An intelligent closed-loop oxygen therapy system based on the synergistic monitoring of brain oxygen and peripheral blood oxygen double targets

Abstract

The application discloses an intelligent closed-loop oxygen therapy system based on brain oxygen and peripheral blood oxygen double-target point cooperative monitoring, and belongs to the technical field of automatic control of medical devices. The system comprises a control system, a brain oxygen monitoring module, a peripheral blood oxygen monitoring module, a brain temperature monitoring module (optional), a flow control valve, a flow metering module, a man-machine interface, a communication module and a power management module. The control system is internally provided with a self-adaptive hierarchical control algorithm based on multi-parameter fusion, takes local brain tissue oxygen saturation rSO2 as a direct control target, takes peripheral blood oxygen saturation SpO2 as a safety reference boundary, and takes frontal lobe brain temperature as an auxiliary decision variable. The actual flow is fed back through the flow metering module, the flow control valve is controlled through PID closed-loop adjustment and step-by-step adjustment, and the oxygen output flow is dynamically adjusted. The algorithm is provided with decision levels such as true and false event identification, hierarchical intervention, intelligent oxygen withdrawal and safety monitoring, can identify true local ischemia, false interference and systemic hypoxia events, and realizes precise brain protection.

Description

Technical Field

[0001] This invention relates to a medical device used in clinical oxygen therapy, specifically a closed-loop control system that uses local brain tissue oxygen saturation (rSO2) as the core control target, integrates peripheral blood oxygen saturation (SpO2) as a safety reference boundary, and simultaneously monitors frontal lobe brain temperature to make multi-parameter intelligent decisions and dynamically adjust oxygen output flow rate. It belongs to the field of automatic control of medical devices. Background Technology

[0002] Oxygen therapy is one of the most commonly used emergency or treatment methods in hospitals. Its purpose is to improve and stabilize a patient's dynamic blood oxygen saturation within a reasonable range by administering oxygen, correcting abnormal oxygenation caused by various reasons, and maintaining vital bodily functions. Currently, oxygen therapy in China is largely unregulated. In most cases, there are no strict medical instructions, and the entire oxygen therapy process lacks rigorous monitoring and necessary flow rate adjustments.

[0003] In recent years, closed-loop oxygen supply systems targeting pulse oxygen saturation (SpO2) have addressed the problem of automatic regulation of systemic hypoxia to some extent. However, existing technologies still have the following significant shortcomings: First, systemic oxygenation indicators cannot accurately reflect the oxygenation status of local organs, especially the brain. In critical care settings such as shock, post-cardiopulmonary bypass surgery, and traumatic brain injury, even if a patient's peripheral blood oxygen saturation (SpO2) appears normal (e.g., 98%-100%), the brain may still be in a state of latent hypoxia due to local microcirculatory disturbances, vasospasm, or insufficient perfusion pressure. This phenomenon of "brain oxygenation disconnecting from systemic oxygenation" is a significant cause of postoperative cognitive impairment (POCD), hypoxic-ischemic encephalopathy, and even brain death. Closed-loop systems that solely target SpO2 cannot address this clinical challenge.

[0004] Second, cerebral oxygen monitoring and oxygen therapy control are independent of each other. Although there are cerebral oxygen saturation monitors based on near-infrared spectroscopy (NIRS) technology in clinical practice, their functions are limited to data monitoring and alarms, and they do not form a control loop with oxygen supply equipment. When cerebral oxygen values ​​deviate, medical staff still need to manually adjust the ventilator or flow meter, which has the defects of adjustment lag, insufficient accuracy, and difficulty in maintaining cerebral oxygen homeostasis.

[0005] Third, single-signal-source control has inherent risks. Whether relying solely on SpO2 or solely on brain oxygen (rSO2) for closed-loop control, there are limitations. A single signal is susceptible to non-physiological factors such as probe displacement, ambient light interference, and patient movement, potentially leading to misjudgments and malfunctions in the control system.

[0006] Fourth, the crucial metabolic parameter of brain temperature is overlooked. Brain temperature is an important indicator reflecting the metabolic level of brain tissue and local blood perfusion. In pathological states such as ischemia, hypoxia, or systemic infection, changes in brain temperature often precede changes in blood oxygen levels. Existing oxygen therapy systems do not incorporate brain temperature monitoring into their control logic, thus missing valuable opportunities for early warning and intervention.

[0007] To address the aforementioned issues, this invention proposes an intelligent closed-loop oxygen therapy system based on the synergistic monitoring of cerebral oxygen and peripheral blood oxygen, aiming to achieve a technological leap from "whole-body oxygen supply" to "precise brain protection". Summary of the Invention

[0008] The purpose of this invention is to provide a safer and more accurate intelligent closed-loop oxygen therapy system based on the synergistic monitoring of cerebral oxygen and peripheral blood oxygen dual targets, in order to solve the problems mentioned in the background art. This invention uses cerebral oxygen (rSO2) as the core control target, peripheral blood oxygen (SpO2) as the safety reference boundary, and introduces frontal lobe brain temperature monitoring as the basis for multi-parameter fusion decision-making, to achieve precise, safe, and intelligent oxygen output.

[0009] To achieve the above objectives, the present invention provides the following technical solution: An intelligent closed-loop oxygen therapy system based on the synergistic monitoring of cerebral oxygen and peripheral blood oxygen dual targets mainly consists of a control system, a cerebral oxygen monitoring module, a peripheral blood oxygen monitoring module, a flow control valve, a flow metering module, a human-machine interface, a communication module, a power management module, and a control algorithm; additionally, a brain temperature monitoring module is optional. Its core feature lies in the fact that the control system incorporates an adaptive hierarchical control algorithm based on multi-parameter fusion. This algorithm uses local brain tissue oxygen saturation (rSO2) as the direct control target, peripheral blood oxygen saturation (SpO2) as the safety reference boundary, and frontal lobe brain temperature as an auxiliary decision variable.

[0010] (I) Composition, structural relationship and function of each module 1. Control system (control module) The control system is the central unit of this invention, based on a control circuit developed using a programmable logic controller (PLC) or a microcontroller (MCU). In a preferred embodiment, an STM32F4 series microcontroller (such as STM32F407 or STM32F429) is used as the main control chip. This chip has the following technical features: based on an ARM Cortex-M4 core, with a maximum operating frequency of 168MHz; a built-in floating-point unit (FPU) to meet the computational requirements of multi-parameter real-time signal processing and complex control algorithm operation; equipped with multiple 12-bit analog-to-digital converters (ADCs) with a maximum sampling rate of 2.4Msps, capable of high-speed synchronous acquisition of multi-channel physiological signals; and built-in multiple universal synchronous / asynchronous transceivers (USARTs), serial peripheral interfaces (SPI), and I / O pins. 2 The C interface is used for data communication with various functional modules.

[0011] The control system connects to the cerebral oxygenation monitoring module, peripheral blood oxygenation monitoring module, brain temperature monitoring module, flow control valve, flow metering module, human-machine interface, and communication module via integrated circuits or wires, respectively, to work collaboratively. Specifically, the cerebral oxygenation monitoring module, peripheral blood oxygenation monitoring module, and brain temperature monitoring module communicate via SPI or I... 2 The C-bus connects to the microcontroller of the control system, transmitting the collected physiological parameter digital signals to the control system; the flow control valve receives flow regulation commands from the control system via pulse width modulation (PWM) signals or 4-20mA analog signals; the flow metering module feeds back real-time flow data to the control system via the SPI interface, forming a closed-loop flow control; the human-machine interface connects to the control system via a parallel bus or serial port; and the communication module connects to the control system via the USART interface to realize remote transmission of monitoring data.

[0012] 2. Brain oxygen monitoring module The brain oxygen monitoring module is the core sensing component of this invention, used to continuously and non-invasively acquire local brain tissue oxygen saturation (rSO2). This module mainly consists of a near-infrared spectroscopy (NIRS) sensor probe, a light source driving circuit, a photoelectric detection circuit, a signal amplification and filtering circuit, an analog-to-digital conversion circuit, and a brain oxygen calculation unit.

[0013] (1) NIRS sensor probe: Employs dual-wavelength near-infrared spectroscopy sensing technology. The sensor probe is attached to the patient's frontal lobe and integrates the following core components: —Near-infrared light-emitting diode (LED) light source: Two near-infrared LEDs with different wavelengths are used, namely 760nm and 850nm. The 760nm wavelength near-infrared light is sensitive to deoxyhemoglobin (deoxy-Hb), while the 850nm wavelength near-infrared light is sensitive to oxyhemoglobin (oxy-Hb). Single-wavelength output power >15mW.

[0014] —Photodetector: Employs a PIN photodiode or avalanche photodiode (APD) to receive near-infrared light signals scattered and absorbed by brain tissue. APD detectors have the advantages of high sensitivity, low dark current, and low noise.

[0015] —Spatial layout of the light source and detector: A reflective detection method is adopted, and the distance between the light source and the photodetector is generally 3-5cm. This distance ensures that the near-infrared light can penetrate the skull and reach the superficial layer of brain tissue (about 1-3cm deep), thereby obtaining local brain tissue oxygenation information.

[0016] (2) Light source driving circuit: A constant current source driving circuit is used to provide a stable and adjustable driving current for the near-infrared LED. The driving circuit adopts a time-division multiplexing method, so that the LEDs with wavelengths of 760nm and 850nm emit light alternately at a certain time interval (usually on the order of milliseconds). The magnitude of the driving current is adjustable to adapt to patients with different skin colors and different tissue thicknesses, ensuring that the signal strength at the receiving end is within the optimal dynamic range.

[0017] (3) Photoelectric detection and signal processing circuit: The photodetector receives the near-infrared light signal scattered by brain tissue and converts it into a weak current signal. This current signal undergoes the following processing in sequence: —Transimpedance amplifier circuit (IV conversion): The integrated transimpedance amplifier (TIA) is used to convert the weak current signal of the photodetector into a voltage signal.

[0018] —Filtering circuit: Includes a low-pass filter and a band-pass filter. The low-pass filter is used to extract the DC component (reflecting tissue baseline absorption), and the band-pass filter is used to extract the AC component (reflecting pulse-related absorption changes).

[0019] —Analog-to-digital converter (ADC): Converts analog voltage signals into digital signals, which are then fed into a microcontroller for processing.

[0020] (4) Brain Oxygen Calculation Unit: The brain oxygen calculation unit is integrated into the microcontroller of the control system. It calculates the local brain tissue oxygen saturation based on the modified Beer-Lambert Law. The specific algorithm is as follows: Based on the light intensity attenuation values ​​measured at two wavelengths, 760nm and 850nm, the concentration changes of deoxyhemoglobin and oxyhemoglobin are calculated respectively, and then the brain oxygen saturation is calculated according to the formula: rSO2=[oxy-Hb] / ([oxy-Hb]+[deoxy-Hb])×100% When the system is first started, it automatically measures and records the patient's individualized cerebral oxygen baseline (rSO2-Base), which is the arithmetic mean of the rSO2 values ​​monitored continuously for 3-5 minutes in the patient's calm state, and uses this as the core reference for subsequent control.

[0021] 3. Peripheral blood oxygen monitoring module The peripheral oxygenation monitoring module is used to dynamically monitor a patient's arterial oxygen saturation (SpO2), pulse rate (PR), and perfusion index (PI). This module mainly consists of a peripheral oxygenation probe, an oxygenation simulation front-end circuit, and an oxygenation calculation unit.

[0022] (1) Peripheral blood oxygen probe: usually a finger clip or finger sleeve type transmission blood oxygen probe is used.

[0023] The peripheral blood oxygen probe contains: —Light-emitting device: Two light-emitting diodes (LEDs) with emission wavelengths of 660nm (red light) and 940nm (infrared light) respectively, which emit light alternately at certain time intervals.

[0024] —Photosensitive receiver: A PIN photodiode is used, placed on the opposite side of the finger, to receive red and infrared light signals transmitted through the finger tissue.

[0025] (2) Blood Oxygen Simulation Front-End Circuit: In the preferred embodiment, the high-performance integrated blood oxygen simulation front-end chip AFE4490 manufactured by TI is used. The AFE4490 chip integrates the following functional modules: —LED transmission channel: Includes programmable LED drive current control, which can drive 660nm red LED and 940nm infrared LED to emit light alternately.

[0026] — Receiver channel: Includes transimpedance amplifier (IV converter), ambient light cancellation circuit, programmable gain amplifier, low-pass filter and 22-bit high-precision analog-to-digital converter (ADC).

[0027] —Diagnostic functions: Includes sensor LED fault detection, probe detachment detection, and other functions.

[0028] The AFE4490 chip communicates with the microcontroller of the control system via the SPI interface, transmitting the acquired and preprocessed photoplethysmography (PPG) digital signal to the microcontroller.

[0029] (3) Blood oxygen calculation unit: Integrated into the microcontroller, it calculates the SpO2 value based on the ratio of AC to DC components at the two wavelengths of red and infrared light, combined with the empirical calibration curve. At the same time, it extracts the pulse rate (PR) and perfusion index (PI) from the PPG signal.

[0030] The data from the peripheral blood oxygen monitoring module does not directly participate in the fine adjustment of oxygen flow under normal conditions, but serves as a core reference for safety boundary verification and event type identification.

[0031] More preferably, the peripheral blood oxygen probe and the NIRS sensor probe of the brain oxygen monitoring module are integrated into a single monitoring unit, which is then applied to the patient's frontal lobe. This improves the integration of components, reduces cable entanglement caused by monitoring different locations, and enhances user-friendliness.

[0032] 4. Brain temperature monitoring module (optional) The brain temperature monitoring module is used to continuously monitor the surface temperature of the patient's frontal lobe, reflecting the changing trend of brain tissue temperature. This module consists of a high-precision thermistor sensor (NTC), signal conditioning circuitry, and a temperature calculation unit.

[0033] (1) NTC Thermistor Sensor: A high-precision medical-grade NTC (negative temperature coefficient) thermistor is used. This thermistor is integrated into the NIRS brain oxygenation sensor patch and attached to the patient's frontal lobe. Key technical parameters include: —Measurement accuracy: ±0.1℃ within the human body temperature range (35℃-42℃); —Thermal time constant: ≤1 second, ensuring rapid response to temperature changes; —Long-term stability: Annual drift is less than 0.05℃; —Operating temperature range: 0℃ to 70℃.

[0034] (2) Signal conditioning circuit: A voltage divider circuit is constructed using precision resistors and NTC thermistors to convert the resistance change caused by temperature change into a voltage change signal. After the voltage signal is filtered and amplified, it is converted from analog to digital by the ADC module of the microcontroller.

[0035] (3) Temperature calculation unit: integrated into the microcontroller, the temperature value is calculated based on the resistance-temperature characteristic curve of the NTC thermistor (Steinhart-Hart equation).

[0036] 5. Flow control valve A flow control valve is used to precisely adjust the oxygen output flow rate according to instructions from the control system. This invention employs an electronic flow proportional valve.

[0037] In the preferred embodiment, Parker's VSO is used. ® LowPro miniature proportional valves or similar products. The technical specifications of this proportional valve are as follows: —Control method: Electromagnetic drive, input voltage selectable: 5VDC, 12VDC or 24VDC; —Response time: Typical value 10ms; —Flow range: Maximum flow rate up to 57 SLPM (standard liters per minute); —Work stress: up to 100 PSIG (approximately 6.9 bar); —Valve type: 2-Way, Normally Closed; —Lifespan: ≥100 million cycles; —Power consumption: Typical value 1.0W, maximum value 2.0W; —Sealing material: FKM (fluororubber), suitable for oxygen media; —Leakage rate: <0.5 sccm (standard cubic centimeters / minute).

[0038] The input signal of the proportional valve is proportional to the output flow rate. The control system controls the opening of the proportional valve through a PWM (Pulse Width Modulation) signal or a 0-10V analog voltage signal. When the control system issues operating commands such as valve opening, flow adjustment, or valve closing, the proportional valve linearly adjusts its opening according to the magnitude of the command signal, achieving stepless and precise regulation of the oxygen flow rate.

[0039] 6. Flow metering module The flow metering module is used to monitor and measure the actual output oxygen flow rate in real time, forming a closed-loop flow control. This module mainly consists of a thermal mass flow sensor and a signal processing circuit.

[0040] (1) Thermal mass flow sensor: Based on the thermal principle of MEMS (Micro-Electro-Mechanical Systems). The sensor integrates the following core components: —Heating resistor: Located in the center of the sensor chip, it generates heat when powered on; —Upstream and downstream thermopile: symmetrically distributed on both sides of the heating resistor, used to detect temperature difference.

[0041] Its working principle is as follows: When gas flows through the sensor, the flowing gas carries away the heat from the heating resistor, causing the temperature distribution on both sides of the heating resistor to become asymmetrical. The temperature difference measured by the upstream and downstream thermopile is proportional to the mass flow rate of the gas. Since heat transfer is only related to the gas mass and heat capacity, the sensor can directly measure the mass flow rate of the gas without the need for temperature and pressure compensation.

[0042] In a preferred embodiment, the flow sensor has a range of 0-15 L / min (or 0-10 L / min) and an accuracy of ±2% of the reading. The sensor communicates via SPI or I... 2 The C interface communicates with the microcontroller of the control system, transmitting real-time flow digital signals to the control system.

[0043] (2) Signal processing circuit: integrated inside the sensor or on the main board of the control system, it amplifies, filters and converts the weak signal output by the sensor to digital, and finally outputs the digital flow value.

[0044] The flow metering module is installed in the oxygen output pipeline downstream of the flow control valve. It monitors the actual oxygen flow rate output by the proportional valve in real time and feeds the data back to the control system. The control system compares the actual flow rate with the target flow rate and uses a PID (proportional-integral-derivative) control algorithm to perform closed-loop regulation of the flow control valve to ensure the accuracy of the output flow rate.

[0045] 7. Human-computer interaction interface The human-computer interface is used for parameter setting, operation control, and information display in this invention. It consists of a combination of an LCD screen and function keys.

[0046] (1) LCD screen: In the preferred embodiment, a 5-inch or 7-inch color TFT LCD touch screen with a resolution of not less than 800×480 is used. The screen is used to display the following information in real time: current brain oxygen saturation (rSO2) value and trend graph, current peripheral blood oxygen saturation (SpO2) value, current frontal lobe brain temperature value, current oxygen output flow rate value, target brain oxygen value and safety control window range, system working status, warning information, etc.

[0047] (2) Function keys: should include at least the following setting keys: —Power switch: Controls the power supply of the system to turn on and off; —Target brain oxygenation setting key: Used to set the target brain oxygenation value (rSO2 target value), which is automatically determined by the system baseline value by default; —Medical order flow / concentration setting key: Used to set the initial oxygen flow rate (L / min); —Oxygen therapy duration setting button: Used to set the duration (h) of this oxygen therapy session.

[0048] Each setting key can be made of an electronic encoder, allowing for quick adjustment and setting of relevant parameters by rotating the encoder left and right; alternatively, it can be input directly via a touchscreen.

[0049] 8. Communication Module The communication module is used to remotely send monitoring data (rSO2, SpO2, brain temperature, blood flow, pulse rate, etc.), analyzed and synthesized monitoring information, and warning information to the medical monitoring terminal.

[0050] The communication module supports the following technologies: —Wired transmission: Connect to the hospital's central monitoring system via RS-232, RS-485, or Ethernet interface; —Wireless transmission: Supports Bluetooth (Bluetooth Low Energy, BLE), WiFi (such as QCA9377 chip) or RF radio frequency technologies.

[0051] The communication module connects to the microcontroller of the control system via a USART or SPI interface.

[0052] 9. Power Management Module The power management module provides a stable and reliable power supply for the entire system. It uses a medical-grade power adapter (input AC 100-240V, output DC 5V / 3A or 12V / 2A). The power management module internally includes: —Voltage conversion circuit: Converts the input voltage to different voltage levels required by each module (such as 3.3V, 5V, 12V, etc.). —Overvoltage protection circuit: Prevents abnormal voltage rise from damaging components; —Overcurrent protection circuit: prevents abnormal increase in current; —Electromagnetic compatibility (EMC) filter circuit: suppresses electromagnetic interference on power lines to meet the EMC requirements of medical devices.

[0053] (II) Control Algorithm (Core of Intelligent Decision Making) The most innovative part of this invention is the control algorithm built into the control system. Its components include, but are not limited to: target brain oxygenation value (rSO2 target value), brain oxygenation safety control window, peripheral blood oxygenation reference threshold, frontal lobe brain temperature reference range, doctor's order flow rate, intervention control time, flow rate correction interval, flow rate correction gradient, and oxygen inhalation duration. The control algorithm adopts a rule-based hierarchical decision architecture, and integrates PID closed-loop regulation and finite state machine (FSM) control strategies.

[0054] The decision-making process of the control algorithm is as follows: First layer: Signal acquisition and preprocessing The control system acquires the following signals in real time at a sampling frequency of not less than 1 Hz: —Brain oxygenation monitoring module: absolute value and trend of rSO2; —Peripheral blood oxygenation monitoring module: SpO2, pulse rate, PI; —Brain temperature monitoring module: absolute value and trend of frontal lobe brain temperature; —Flow metering module: Real-time oxygen output flow rate.

[0055] All signals are processed by digital filtering (such as moving average filtering or median filtering) to remove high-frequency noise and occasional interference before entering the decision logic layer.

[0056] Second layer: Individualized goal definition (1) Core target value: i.e., the patient's individualized cerebral oxygen baseline value (rSO2-Base). After the system is started, it is continuously monitored for 3-5 minutes in a calm state, and the arithmetic mean of rSO2 is automatically calculated as the baseline value.

[0057] (2) Brain oxygen safety control window: Based on the baseline value, an "absolute safety window" is defined. The preferred scheme is: the safety control window is defined as ±5% of the baseline value (for example, if the baseline value is 65%, then the safety control window is 60%-70%). The safety control window is written into the control algorithm as the sole direct control target for oxygen flow regulation.

[0058] (3) Peripheral blood oxygen safety boundary: SpO2 below 88% is defined as the trigger threshold for "systemic hypoxia event".

[0059] (4) Brain temperature reference range: Based on the patient's basal body temperature, an auxiliary warning is triggered when the frontal lobe brain temperature increases by more than 1°C or decreases by more than 1°C from the baseline value.

[0060] The third layer: "True / False" event identification (state machine level 1) When the system detects a significant deviation in the rSO2 value (e.g., deviating from the safety control window), the control system will not immediately adjust the oxygen flow rate, but will first enter the event identification procedure. Event identification is based on the following three parallel decision paths: —Pathway A (True Local Ischemia Event): rSO2 is below the lower limit of the safety control window, and SpO2 ≥ 90% (no systemic hypoxia), and there is no drastic change in brain temperature (change from baseline < 1℃).

[0061] —Diagnosis: Insufficient local cerebral perfusion or imbalance between cerebral oxygen supply and demand.

[0062] —Action: Enter the “Stratified Intervention” procedure (Level 4) and start oxygen flow regulation.

[0063] —Path B (artifacts or interference events): rSO2 is below the lower limit of the safety control window, but the SpO2 signal is stable and ≥90%, while the brain temperature data is normal, and the perfusion index (PI) fluctuates drastically (indicating poor probe contact).

[0064] — Judgment: The NIRS probe is being interfered with (e.g., displacement, changes in ambient light, patient agitation).

[0065] —Action: Suspend all oxygen regulation commands, issue a prompt "Please check brain oxygen probe" on the human-machine interface, and restore the oxygen flow rate to the flow rate value of the previous stable cycle (average flow rate value 3 minutes ago).

[0066] —Pathway C (systemic hypoxia event): rSO2 is below the lower limit of the safety control window, and SpO2 is simultaneously below 88%.

[0067] —Diagnosis: Severe systemic hypoxemia.

[0068] —Action: Activate "Safety Coverage Mode", stop performing fine regulation of brain oxygenation, directly switch to emergency high oxygen supply (supply oxygen according to the doctor's order flow limit), and issue the highest priority audible and visual alarm, while remotely transmitting it to the nursing terminal through the communication module.

[0069] Fourth layer: Layered intervention and closed-loop regulation (state machine level 2) For procedures identified as "true ischemic events," the following stratified adjustments are initiated only: (1) Alert zone (rSO2 decreases by 5%-10% from baseline, but is still above the lower limit of the safety control window): Do not adjust the oxygen flow rate, increase the monitoring frequency to once every 10 seconds, and display "slight decrease in brain oxygen, observe closely" on the interface.

[0070] (2) Intervention zone (rSO2 is below the lower limit of the safety control window, i.e., a decrease of >10% from the baseline): The system starts the "trial oxygenation" program.

[0071] The adjustment method employs a strategy combining PID (proportional-integral-derivative) control algorithm with stepped adjustment: — PID control: Using the target value (baseline value) of rSO2 as the setpoint and the current measured value of rSO2 as the process variable, the deviation value e(t) is calculated as: target value of rSO2 - measured value of rSO2. The output u(t) of the PID controller is calculated according to the following formula: u(t)=Kp×e(t)+Ki×∫e(t)dt+Kd×de(t) / dt Wherein Kp, Ki, and Kd are the proportional coefficient, integral coefficient, and differential coefficient, respectively, which are tuned through clinical pre-trials. The preferred parameter ranges are: Kp = 0.5-2.0, Ki = 0.01-0.1, Kd = 0.1-0.5.

[0072] —Step-by-step regulation: To prevent excessive PID output from causing large fluctuations in flow rate, the theoretical regulation value calculated by the PID is converted into a step-by-step regulation. Each regulation step value is the flow correction gradient (0.25-0.5 L / min). After each regulation, wait for an intervention control time (0.5-1 minute when increasing the flow rate, and 1-3 minutes when decreasing the flow rate) and observe the response of rSO2.

[0073] —Adjustment direction determination: —If rSO2 is below the lower limit of the safety control window, increase the oxygen output flow rate; —If rSO2 is higher than the upper limit of the safety control window, reduce the oxygen output flow rate.

[0074] —Adjustment Range Limitations: Oxygen flow rate adjustments are strictly limited to the flow correction range. The flow correction range is automatically matched based on the prescribed flow rate value. —Low flow orders (flow rate 0.5-2 L / min) → Flow rate correction range 0.1-2 L / min; —Medium flow prescription (prescribed flow rate 3-4 L / min) → Flow correction range 0.1-4 L / min; —High-flow medical orders (flow rate 5-8L / min) → Flow rate correction range 0.1-8L / min.

[0075] (3) Refractory hypoxia zone (rSO2 decreases by more than 15% from baseline or absolute value is less than 50%): The system continues to increase the oxygen flow rate to the preset upper limit within the flow rate correction range. If the flow rate has reached the upper limit and rSO2 does not improve within 3 minutes (recovery is less than 2%), the system will no longer blindly increase the oxygen flow rate, but will issue a clear etiology prompt of "brain oxygen continues to not rise, it is recommended to investigate blood pressure / cardiac output / intracranial pressure", and maintain the current highest safe flow rate value, waiting for clinical intervention.

[0076] In addition, the system has a clear recovery and exit mechanism for the "artifacts or interference events (path B)" and "systemic hypoxia events (path C)" identified in the "third layer" judgment: —Path B (artifact event) exit mechanism: When the system detects that the signal quality has recovered (such as stable PI value and rSO2 signal strength meeting the standard) and continues for more than 10 seconds, it will automatically release the "suspended" state and re-enter the normal oxygen therapy monitoring and adjustment mode.

[0077] —Pathway C (Systemic Hypoxia Event) Exit Mechanism: In "Safe Coverage Mode," the system continuously monitors SpO2. When SpO2 rises and stabilizes at ≥90% for more than 30 seconds, the system automatically exits "Safe Coverage Mode" and switches back to the fine-tuning mode targeting rSO2.

[0078] Fifth layer: Intelligent oxygen withdrawal and steady-state maintenance (state machine level 3) Once rSO2 recovers and stabilizes within the safety control window (remaining stable for more than 3 minutes), the system initiates the "slow active oxygen withdrawal" procedure: —Gradually reduce the oxygen flow rate at a rate of 0.25-0.5 L / min per minute (i.e., one flow correction gradient per minute); —Observe the response of rSO2 after each reduction. If rSO2 remains stable within the safe control window, continue to reduce it. —If rSO2 falls below the lower limit of the safety control window again during the adjustment process, stop oxygen removal immediately and restore to the previous stable flow rate value; —Until it drops to the initial prescribed flow rate or the minimum effective flow rate (the minimum flow rate that stabilizes rSO2 within the safety control window).

[0079] This process ensures that "oxygen is replenished immediately upon deficiency and removed immediately upon recovery," minimizing the risk of oxygen poisoning.

[0080] Sixth layer: Brain temperature-assisted decision-making During the intervention, the system simultaneously monitored changes in frontal lobe brain temperature: —If a decrease in rSO2 is accompanied by an abnormal increase in brain temperature (more than 1°C above baseline), the system will add the auxiliary judgment "Pay attention to infection or increased brain metabolism" to the prompt message. —If a decrease in rSO2 is accompanied by an abnormal decrease in brain temperature (a decrease of >1°C from baseline), the system will add the auxiliary judgment "severe inadequacy of perfusion" to the prompt message; —Brain temperature data does not directly participate in the automatic regulation of oxygen flow; it is only used as auxiliary diagnostic information for clinicians' reference.

[0081] Seventh layer: Security monitoring and fault handling (1) When the blood oxygen monitoring module malfunctions (including but not limited to the blood oxygen probe falling off the monitoring site, the blood oxygen probe being damaged, or the blood oxygen value not being able to be obtained normally), the control system will automatically restore the output flow rate to the initially set medical order flow rate, stop the automatic intervention adjustment of the output flow rate, avoid safety hazards, and issue a prompt "Please check the blood oxygen probe" on the human-machine interface.

[0082] (2) When the signal quality of the brain oxygen monitoring module decreases (e.g., the signal strength is below the threshold), the control system suspends all adjustment commands, maintains the current flow value, and issues a prompt that “brain oxygen signal quality is poor, please check the probe”.

[0083] (3) In case of system failure (such as microcontroller crash, proportional valve drive abnormality, etc.), the hardware watchdog circuit automatically resets the system, and at the same time the flow control valve returns to the normally closed state, cutting off oxygen output to ensure patient safety.

[0084] Eighth layer: Respiratory rate learning function (optional expansion) As an optional extension, the control algorithm can learn the patient's respiratory rate. During dynamic monitoring of the patient's respiratory rate, the algorithm records the patient's inhalation, exhalation, and the duration of these actions. The control algorithm performs deep learning based on these changes in respiratory rate, establishing a respiratory rate pattern that automatically synchronizes the oxygen output of the flow control valve with the patient's breathing. Simultaneously, the respiratory rate sensor continues to dynamically monitor the patient's respiratory rate. If the patient's respiratory rate pattern changes, the flow control valve switches back to a real-time response mode based on respiratory rate and begins a new learning process.

[0085] (III) Optimization range of parameters for each element of the control algorithm

[0086] (iv) Working method of the present invention The specific parameters of each component of the control algorithm are pre-written into the control system. When applying oxygen therapy according to this invention: 1. Medical staff attach the NIRS brain oxygen sensor and brain temperature probe to the patient's forehead, clip the finger pulse oxygen probe to the patient's fingertip, and connect the oxygen input and output lines.

[0087] 2. Turn on the power. The system will start automatically, automatically measure and record the patient's individualized cerebral oxygen baseline (rSO2-Base).

[0088] 3. Based on the patient's condition, medical staff set the target brain oxygen value (the baseline value automatically measured by the system is used by default), the prescribed flow rate, and the duration of oxygen inhalation on the human-computer interaction interface.

[0089] 4. After the settings are complete, the system will enter oxygen therapy mode.

[0090] 5. The brain oxygen monitoring module, peripheral blood oxygen monitoring module, and brain temperature monitoring module simultaneously perform multi-parameter dynamic monitoring.

[0091] 6. The control algorithm of the control system runs automatically according to the above eight-level decision-making process.

[0092] 7. Once the set oxygen therapy duration is reached, the flow control valve will automatically close, and the oxygen therapy session will end.

[0093] (v) Beneficial effects of the present invention 1. Precise brain protection: It directly controls brain oxygen saturation, solving the problem of "hidden brain hypoxia" under normal systemic oxygenation.

[0094] 2. Dual-target synergy enhances safety: By introducing peripheral blood oxygen as a reference boundary, it can effectively distinguish between "systemic hypoxia" and "local cerebral hypoxia" and prevent misregulation caused by interference from a single cerebral oxygen signal source.

[0095] 3. Multi-parameter fusion enhances diagnostic value: The innovative introduction of frontal lobe brain temperature monitoring provides richer dimensions for the system to judge the pathophysiological state of cerebral ischemia and hypoxia.

[0096] 4. Intelligent closed-loop management: From identifying “real / false” events to tiered intervention and then to proactive oxygen withdrawal, the entire process is automated and intelligently managed.

[0097] 5. Precise PID control: By integrating the PID closed-loop control algorithm with the step-by-step adjustment strategy, a smooth, precise, and stable output of oxygen flow is achieved, avoiding adverse stimulation to patients caused by sudden changes in flow. Attached Figure Description

[0098] Figure 1 This is a diagram illustrating the working principle and module connection of the system of the present invention.

[0099] Figure 2 This is a flowchart of the decision tree for the intelligent decision control algorithm of this invention. Detailed Implementation

[0100] The present invention will be further described in detail below with reference to the accompanying drawings and embodiments. Those skilled in the art should understand that the following embodiments are merely illustrative of the technical solutions of the present invention and not intended to limit it.

[0101] Example 1: Preparation of the present invention The main components used in the invention are as follows: —Control system: The microcontroller uses ST's STM32F407ZGT6 chip (ARM Cortex-M4 core, 168MHz clock speed, built-in FPU), equipped with an external 8MHz crystal oscillator and a 32.768kHz real-time clock crystal oscillator.

[0102] —Brain oxygen monitoring module: The NIRS sensor probe uses dual-wavelength near-infrared LEDs (wavelengths of 760nm and 850nm, single-wavelength power >15mW), and the photodetector uses an avalanche photodiode (APD); the light source driving circuit uses a constant current source driving chip; and the signal conditioning circuit uses an integrated transimpedance amplifier.

[0103] —Peripheral blood oxygen monitoring module: The finger clip blood oxygen probe uses dual-wavelength LEDs (660nm and 940nm); the blood oxygen simulation front end uses the TI AFE4490 chip.

[0104] —Brain temperature monitoring module: The NTC thermistor uses a high-precision medical-grade NTC chip (accuracy ±0.1℃, thermal time constant ≤1 second).

[0105] —Flow control valve: Parker VSO® LowPro miniature proportional valve (10ms response time, 57SLPM maximum flow rate).

[0106] —Flow metering module: adopts MEMS thermal mass flow sensor (range 0-15L / min, accuracy ±2%).

[0107] —Human-computer interaction interface: It adopts a 5-inch color TFT touch LCD screen (resolution 800×480) and three electronic encoder function keys.

[0108] —Communication module: It adopts a WiFi module (model QCA9377) and a Bluetooth BLE 4.0 module.

[0109] —Power Management Module: Uses a medical-grade AC-DC power adapter (input 100-240V, output DC 5V / 3A), with integrated overvoltage, overcurrent, and EMC filtering protection circuits.

[0110] After each component has passed inspection, according to Figure 1 The working principle diagram shown is as follows: PCB (printed circuit board) is prepared using general electronic product manufacturing processes. Surface mount technology (SMT) is used for component placement and soldering. Various components such as flow control valve and flow metering module are assembled, and a housing and control buttons are added. After the whole machine is assembled, an aging test (72 hours of continuous power-on operation) is performed.

[0111] Following the embedded software development process, embedded software was written in C language within the Keil MDK development environment. The software includes the following functional modules: hardware driver layer (ADC driver, SPI driver, PWM driver, USART driver, etc.), signal processing layer (digital filtering, feature extraction), control algorithm layer (event identification, state machine management, PID controller), human-machine interaction layer (display driver, touchscreen processing), and communication protocol layer. The software was burned into the microcontroller of the complete product and underwent full performance testing, passing all tests.

[0112] Example 2: Target Brain Oxygen Value Setting Method When applying oxygen therapy using this invention, the target brain oxygenation value can be set for different patients by operating the function keys on the human-computer interaction interface according to the patient's condition characteristics: 1. Patients recovering from anesthesia: The target brain oxygenation level is set at 96% (baseline value), and the brain oxygenation safety control window is automatically defined as 91%-100%; 2. For newborns receiving routine oxygen therapy: the target brain oxygen saturation is set at 93%, and the safe control window for brain oxygenation is automatically defined as 88%-98%. 3. Patients with acute respiratory distress syndrome: The target brain oxygenation value is set at 92%, and the brain oxygenation safety control window is automatically defined as 87%-97%; 4. For individuals at risk of hypercapnia: the target brain oxygenation level is set at 90%, and the safe control window for brain oxygenation is automatically defined as 85%-95%; 5. For patients receiving routine oxygen therapy: the target brain oxygen value is set to 95%, and the brain oxygen safety control window is automatically defined as 90%-100%.

[0113] Example 3: Clinical application in post-cardiac surgery patients Scenario: A patient is transferred to the ICU after coronary artery bypass grafting. Medical staff attach the NIRS brain oxygen sensor and brain temperature probe to the patient's forehead and clip the finger pulse oxygen probe to the patient's fingertip. The prescribed flow rate is set to 3 L / min on the human-machine interface. After system startup, the system automatically measures brain oxygen levels at 65% (i.e., rSO2-Base = 65%), and the brain oxygen safety control window automatically locks at 60%-70% (baseline ± 5%).

[0114] Event Triggering and Intelligent Identification: Two hours post-surgery, due to transient low cardiac output, the system monitored rSO2 gradually decreasing from 65% to 55% (a decrease of approximately 15%, falling below the lower limit of the safety control window of 60%). The system simultaneously read peripheral blood SpO2, showing 97% (normal, ≥90%), with no significant change in brain temperature. Based on "Pathway A," the system determined this to be a true focal ischemic event.

[0115] Quasi-stratified intervention: The system enters the "intervention zone" and initiates the "tentative oxygenation" program. The oxygen flow rate is increased every minute (intervention control time = 1 min) in increments of 0.5 L / min (flow rate correction gradient = 0.5 L / min). When the flow rate increases from 3 L / min to 4.5 L / min, rSO2 begins to rise, stabilizing at 63% after 3 minutes (returning to the safe control window of 60%-70%). The system maintains the current flow rate of 4.5 L / min and displays the message "Brain oxygenation improved, current flow rate 4.5 L / min" on the interface.

[0116] Intelligent oxygen withdrawal and steady-state maintenance: After another 10 minutes, the patient's cardiac function recovered, and rSO2 stabilized at 66% (stable for >3 minutes within the safety control window). The system initiated the "slow active oxygen withdrawal" program, reducing the oxygen flow rate smoothly to the initial commanded level of 3L / min at a rate of 0.5L / min per minute (flow correction gradient = 0.5L / min). Throughout the process, rSO2 remained stable within the safety control window of 60%-70%, and SpO2 remained ≥96%.

[0117] Results and significance: Under the background of normal systemic oxygenation, this system accurately captured and automatically corrected a transient cerebral hypoxia event, avoiding the risk of medical staff neglecting cerebral oxygenation due to reliance on SpO2. It also eliminated the possibility of probe interference and achieved precise individualized brain protection of "supplementing oxygenation when hypoxia occurs and withdrawing it when recovery occurs".

[0118] Example 4: Multi-parameter fusion to eliminate interference Scenario: A COPD patient who has been receiving long-term oxygen therapy experiences a sudden drop in NIRS brain oxygenation from 60% to 48% during oxygen therapy, while the human-machine interface displays abnormal fluctuations in brain oxygen signal intensity data.

[0119] System Response: The system simultaneously checked peripheral blood SpO2, which showed 93% (stable, ≥90%), with no abnormal brain temperature. However, the perfusion index (PI) fluctuated drastically. Based on "Path B," the system determined this to be an artifact event and did not increase the oxygen flow rate. A message popped up stating, "Please check the brain oxygen probe; signal quality is poor." The oxygen flow rate automatically returned to the value of the previous stable cycle (the average flow rate value from 3 minutes prior).

[0120] Result: After the nurse re-secured the probe, the rSO2 reading recovered to 60%. The system avoided an erroneous high-flow oxygen supply due to a loose probe, ensuring patient safety.

[0121] Example 5: Safety Coverage for Systemic Hypoxia Scenario setting: A patient with respiratory failure experiences a sharp drop in SpO2 during oxygen therapy due to airway obstruction caused by sputum plugs.

[0122] System Response: The system detected that rSO2 dropped from 62% to 50% (below the lower limit of the safety control window), and simultaneously SpO2 dropped from 95% to 85% (<88% trigger threshold). Based on "Path C", the system determined it to be a systemic hypoxia event and immediately activated "Safety Coverage Mode": it stopped performing fine-tuning of cerebral oxygen and directly switched to emergency hyperoxia supply (oxygen supply at the doctor's prescribed flow rate of 8L / min); it issued the highest priority audible and visual alarm; and at the same time, it remotely transmitted the warning information to the nursing terminal via the communication module.

[0123] Result: After receiving the alarm, medical staff rushed to the bedside to clear the airway. The patient's SpO2 returned to normal. After confirming that SpO2 had risen to ≥90%, the system automatically exited the "safe coverage mode" and resumed fine regulation of brain oxygen.

[0124] The above figures and embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the present invention, and all such modifications and substitutions should be covered within the scope of the claims of the present invention.

Claims

1. An intelligent closed-loop oxygen therapy system based on the synergistic monitoring of cerebral oxygen and peripheral blood oxygen, characterized in that, Includes a control system, a brain oxygen monitoring module, a peripheral blood oxygen monitoring module, a flow control valve, and a flow metering module; The brain oxygen monitoring module is used to continuously and non-invasively acquire the local brain tissue oxygen saturation rSO2. The peripheral oxygenation monitoring module is used to dynamically monitor arterial oxygen saturation SpO2, pulse rate (PR), and perfusion index PI. The flow metering module is installed in the oxygen output pipeline downstream of the flow control valve, and is used to monitor the actual output oxygen flow rate in real time and feed it back to the control system. The control system is connected to the brain oxygen monitoring module, peripheral blood oxygen monitoring module, flow control valve and flow metering module respectively, and has an built-in adaptive hierarchical control algorithm; The control system is configured as follows: Obtain rSO2, SpO2, PI and actual oxygen output flow rate; rSO2 was continuously monitored for 3-5 minutes in a patient's calm state. The arithmetic mean was calculated as the individualized rSO2 baseline value, and the brain oxygen safety control window was determined by ±5% of the baseline value. When rSO2 deviates from the brain oxygen safety control window, event identification is first performed based on rSO2, SpO2, and PI to distinguish between true local ischemic events, artifacts or interference events, and systemic hypoxia events; among which, PI is used to assist in the identification of artifacts or interference events. Closed-loop flow regulation with rSO2 as the control target is initiated only for true local ischemia events. The closed-loop flow regulation adopts a combination of PID control and step-by-step regulation, and limits the oxygen flow rate within the flow correction range. Suspend oxygen flow adjustment for false alarms or interference events, and restore oxygen flow to the average flow value 3 minutes before the previous stable cycle; In the event of systemic hypoxia, activate the safety coverage mode, supply oxygen at the doctor's prescribed flow rate limit and issue an alarm; Once rSO2 recovers and stabilizes within the brain oxygen safety control window for more than 3 minutes, a slow active oxygen withdrawal procedure is executed.

2. The intelligent closed-loop oxygen therapy system based on dual-target synergistic monitoring of cerebral oxygenation and peripheral blood oxygenation as described in claim 1, characterized in that, It also includes a brain temperature monitoring module, which is used to continuously monitor the surface temperature of the patient's frontal lobe and transmit the collected brain temperature signal to the control system, which uses the frontal lobe brain temperature as an auxiliary decision variable.

3. The intelligent closed-loop oxygen therapy system based on dual-target synergistic monitoring of cerebral oxygenation and peripheral blood oxygenation as described in claim 2, characterized in that, The event identification includes: if rSO2 is below the lower limit of the safety control window, SpO2 is ≥90%, and the frontal lobe brain temperature changes by <1℃ from the baseline value, it is determined to be a true local ischemic event; if rSO2 is below the lower limit of the safety control window, SpO2 is stable and ≥90%, the frontal lobe brain temperature is normal, but PI fluctuates drastically, it is determined to be an artifact or interference event; if rSO2 is below the lower limit of the safety control window and SpO2 is simultaneously below 88%, it is determined to be a systemic hypoxia event.

4. The intelligent closed-loop oxygen therapy system based on dual-target synergistic monitoring of cerebral oxygenation and peripheral blood oxygenation as described in claim 2, characterized in that, The brain temperature monitoring module includes an NTC thermistor, a signal conditioning circuit, and a temperature calculation unit. The NTC thermistor is integrated into a near-infrared spectral brain oxygen sensor patch and is attached to the patient's frontal lobe. The temperature calculation unit calculates the temperature value based on the resistance-temperature characteristic curve of the NTC thermistor.

5. The intelligent closed-loop oxygen therapy system based on dual-target synergistic monitoring of cerebral oxygenation and peripheral blood oxygenation as described in claim 1, characterized in that, The flow metering module includes a thermal mass flow sensor based on the MEMS thermal principle and a signal processing circuit. The thermal mass flow sensor integrates a heating resistor located in the center of the chip, as well as an upstream thermopile and a downstream thermopile symmetrically distributed on both sides of the heating resistor. The thermal mass flow sensor has a range of 0-15L / min or 0-10L / min and an accuracy of ±2% of the reading. The sensor communicates with the control system via an SPI or I²C interface. The control system compares the actual flow rate with the target flow rate and performs closed-loop regulation of the flow control valve using a PID control algorithm.

6. The intelligent closed-loop oxygen therapy system based on dual-target synergistic monitoring of cerebral oxygenation and peripheral blood oxygenation as described in claim 2, characterized in that, The adaptive hierarchical control algorithm includes the following decision levels: The first layer, signal acquisition and preprocessing: The control system acquires rSO2, SpO2, pulse rate, PI, frontal lobe brain temperature and real-time oxygen output flow rate in real time at a sampling frequency of not less than 1Hz, and processes them through digital filtering. The second layer is the definition of individualized targets: after the system is started, it continuously monitors the patient in a calm state for 3-5 minutes, automatically calculates the arithmetic mean of rSO2 as the baseline value, and defines the cerebral oxygen safety control window with ±5% of the baseline value; the systemic hypoxia event trigger threshold is defined with SpO2 < 88%. Auxiliary warning is triggered when the frontal lobe brain temperature increases by more than 1°C or decreases by more than 1°C from the baseline value. The third layer is the identification of true / false events: when the rSO2 value shifts significantly, the event identification procedure is first entered, and true local ischemia events, artifacts or interference events, and systemic hypoxia events are determined based on three parallel decision paths: path A, path B, and path C. The fourth layer, stratified intervention and closed-loop regulation: stratified regulation is initiated only for true local ischemia events, including the alert zone, intervention zone and refractory hypoxia zone; The fifth layer, intelligent oxygen withdrawal and steady-state maintenance: When rSO2 recovers and stabilizes within the safety control window for more than 5 minutes, the slow active oxygen withdrawal program is initiated; The sixth layer, brain temperature-assisted decision-making: synchronously monitors changes in frontal lobe brain temperature and outputs auxiliary diagnostic information; The seventh layer is for security monitoring and fault handling.

7. The intelligent closed-loop oxygen therapy system based on dual-target synergistic monitoring of cerebral oxygen and peripheral blood oxygen as described in claim 6, characterized in that, In the third layer: Pathway A is a true local ischemic event: rSO2 is below the lower limit of the safety control window, SpO2 is ≥90%, and brain temperature changes from baseline by <1℃. It is judged as local cerebral insufficiency or cerebral oxygen supply and demand imbalance, and enters the stratified intervention procedure. Path B is an artifact or interference event: rSO2 is below the lower limit of the safety control window, but the SpO2 signal is stable and ≥90%, the brain temperature data is normal, and the perfusion index PI fluctuates drastically. It is determined that the NIRS probe is interfered with, all oxygen regulation commands are suspended, a prompt is issued on the human-machine interface, and the oxygen flow rate is restored to the average flow rate value 3 minutes before the previous stable cycle. Path C is a systemic hypoxemia event: rSO2 is below the lower limit of the safety control window, and SpO2 is simultaneously below 88%, which is judged as severe systemic hypoxemia. The safety coverage mode is activated, directly switching to oxygen supply according to the doctor's order flow rate limit, and issuing the highest priority audible and visual alarm. At the same time, it is remotely transmitted to the nursing terminal through the communication module.

8. The intelligent closed-loop oxygen therapy system based on dual-target synergistic monitoring of cerebral oxygen and peripheral blood oxygen as described in claim 6, characterized in that, In the fourth layer: The warning zone is defined as rSO2 being below the lower limit of the safety control window and decreasing by no more than 10% from the baseline. The intervention zone was defined as rSO2 below the lower limit of the safety control window and a decrease of >10% from baseline. The PID control uses the target value of rSO2 as the setpoint and the current measured value of rSO2 as the process variable. The deviation value e(t) is calculated as rSO2 target value - rSO2 measured value. The PID controller outputs u(t) = Kp×e(t) + Ki×∫e(t)dt + Kd×de(t) / dt, where Kp = 0.5-2.0, Ki = 0.01-0.1, and Kd = 0.1-0.

5. Each adjustment step of the stepped adjustment is a flow correction gradient of 0.25-0.5 L / min. After each adjustment, the waiting time for intervention control is 0.5-1 minute when increasing the flow rate and 1-3 minutes when decreasing the flow rate. The refractory hypoxia zone is defined as a decrease in rSO2 of >15% from baseline or an absolute value of <50%.

9. The intelligent closed-loop oxygen therapy system based on dual-target synergistic monitoring of cerebral oxygenation and peripheral blood oxygenation as described in claim 8, characterized in that, The flow correction range is automatically matched according to the doctor's order flow rate: low flow orders of 0.5-2 L / min correspond to a flow correction range of 0.1-2 L / min; medium flow orders of 3-4 L / min correspond to a flow correction range of 0.1-4 L / min; and high flow orders of 5-8 L / min correspond to a flow correction range of 0.1-8 L / min. Furthermore, when rSO2 is below the lower limit of the brain oxygen safety control window, the oxygen output flow rate is increased, and when rSO2 is above the upper limit of the brain oxygen safety control window, the oxygen output flow rate is decreased.

10. The intelligent closed-loop oxygen therapy system based on dual-target synergistic monitoring of cerebral oxygen and peripheral blood oxygen as described in claim 6, characterized in that, The slow, active oxygen removal procedure in the fifth layer is as follows: Gradually reduce the oxygen flow rate at a rate of 0.25-0.5 L / min. Observe the rSO2 reaction after each reduction. If rSO2 remains stably within the safe control window, continue to reduce the level. If rSO2 falls below the lower limit of the safety control window again during the adjustment process, oxygen withdrawal should be stopped immediately and the flow rate restored to the previous stable value. Until it drops to the initial prescribed flow rate or the lowest effective flow rate that stabilizes rSO2 within the brain oxygen safety control window.

11. The intelligent closed-loop oxygen therapy system based on dual-target synergistic monitoring of cerebral oxygenation and peripheral blood oxygenation as described in claim 7, characterized in that, The exit mechanism for path B is as follows: when the system detects that the signal quality has recovered and lasted for more than 10 seconds, it automatically releases the suspended state and re-enters the normal oxygen therapy monitoring and adjustment mode. The exit mechanism for path C is as follows: In the safe coverage mode, when SpO2 rises and stabilizes at ≥90% for more than 30 seconds, the system automatically exits the safe coverage mode and switches back to the fine control mode targeting rSO2.

12. The intelligent closed-loop oxygen therapy system based on dual-target synergistic monitoring of cerebral oxygenation and peripheral blood oxygenation as described in claim 2, characterized in that, The control system is also configured to perform brain temperature-assisted decision-making: if a decrease in rSO2 is accompanied by a frontal lobe brain temperature increase of >1°C from the baseline value, an auxiliary judgment of infection or increased brain metabolism is added to the prompt information; if a decrease in rSO2 is accompanied by a decrease in frontal lobe brain temperature of >1°C from the baseline value, an auxiliary judgment of severe inadequate perfusion is added to the prompt information; the frontal lobe brain temperature does not directly participate in the automatic regulation of oxygen flow.

13. The intelligent closed-loop oxygen therapy system based on dual-target synergistic monitoring of cerebral oxygen and peripheral blood oxygen as described in claim 6, characterized in that, The control system is also configured to perform safety monitoring and fault handling: when the peripheral blood oxygen monitoring module malfunctions, the output flow rate is restored to the initially set doctor's order flow rate value, the automatic intervention adjustment is stopped, and a prompt is made to check the blood oxygen probe. When the signal strength collected by the brain oxygen monitoring module is lower than the preset threshold, all adjustment commands are suspended, the current flow value is maintained, and a message indicating poor brain oxygen signal quality is displayed. When the system malfunctions, it is automatically reset through the hardware watchdog circuit and the flow control valve is restored to the normally closed state to cut off oxygen output.

14. A smart closed-loop oxygen therapy control method based on dual-target synergistic monitoring of cerebral oxygen and peripheral blood oxygen, characterized in that, Includes the following steps: S1. Continuously collect local brain tissue oxygen saturation rSO2, peripheral blood oxygen saturation SpO2, perfusion index PI, and actual oxygen output flow rate; S2. Monitor rSO2 continuously for 3-5 minutes while the patient is at rest, calculate the arithmetic mean as the individualized rSO2 baseline value, and define the brain oxygen safety control window with ±5% of the baseline value. S3. When rSO2 deviates from the brain oxygen safety control window, first perform event identification based on rSO2, SpO2 and PI to distinguish between true local ischemia events, artifacts or interference events and systemic hypoxia events. S4. Closed-loop flow regulation targeting rSO2 is initiated only for true local ischemia events. The closed-loop flow regulation adopts a combination of PID control and step-by-step regulation, and limits the oxygen flow rate within the flow correction range. S5. For false alarms or interference events, suspend oxygen flow adjustment and prompt for checking the brain oxygen probe; for systemic hypoxia events, activate the safety coverage mode, supply oxygen according to the doctor's prescribed flow limit and issue an alarm. S6. Once rSO2 recovers and stabilizes within the brain oxygen safety control window, execute the slow active oxygen withdrawal procedure.

15. The method according to claim 14, characterized in that, It also includes collecting frontal lobe brain temperature and using it as an auxiliary decision variable, wherein the frontal lobe brain temperature does not directly participate in the automatic regulation of oxygen flow.

Images