaptive oxygen supply method for the human body's blood oxygen level, by real-time monitoring of blood oxygen data, precise adjustment of the oxygen supply amount is achieved. Compared with the traditional oxygen supply method based on respiratory detection, it not only improves the accuracy of oxygen supply but also extends the usage time and battery life of the system. At the same time, it integrates remote parameter monitoring and danger alarm functions, using the 4G network to collect and monitor environmental parameters, human blood oxygen data, GPS and other information in real time. When an abnormal situation is detected, the system will automatically send an alarm reminder and provide necessary distress information, significantly enhancing the safety of users in the high-altitude environment.

[0092] As Figure 1 shown, the adaptive oxygen generation control module 300 consists of an environmental parameter acquisition unit 310 and a control adjustment unit 320;

[0093] Among them: The environmental parameter acquisition unit 310, based on sensors, real-time collects environmental monitoring data, where the environmental monitoring data includes environmental air pressure and environmental oxygen concentration;

[0094] The control adjustment unit 320 is used to load the environmental monitoring data, calculate the compressor pressurization time based on the environmental monitoring data, and adjust the pressurization time and oxygen generation-related data in combination with the compressor pressurization time and the PID control algorithm.

[0095] The output mode generation module 330 generates a personalized oxygen output mode based on the pressurization time and oxygen generation-related data.

[0096] The output mode generation module 330 includes:

[0097] The constant oxygen supply unit 331 generates a constant oxygen concentration oxygen supply mode based on a preset oxygen concentration value;

[0098] In the constant oxygen concentration oxygen supply mode, users can achieve stable oxygen output by setting a preset oxygen concentration value. This mode is particularly suitable for application scenarios that require maintaining a specific oxygen concentration for a long time, such as long-term outdoor activities in high-altitude environments or patients with specific health needs. The device first collects the current air pressure P env and oxygen concentration C env, while reading the target oxygen concentration C set by the user t . Through the built-in PID control algorithm, the device adjusts the pressurization time according to the actually output oxygen concentration C o to ensure that the output is stable at the preset target value C t . If the actual concentration deviates from the target value, the device will adjust the working time of the compressor in real time to keep the output concentration at the set ideal level.

[0099] The adaptive oxygen supply unit 332 generates an adaptive oxygen supply mode based on a time neural network model combined with the actual blood oxygen level;

[0100] The portable oxygen generation host 400 of the embodiment of the present invention is designed with two innovative oxygen output modes, namely the constant oxygen concentration oxygen supply mode and the adaptive oxygen supply mode based on the actual blood oxygen level. These modes are designed to provide efficient and personalized oxygen supply solutions for users in different environments and usage scenarios. The introduction of these two oxygen supply modes enables the portable oxygen generation host 400 of the present invention to provide efficient and safe oxygen supply solutions under different user needs and usage environments. The constant oxygen concentration oxygen supply mode ensures a stable oxygen output for a long time and is suitable for long-term outdoor activities and continuous medical treatments; while the adaptive oxygen supply mode can dynamically adjust the oxygen supply according to the actual physiological state of the user and environmental changes to ensure the best blood oxygen support in various situations. The combination of these functions not only improves the practicability and flexibility of the device, but also provides an important guarantee for the health and safety of users.

[0101] Among them, when generating the adaptive oxygen supply mode based on the time neural network model combined with the actual blood oxygen level, the real-time blood oxygen value, historical blood oxygen value, and environmental oxygen concentration data are input into the time neural network model, and the output of the time neural network model is the adjusted oxygen concentration parameter C t , and the portable oxygen generation host 400 dynamically adjusts the oxygen output concentration according to the output oxygen concentration.

[0102] The embodiment of the present invention provides a method for combining the compressor pressurization time, PID control algorithm to adjust the pressurization time, and oxygen generation related data, as Figure 2 shown, the method for combining the compressor pressurization time, PID control algorithm to adjust the pressurization time, and oxygen generation related data specifically includes:

[0103] S101, load the environmental monitoring data and identify the environmental air pressure P env and environmental oxygen concentration C env ;

[0104] S102, set the target oxygen concentration C t , and the target oxygen concentration is pure oxygen or a preset concentration;

[0105] S103. Calculate the required pressurization time T using formula (1) according to the collected ambient air pressure and oxygen concentration p ;

[0106]

[0107] wherein, in formula (1), k is an empirical constant, and P s is the standard atmospheric pressure;

[0108] S104. Adjust the operating time of the compressor according to the calculated pressurization time T p to ensure that the expected oxygen concentration is reached during the oxygen generation process;

[0109] S105. Through a feedback mechanism, the system continuously monitors the actual concentration C o of the output oxygen and compares it with the target oxygen concentration C t . If C o does not meet C t , then further adjust the pressurization time and oxygen generation related data through the PID control algorithm. The feedback mechanism monitors the actual concentration C o of the output oxygen by establishing a feedback network, and feeds back the monitored actual concentration C o of the output oxygen to the PID controller. The PID controller calculates the error between the target oxygen concentration C t and the actual concentration C o , and updates the parameters of the feedback network according to the backpropagation rule to assist the PID control algorithm in adjusting the pressurization time and oxygen generation related data;

[0110]

[0111] wherein, formula (2) is used to adjust the pressurization time and oxygen generation related data, eliminate the static error of the pressurization time and oxygen generation related data, and improve the deadbeat of the compressor. In formula (2), ΔT p represents the output signal of the PID controller, e(t) = C t -C o is the error, and K p , K i , K d are the proportional, integral, and differential coefficients respectively.

[0112] In this embodiment, an adaptive oxygen generation control algorithm based on ambient air pressure and ambient oxygen concentration is provided to ensure stable oxygen output at different altitudes. This algorithm dynamically adjusts the pressurization time and other key parameters during the oxygen generation process by continuously monitoring the ambient air pressure and oxygen concentration, achieving efficient and stable oxygen supply.

[0113] At different altitudes, the ambient air pressure and oxygen concentration change significantly. The adaptive algorithm continuously adjusts the pressurization time and other oxygen generation parameters, so that the output oxygen concentration clock is stably within the target range.

[0114] The adaptive oxygen generation control algorithm based on ambient air pressure and ambient oxygen concentration provides a clear control process, covering the whole process of ambient parameter acquisition, calculation of pressurization time, PID control adjustment and adaptive regulation. The algorithm pseudocode is shown in the following table:

[0115]

[0116] In the embodiments of the present invention, through the intelligent adaptive oxygen generation control algorithm, the portable oxygen generation host 400 can efficiently and stably provide high-purity oxygen in various environments, and is especially suitable for areas with drastic changes in air pressure and oxygen concentration such as plateaus. The adaptive algorithm not only improves the oxygen generation efficiency, but also ensures the use safety and comfort of users in different environments.

[0117] The embodiments of the present invention provide a training method for a time neural network model, as Figure 3 shown, the training method of the time neural network model specifically includes:

[0118] S201, taking the Transformer model as the initial model of the time neural network model. The Transformer model includes two groups of encoders and decoders. The encoders and decoders respectively contain self-attention mechanisms and feed-forward neural networks. The self-attention mechanism calculates the attention weights of each position through linear transformations of queries, keys and values, so as to extract global context information;

[0119] S202, using the sharpness-aware minimization of the SAMformer model to optimize and improve the Transformer model;

[0120] S203, obtaining sequential training data and test data, taking the training data as input, iteratively training the Transformer model through the training data, and outputting a converged time neural network model;

[0121] S204, loading the test data, taking the test data as input, executing the time neural network model, and outputting the test result;

[0122] S205, judging whether the test result is qualified based on a preset test accuracy;

[0123] S206, if the test result is qualified, outputting the trained time neural network model;

[0124] If the test result is unqualified, return to S203.

[0125] In this embodiment, the Time Neural Network (TNN) model plays a key role in the adaptive oxygen supply mode based on the actual blood oxygen level. It analyzes the user's real-time blood oxygen value, historical blood oxygen value, and environmental oxygen concentration to predict the user's oxygen demand for a period of time in the future and dynamically adjusts the oxygen output concentration. The device obtains real-time blood oxygen value, historical blood oxygen value, and environmental oxygen concentration data through Bluetooth communication technology, and performs data preprocessing on them, including normalization and filtering, to ensure the quality and consistency of the input data. The time neural network model takes the Transformer model as the main framework and uses the sharpness-aware minimization of the SAMformer model to optimize and improve the Transformer model. The Transformer model includes two groups of encoders and decoders. The encoders and decoders respectively contain self-attention mechanisms and feed-forward neural networks. When the portable oxygen generator host 400 dynamically adjusts the oxygen output concentration, it first preset the target oxygen concentration, and then based on the environmental monitoring data collected by the sensor in real time, according to the collected environmental air pressure and oxygen concentration, calculate the required pressurization time T p , according to the calculated pressurization time T p , adjust the running time of the compressor to ensure that the expected oxygen concentration is reached during the oxygen generation process. Through the feedback mechanism, the system real-time monitors the actual concentration C of the output oxygen o , and compare it with the target oxygen concentration C t . If C o does not meet C t , then further adjust the pressurization time and oxygen generation related data through the PID control algorithm.

[0126] In this embodiment, the method for optimizing and improving the Transformer model by using the sharpness-aware minimization of the SAMformer model includes:

[0127] S2021, define a linear generation model based on the SAMformer model. The linear generation model is used to simulate time series prediction. Among them, the output of the linear generation model is expressed as:

[0128] Y = XW + ε (3)

[0129] Among them, X represents the input matrix (sequence length L × sequence dimension D), W represents the model weight parameter, ε represents the loss, and Y is the model output;

[0130] S2022, introduce a simplified Transformer encoder into the Transformer model. The Transformer encoder is composed of an attention mechanism, a residual connection layer, and a linear layer. The residual connection layer and the linear layer are one layer and three layers respectively. The Transformer encoder is expressed as:

[0131] f(X) = [X + A(X)XW V W O W(4)

[0132] wherein, d m is the model dimension, and A(X) is the attention matrix of the input sequence:

[0133]

[0134] wherein, softmax makes A(X) have correct randomness, and each row describes a probability distribution. Using the SAMformer model;

[0135] In S2023, the Transformer encoder is modified based on formula (4), reversible instance normalization is applied to the input X, and the SAMformer model is used to optimize the Transformer model, making the Transformer model tend to a flatter local minimum, completing the optimization of the Transformer model. SAMformer is based on formula (4) and has made two important modifications. First, reversible instance normalization is applied to the input X because this technique has been proven effective in dealing with the offset between time series training data and test data. Second, the SAMformer (Sharpness-Aware Minimization) is used to optimize the model to make it tend to a flatter local minimum. The SAMformer model architecture of a shallow transformer model with an encoder is as Figure 4 shown.

[0136] SAMformer uses the D×D matrix represented in formula (4) for channel attention instead of the spatial or temporal attention of the L×L matrix used in other models. It ensures the invariance of feature permutation, eliminates the need for position encoding that is usually required before the attention layer, and reduces the time and memory complexity because D ≤ L in most actual datasets. Our channel attention checks the average influence of each feature on other features at all time steps.

[0137] An embodiment of the present invention provides a portable oxygen generation host 400, as Figure 5 shown, the portable oxygen generation host 400 specifically includes:

[0138] An oxygen supply tube for supplying oxygen to a user;

[0139] Medical molecular sieves installed in the main housing for filtering and purifying air;

[0140] An air compressor fixedly installed inside the main housing, which is used to compress air and assist in oxygen production;

[0141] Molecular adsorption tower A;

[0142] Molecular adsorption tower B. Molecular adsorption tower A and molecular adsorption tower B are respectively fixedly installed inside the main housing. Molecular adsorption tower A and molecular adsorption tower B are respectively filled with adsorbents. Molecular adsorption tower A and molecular adsorption tower B cooperate with each other and are used to alternately perform adsorption and desorption operations to achieve continuous oxygen production.

[0143] Molecular adsorption tower A and molecular adsorption tower B form a dual-column pressure swing adsorption (PSA) technology to separate oxygen from the air and provide high-purity oxygen output. The main unit is designed to be compact and portable, suitable for high altitudes and other environments where portable oxygen supply is required. Its miniaturized design ensures the portability of the device, enabling users to obtain oxygen supply anytime and anywhere, especially suitable for use in high-altitude areas.

[0144] The dual-column pressure swing adsorption technology operates alternately through two adsorption towers (molecular adsorption tower A and molecular adsorption tower B), performing adsorption and desorption operations respectively to achieve continuous oxygen production. First, in the adsorption stage, air is compressed by the air compressor to a certain pressure and then enters molecular adsorption tower A or molecular adsorption tower B. Molecular adsorption tower A and molecular adsorption tower B are filled with specific adsorbents (such as molecular sieves). Under high-pressure conditions, the adsorbent has a stronger adsorption capacity for nitrogen in the air, while oxygen is enriched due to the selective adsorption of the adsorbent and is discharged from the upper end of the adsorption tower as high-purity oxygen product. In the desorption stage, the pressure of molecular adsorption tower A and molecular adsorption tower B is rapidly reduced to near atmospheric pressure or lower. Under low-pressure conditions, the adsorbent releases the previously adsorbed nitrogen, and this process is called desorption or regeneration. The desorbed nitrogen is discharged through the exhaust port. After the desorption stage ends, molecular adsorption tower A and molecular adsorption tower B are ready to perform adsorption operations again. Molecular adsorption tower A and molecular adsorption tower B alternately perform adsorption and desorption operations. When one adsorption tower is in the adsorption stage, the other adsorption tower is in the desorption stage. This alternating operation mode ensures that the system can continuously produce high-purity oxygen. At the moment of tower switching, gas equalization is carried out between molecular adsorption tower A and molecular adsorption tower B to transfer part of the high-pressure gas in one tower to the other tower to improve energy efficiency and reduce energy consumption.

[0145] A radiator, the radiator is connected to the air compressor, and a solenoid valve is connected to the end of the radiator away from the air compressor. The solenoid valve is respectively connected to molecular adsorption tower A and molecular adsorption tower B;

[0146] An air storage tank, which is respectively connected to molecular adsorption tower A and molecular adsorption tower B, and one end of the air storage tank is also connected to an oxygen supply pipe.

[0147] In this embodiment, the dual-tower pressure swing adsorption technology can efficiently separate oxygen and nitrogen in the air and provide a relatively high-purity oxygen output. Through the alternating operation of molecular adsorption tower A and molecular adsorption tower B, continuous and uninterrupted oxygen supply is achieved, which is very suitable for the application scenarios of portable and miniaturized designs. Utilizing the selective adsorption characteristics of the adsorbent, no high temperature or chemical reactions are required, the operating cost is relatively low, and it is environmentally friendly. The system is easy to operate, highly automated, and has good stability, and is suitable for various environments that require stable oxygen supply. Especially when used in high-altitude environments, its stable oxygen output is particularly important. The portable oxygen generator main unit 400 of the dual-tower pressure swing adsorption (PSA) can be matched with nasal inhalation, face mask, and earphone types of user oxygen inhalation wearing methods.

[0148] To sum up, the present invention provides a portable long-term health monitoring and adaptive oxygen supply system in a high-altitude environment. In the embodiments of the present invention, through the coordinated cooperation of the blood oxygen monitoring module 100, the monitoring cloud platform 200, the adaptive oxygen generation control module 300, and the portable oxygen generator main unit 400, and based on the adaptive oxygen supply method based on the human blood oxygen level, by real-time monitoring of blood oxygen data, precise adjustment of the oxygen supply amount is achieved. Compared with the traditional oxygen supply method based on respiratory detection, it not only improves the oxygen supply accuracy, but also extends the service time and battery life of the system. At the same time, it integrates remote parameter monitoring and danger alarm functions, and uses the 4G network to collect and monitor environmental parameters, human blood oxygen data, GPS and other information in real time. When an abnormal situation is detected, the system will automatically send an alarm reminder and provide necessary distress information, significantly improving the safety of users in high-altitude environments.

[0149] In the embodiments of the present invention, a portable oxygen generator main unit 400 with a portable and miniaturized design is provided. The portable oxygen generator main unit 400 is adapted to the external environment, enabling it to be used in various external environments, facilitating carrying and operation. At the same time, a high-precision adaptive oxygen generation control method based on ambient air is proposed, which can automatically adjust the oxygen generation parameters according to different altitudes and ambient oxygen concentrations to ensure high-precision and stable oxygen output, with significant innovation and practicality.

[0150] It should be noted that for the foregoing embodiments, for the sake of simple description, they are all expressed as a series of action combinations. However, those skilled in the art should know that the present invention is not limited by the described action sequence, because according to the present invention, certain steps may be in other sequences or carried out simultaneously. Secondly, those skilled in the art should also know that the embodiments described in the specification are all preferred embodiments, and the actions and modules involved are not necessarily essential to the present invention.

[0151] Furthermore, the above embodiments can be used in scenarios such as plateau tourist attractions, plateau border defense posts, plateau scientific research stations, plateau medical institutions, etc., so as to provide continuous health monitoring and oxygen supply for the residents and tourists in the plateau area. The above embodiments are only used to illustrate the technical solutions of the present invention, rather than limiting the protection scope of the invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all embodiments. Based on these embodiments, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention. Although the present invention has been described in detail with reference to the above embodiments, those of ordinary skill in the art can still, without conflict and without creative efforts, combine, add, delete or make other adjustments to the features in the embodiments of the present invention according to the situation, so as to obtain different technical solutions that essentially do not deviate from the concept of the present invention, and these technical solutions also belong to the scope of protection of the present invention.

Claims

1. A portable long-term health monitoring and adaptive oxygen supply system in a plateau environment, characterized in that, The portable long-term health monitoring and adaptive oxygen supply system in the plateau environment includes: A blood oxygen monitoring module, which is used to monitor and obtain the user's blood oxygen monitoring data in real time and upload the blood oxygen monitoring data in real time; A monitoring cloud platform, which is communicatively connected to the blood oxygen monitoring module. The monitoring cloud platform is used to obtain blood oxygen monitoring data, oxygen generation-related data, and environmental monitoring data, analyze and store the blood oxygen monitoring data and oxygen generation-related data. The blood oxygen monitoring module establishes a data transmission channel with the monitoring cloud platform through a low-power Bluetooth module, and the sampling frequency ≥ 10Hz; An adaptive oxygen generation control module, which is used to monitor the environmental monitoring data and blood oxygen monitoring data in real time, dynamically adjust the pressurization time and oxygen generation-related data, realize adaptive oxygen supply, and generate a personalized oxygen output mode; A portable oxygen generation host, which is used to obtain a personalized oxygen output mode, and use the dual-tower pressure swing adsorption technology to separate oxygen from the air and provide high-purity oxygen output.

2. The portable long-term health monitoring and adaptive oxygen supply system in a plateau environment according to claim 1, wherein: The monitoring cloud platform includes: A data acquisition unit, which is communicatively connected to the adaptive oxygen generation control module, the blood oxygen monitoring module, and the portable oxygen generation host respectively. The data acquisition unit is used to obtain environmental monitoring data and blood oxygen monitoring data; A GPS module, which is used to locate the user's position in real time. The GPS module is communicatively connected to the blood oxygen monitoring module; A plateau warning unit, which analyzes and monitors the environmental monitoring data and blood oxygen monitoring data based on preset alarm rules, and performs adaptive warning of altitude sickness through the fitting curve of the relationship between blood oxygen value, heart rate value, altitude sickness and time, and triggers a plateau warning instruction.

3. The portable long-term health monitoring and adaptive oxygen supply system in a plateau environment according to claim 2, characterized in that: The adaptive oxygen generation control module is composed of an environmental parameter acquisition unit and a control adjustment unit; Among them: the environmental parameter acquisition unit collects environmental monitoring data in real time based on sensors. Among them, the environmental monitoring data includes environmental air pressure and environmental oxygen concentration; The control adjustment unit is used to load the environmental monitoring data, calculate the compressor pressurization time based on the environmental monitoring data, and adjust the pressurization time and oxygen generation-related data in combination with the compressor pressurization time and the PID control algorithm.

4. The portable long-term health monitoring and adaptive oxygen supply system in a plateau environment according to claim 3, characterized in that: The method of adjusting the pressurization time and oxygen generation-related data in combination with the compressor pressurization time and the PID control algorithm specifically includes: Load environmental monitoring data and identify the ambient air pressure P in the environmental monitoring data env and ambient oxygen concentration C env ; Set the target oxygen concentration C t , where the target oxygen concentration is pure oxygen or a preset concentration; According to the collected ambient air pressure and oxygen concentration, use formula (1) to calculate the required pressurization time T p ; Among them, in formula (1), k is an empirical constant, and P s is the standard atmospheric pressure; According to the calculated pressurization time T p , adjust the operating time of the compressor to ensure the expected oxygen concentration during the oxygen production process; Through the feedback mechanism, the system monitors the actual concentration C of the output oxygen in real time o , and compares it with the target oxygen concentration C t . If C o does not meet C t , the pressurization time and oxygen generation related data are further adjusted through the PID control algorithm. The feedback mechanism monitors the actual concentration C of the output oxygen by establishing a feedback network o , and feeds back the monitored actual concentration C of the output oxygen o to the PID controller. The PID controller calculates the error between the target oxygen concentration C t and the actual concentration C o , and updates the parameters of the feedback network according to the backpropagation rule to assist the PID control algorithm in adjusting the pressurization time and oxygen generation related data; Among them, formula (2) is used to adjust the pressurization time and the oxygen production related data, eliminate the static error of the pressurization time and the oxygen production related data, and improve the dead-band of the compressor. In formula (2), ΔT p represents the output signal of the PID controller, and e(t) = C t - C o is the error, and K p , K i , K d are the proportional, integral, and differential coefficients respectively.

5. The portable long-term health monitoring and adaptive oxygen supply system in a plateau environment according to claim 3, characterized in that: The adaptive oxygen generation control module further includes: An output mode generation module, which generates a personalized oxygen output mode based on the pressurization time and oxygen generation-related data.

6. The portable long-term health monitoring and adaptive oxygen supply system in a plateau environment according to claim 5, characterized in that: The output mode generation module includes: A constant oxygen supply unit, which generates a constant oxygen concentration supply mode based on a preset oxygen concentration value; An adaptive oxygen supply unit, which generates an adaptive oxygen supply mode based on a time neural network model combined with the actual blood oxygen level; Among them, when generating an adaptive oxygen supply mode based on a time neural network model combined with the actual blood oxygen level, the input to the time neural network model includes real-time blood oxygen value, historical blood oxygen value, and ambient oxygen concentration data, and the output of the time neural network model is the adjusted oxygen concentration parameter C t , according to the output oxygen concentration, the portable oxygen generator host dynamically adjusts the oxygen output concentration. The time neural network model takes the Transformer model as the main framework. The Transformer model includes two groups of encoders and decoders. The encoders and decoders respectively contain self-attention mechanisms and feed-forward neural networks. When the portable oxygen generator host dynamically adjusts the oxygen output concentration, it first presets the target oxygen concentration, and then based on the sensors, it real-time collects environmental monitoring data. According to the collected ambient air pressure and oxygen concentration, it calculates the required pressurization time T p , according to the calculated pressurization time T p , adjusts the running time of the compressor to ensure that the expected oxygen concentration is reached during the oxygen generation process. Through the feedback mechanism, the system real-time monitors the actual concentration C of the output oxygen o , and compares it with the target oxygen concentration C t . If C o does not meet C t , then the pressurization time and oxygen generation related data are further adjusted through the PID control algorithm.

7. The portable long-term health monitoring and adaptive oxygen supply system in a plateau environment according to claim 6, wherein: The training method of the time neural network model specifically includes: Taking the Transformer model as the initial model of the time neural network model, the Transformer model includes two groups of encoders and decoders. The encoders and decoders respectively include self-attention mechanisms and feed-forward neural networks. The self-attention mechanism calculates the attention weights of each position through the linear transformation of queries, keys, and values, so as to extract global context information; Optimizing and improving the Transformer model using sharpness-aware minimization of the SAMformer model; Obtain temporal training data and test data. Using the training data as input, iteratively train the Transformer model with the training data and output a converged temporal neural network model; Load the test data. Using the test data as input, execute the temporal neural network model, output the test results, and determine whether the test results are qualified based on a preset test accuracy. If the test results are qualified, output the trained temporal neural network model.

8. The portable long-term health monitoring and adaptive oxygen supply system in a plateau environment according to claim 7, wherein: The method of optimizing and improving the Transformer model using sharpness-aware minimization of the SAMformer model includes: Define a linear generation model based on the SAMformer model. The linear generation model is used to simulate temporal prediction. Among them, the output of the linear generation model is expressed as: Y = XW + ε (3) Where X represents the input matrix (sequence length L × sequence dimension D), W represents the model weight parameter, ε represents the loss, and Y is the model output; Introduce a simplified Transformer encoder into the Transformer model. The Transformer encoder consists of an attention mechanism, a residual connection layer, and a linear layer. The residual connection layer and the linear layer are one layer and three layers respectively. The Transformer encoder is expressed as: f(X) = [X + A(X)XW V W O W(4) where \(W\in\mathbb{R}\) L×H , d m is the model dimension, and \(A(X)\) is the attention matrix of the input sequence: Among them, Use the SAMformer model to modify the Transformer encoder based on formula (4), apply invertible instance normalization to the input X, and use the SAMformer model to optimize the Transformer model to make the Transformer model tend to a flatter local minimum, completing the optimization and improvement of the Transformer model.

9. The portable long-term health monitoring and adaptive oxygen supply system in a plateau environment according to claim 1, characterized in that: The portable oxygen generator host includes: An oxygen supply pipe for supplying oxygen to users; Medical molecular sieve installed in the main body housing, which is used to filter and purify the air; An air compressor fixedly installed in the main body housing, which is used to compress air and assist in oxygen generation; Molecular adsorption tower A; Molecular adsorption tower B. Molecular adsorption tower A and molecular adsorption tower B are respectively fixedly installed in the main body housing. Molecular adsorption tower A and molecular adsorption tower B are respectively filled with adsorbents. Molecular adsorption tower A and molecular adsorption tower B cooperate with each other and are used to alternately perform adsorption and desorption operations to achieve continuous oxygen production.

10. The portable long-term health monitoring and adaptive oxygen supply system in a plateau environment according to claim 9, characterized in that: The portable oxygen generator host also includes: A radiator, which is connected to the air compressor. One end of the radiator away from the air compressor is connected to a solenoid valve, and the solenoid valve is respectively connected to molecular adsorption tower A and molecular adsorption tower B; A gas storage tank, which is respectively connected to molecular adsorption tower A and molecular adsorption tower B, and one end of the gas storage tank is also connected to the oxygen supply pipe.

Images