Oxygen monitoring and control system

Abstract

Systems and methods are disclosed for regulating supplemental oxygen delivery to a user using physiological, environmental, and device‑state data. An oxygen delivery system includes an electronically controllable oxygen‑flow valve, one or more sensors, and a controller configured to determine a current operating context based on data received from the sensor(s), retrieve a context‑trim value corresponding to the current operating context from memory, compute an oxygen‑flow command based on the data received from the sensor(s) and the context-trim value, and actuate the valve based on the oxygen-flow command to maintain a target blood oxygen saturation level.

Description

OXYGEN MONITORING AND CONTROL SYSTEM PRIORITY

[0001] This patent application claims the benefit of U. S. Provisional Patent Application No. 63 / 759,821, entitled " OXYGEN MONITORING AND CONTROL SYSTEM”, filed on February 18, 2025, which is incorporated herein by reference in its entirety.TECHNICAL FIELD

[0002] The disclosure relates to devices, systems, and methods for measuring and controlling blood oxygen in a subject.BACKGROUND

[0003] Long-term oxygen therapy (LTOT) can be used by a substantial number of subjects worldwide for various medical indications, and supplemental oxygen can be provided through several established delivery modalities. These include home-based oxygen concentrators, pre-filled compressed-gas cylinders used with oxygen-conserving regulators, and liquid-oxygen systems. Each system has characteristics associated with its design, such as weight, portability, mode of operation, and the manner in which oxygen can be supplied to a subject. However, existing supplemental oxygen systems may vary in their responsiveness to physiological changes and environmental conditions.SUMMARY

[0004] The present disclosure relates to devices, systems, and methods for regulating supplemental oxygen delivery to a subject using physiological, environmental, and device-state information. In various aspects, an oxygen delivery’ system comprises an oxygen source, an electronically controllable oxygen-flow valve, one or more physiological and environmental sensors, a memory storing context-dependent adjustment information, and a controller configured to determine and actuate an oxygen-flow command to maintain the subject’s blood oxygen saturation within a desired range. The controller may determine a current operating context based on sensor data and select a corresponding adjustment value from a plurality of stored context-trim entries. The controller may further generate an oxygen-flow command based on both the received sensor data and the selected adjustment value, and may actuate the electronically controllable oxygen-flow valve in accordance with the oxygen-flow command.1121390.09891014899-1153-2431.2

[0005] In certain aspects, the oxygen delivery system may operate in a plurality of modes associated with different combinations of physiological and environmental inputs, such as a constant flow mode, titration mode, titration and breath mode, andtitration-and-activity mode. In some examples, the controller may execute a proportionalintegral-derivative algorithm, wherein proportional, integral, and derivative terms are computed based on the deviation between a measured blood oxygen saturation level and a target blood oxygen saturation level, as well as the rate of change of the measured level. Additional adaptive adjustments may be based on motion, respiratory effort, heart-rate variability, and other relevant factors.

[0006] In some examples, the oxygen delivery system includes a dual-lumen cannula comprising a flow lumen for oxygen delivery and a sense lumen coupled to a pressure sensor. The controller may detect inhalations and exhalations based on a pressure signal obtained via the sense lumen and may adjust the oxygen-flow command in coordination with the subject’s breathing cycle. In certain aspects, the controller may adjust a rate at which the oxygen-flow valve opens based on measured inspiratory’ effort.

[0007] In other examples, the system comprises a flow sensor and a tank-pressure sensor. The controller may’ determine a size of a compressed oxygen tank and a remaining quantity of oxygen based on correlating measured flow with a rate of change of tank pressure over time, and may provide a display of the remaining quantity of oxygen.

[0008] In various aspects, the oxygen delivery system may include a user- adjustable unified control knob for selecting a continuous-flow setting or an automatic electronic mode. A watchdog safety circuit may monitor the controller and actuate a mechanical bypass path based on detection of a failure, with the unified control knob providing a corresponding indication of the bypass state.

[0009] In certain aspects, a machine-learning framework may be provided for device management, dashboard display, and patient monitoring and management. The framework may include one or more machine-learning models trained using physiological time-series data, device telemetry, environmental signals, patient activity data, and labeled clinical outcomes. The framework may be hosted on a remote server or integrated within a device associated with the oxygen delivery system. The machine-learning model may generate device-management actions, dashboard indicators, and patient-status assessments based on processed training data2121390.09891014899-1153-2431.2BRIEF DESCRIPTION OF THE DRAWINGS

[0010] The following detailed description can be better understood when read in conjunction with the appended drawings. In the drawing, for purposes of explanation, numerous specific details are set forth in order to provide an understanding of the variations in implementing the disclosed technology. However, the instant disclosure may take many different forms and should not be construed as limited to the specific examples disclosed in the drawings. In the drawings:

[0011] FIG. 1A depicts a PID (proportional, integral, derivative) control loop, according to some variations of the disclosure;

[0012] FIG. IB depicts a block diagram of a closed loop oxygen delivery system, according to some variations of the disclosure;

[0013] FIG. 2 depicts a block diagram depicting details of an internal configuration of an oxygen flow control unit (OFCU 1) and its relationship with external components of the closed loop oxygen delivery system, according to some variations of the disclosure;

[0014] FIG. 3 depicts the data display and control interfaces of an example OFCU 1, according to some variations of the disclosure;

[0015] FIG. 4A depicts the internal layout of the key components of the OFCU 1, according to some variations of the disclosure;

[0016] FIG. 4B depicts the internal layout of the key components of an OFCU 1, according to some variations of the disclosure;

[0017] FIG. 4C depicts an external configuration of a multi-component OFCU 1, home / stationary concentrator mounted module, according to some variations of the disclosure;

[0018] Figure 5 illustrates a flow diagram of the Device Operating Modes;

[0019] Figure 6A depicts graphs for oxygen saturation of arterial blood (SpO2) reported from a pulse oximeter and supply of oxygen flow from an oxygen flow control unit (OFCU 1), according to some variations of the disclosure;

[0020] Figure 6B describes operational guidelines depicted in Figure 6A;

[0021] Figure 7A depicts graphs for oxygen saturation of arterial blood (SpO2), pulse rate (PR), respiratory rate (RR), and oxygen flow rate (FLOW) from running the algorithm logic via a simulation engine that mimics lung activity for one hour of typical usage;

[0022] Figure 7B depicts a graph for the SpO2 range and the time spent in each SpO2range for the data in Figure 7A.3121390.098910\4899-1153-2431.2

[0023] Figure 8A depicts graphs for oxygen saturation of arterial blood (SpO2), pulse rate (PR), respiratory’ rate (RR), and oxygen flow rate (FLOW) from running the algorithm logic via a simulation engine that mimics lung activity for one hour of typical usage, with a desaturation event;

[0024] Figure 8B depicts a graph for the SpO2 range and the time spent in each SpO2range for the data in Figure 8A;

[0025] Figure 9A depicts graphs for oxygen saturation of arterial blood (SpO2), pulse rate (PR), respiratory rate (RR), and oxygen flow rate (FLOW) from running the algorithm logic via a simulation engine that mimics four weeks of lung activity for Short-Term Oxygen Therapy (STOT);

[0026] Figure 9B depicts a graph for the SpO2 range and the time spent in each SpO2range for the data in Figure 9A;

[0027] Figure 9C depicts a graph for the number of desaturation events per day (with and without activity ) for the data in Figure 9A;

[0028] Figure 10A depicts graphs for oxygen saturation of arterial blood (SpO2), pulse rate (PR), respiratory’ rate (RR), and oxygen flow rate (FLOW) from running the algorithm logic via a simulation engine that mimics twelve weeks of lung actiy ity for a patient with Chronic Obstructive Pulmonary Disease (COPD);

[0029] Figure 10B depicts a graph for the SpO2 range and the time spent in each SpO2range for the data in Figure 10A;

[0030] Figure 1 OC depicts a graph for the number of desaturation events per day (with and without activity) for the data in Figure 10A;

[0031] Figure 11 is a schematic of the System Architecture that incorporate Artificial Intelligence (Al) and machine learning (ML), according to some variations of the disclosure;

[0032] Figure 12 is a schematic showing tailored response for oxygen delivery based on integration of feedback response over time and algorithmic logic, according to some variations of the disclosure;

[0033] FIG. 13 is an example flowchart illustrating an example process for providing oxygen to a subject, according to some variations of the disclosure.

[0034] FIG. 14 illustrates an example computer system, according to some variations of the disclosure.

[0035] FIG. 15 illustrates an example of a machine learning framework, according to some variations of the disclosure.4121390.098910\4899-1153-2431.2DETAILED DESCRIPTION

[0036] In various aspects, an OFCU configured to provide a closed-loop oxygen supply system is described. The OFCU can comprise the following components: an oxygen source and a proportional valve. The proportional valve can be coupled to a microprocessor that controls a communication system, which receives data from blood oxygen sensors. In various aspects, oxygen can be delivered to the user. The OFCU described herein may alternatively be referred to as an “Automatic Supplemental Oxygen Control (ASOC)” unit.

[0037] This system can regulate the flow of oxygen to the subject through the proportional valve that calibrates the flow of oxygen to the subject. The OFCU provides components for inputting subject information such as the subject's desired target blood oxygen level. The target blood oxygen level and the actual blood oxygen level of the subject, as detected by the pulse / oximetry sensors, are used by the microprocessor to calculate the appropriate proportionally controlled valve settings required to deliver to the subject the necessary amount of oxygen to maintain the target oxygen level. When the sensors detect that the subject's blood oxygen levels are lower than the target oxygen level, the microprocessor signals the controller to increase flow through the controlled valve. When the blood oxygen levels are higher than the target oxygen level the microprocessor instructs the controller to decrease flow through the controlled valve.

[0038] In variations as depicted in FIG. 1 A, the processor can be operating a PID (proportional, integral, derivative) control loop. An OFCU (such as the OFCU 1 described below in more detail with regard to FIG. IB) measures the oxygen levels and the processor calculates the difference between desired target oxygen levels and actual oxygen levels (Proportional). The processor also can calculate the velocity of the difference in levels (mathematical first Derivative). The processor also can calculate the long-term difference in target and actual values and integrates this small difference over time (Integral). If the subject changes from sedentary (sitting or lying down) to walking quickly, the blood oxygen deficit can change quickly. The first derivative portion of the control loop calculates the deficit will become very large very quickly and opens (or closes) the valve more aggressively than the Proportional portion of the algorithm would alone. If the subject is sedentary for a long time and the difference between target oxygen level and actual is slightly off, the mathematical integral of this difference will grow over time and cause further adjustments to the Oxygen flow rate to keep the error even lower. If oxygen is already present in in the tube, the change in flow and flow rate can be instantaneous.5121390.098910\4899-1153-2431.2

[0039] The OFCU may also calculate, record, and store data related to oxygen consumption, the subject's blood oxygen level over time, respiration rate and other respiratory indicia. The subject or a healthcare professional can monitor oxygen use based on this data. In some aspects, as depicted in FIG. 2, the OFCU 1 can be designed to be coupled to an oxygen tube between a low-pressure oxygen source and an oxygen delivery system. In this aspect the OFCU 1 receives oxygen flowing from the oxygen source through the oxygen tube via an oxygen inlet port of device 4 attached to the proportional valve of the OFCU 1. The oxygen flow rate can be regulated by the proportional controlled valve and exits via an oxygen outlet port outflow 16. The oxygen outlet port can be attached to an oxygen tube leading to an oxygen delivery system such as a transtracheal tube, a nasal cannula worn by the subject or a high- altitude training mask. The OFCU 1 can be designed to be highly portable, transported in the subject's pocket or handbag or alternatively, connected to the oxygen regulator attached to the O2 oxygen tank, or small and light enough to attach to a subject's belt or clothing when in use. Breath-synchronous ramp shaping improves perceived comfort, reducing the sensation of abrupt "‘bolus hits" by tailoring onset and cutoff of oxygen flow to the subject's measured inspiratory effort.

[0040] Various aspects of an OFCU 1 are configured as a closed loop oxygen delivery system as depicted in FIG. IB. In the figure, oxygen flows through the system from left to right along the path indicated by the heavy line, originating with source 2, which includes an oxygen source comprising an oxygen reservoir or generator and a pressure regulator. The oxygen flows from the oxygen source to the OFCU 1. Oxygen then flows from the OFCU 1 to the patient via an oxygen delivery device, for example a cannula or transtracheal device, as depicted in device 4. A physiological monitoring device such as a pulse / oximetry sensor, shown in sensor 3, monitors the level of oxygen present in the patient and relays this information to the OFCU 1 via a wired or wireless connection to close the loop connecting oxygen delivery to oxygen demand as indicated by the heavy solid line. Additional wired or wireless inputs and outputs may transfer data to or from the OFCU 1 from additional external devices (Smartphones and / or Smartwatches and Bluetooth-enabled hearing aids as shown in device 5 by the dashed line.

[0041] Additional information related to remaining oxygen in portable tanks can be provided to the OFCU 1 using a Bluetooth connection to the Digital Tank Pressure Sensor 6.This capability provides the patient with oxygen reserve time in minutes versus the current pounds per square inch (psi) that can be of little use to patients. In some embodiments, the controller estimates tank size and remaining oxygen by correlating an integrated delivered 6121390.098910\4899-1153-2431.2volume (e.g., jflowdt measured by a calibrated flow sensor) with a pressure drop AP over time measured by a tank-side pressure sensor. Because pressure can be proportional to gas quantity for compressed oxygen under typical operating conditions, this method applies at any fill level. In certain embodiments, this estimation is not applicable to liquid oxygen reservoirs whose tank pressure remains relatively constant; liquid oxygen may require alternative methods (e.g., gravimetric).

[0042] There can be an additional wireless (Wi-Fi) connection between the OFCU 1 and the patient's Home / stationary Concentrator 7. This provides control over the flow rate for supplemental oxygen from the Home / stationary Concentrator, a device that provides the majority of the patient's supplemental oxygen when they are home.

[0043] In some examples, the regulated oxygen delivery system of FIG. IB employs a dual-lumen cannula comprising a flow lumen configured to deliver oxygen to the subject and a sense lumen fluidly isolated from the flow lumen and coupled to a pressure sensor. The pressure signal measured via the sense lumen provides a high-fidelity respiratory waveform from which the controller can detect inspiration onset (pressure decrease) and inspiration termination (return toward baseline or reversal in slope), enabling synchronized oxygen delivery that begins during inspiration (e g., inhalation) and that is reduced prior to exhalation. In certain embodiments, the controller further shapes a comfort-optimized valve ramp-up and ramp-down profile in proportion to inspiratory effort inferred from the pressure signal (e.g., larger effort —> faster ramp; smaller effort —> slower ramp). In some embodiments, pressure-sensing latency can be on the order of milliseconds, whereas pulse oximetry signals may be lagging; thus, breath-synchronous gating and anticipatory adjustments can be performed using pressure, motion, and other leading indicators.

[0044] In some embodiments of the oxygen delivery system of FIG. IB, a single user control (e.g., unified control knob 29) can be adjustable by a user, allowing the user to select among a plurality of continuous-flow settings and an automatic electronic mode. The unified control knob 29 may provide a single-location indication of the activeoxygen-delivery state. For example, the unified control knob 29 may include an interface that can be configured to display information indicating the active (e.g., current) oxy gen-delivery state of the OFCU 1. If the OFCU 1 can be operating according to a first continuous flow setting (e.g., the user has used the unified control knob 29 to select the first continuous flow setting), the interface may display information indicating that the OFCU 1 can be operating according to the first continuous flow setting. Likewise, if the OFCU 1 can be operating according to the automatic electronic mode (e.g., the user has used the unified control knob 7121390.098910\4899-1153-2431.229 to select the automatic electronic mode), the interface may display information indicating that the OFCU 1 can be operating according to the automatic electronic mode.

[0045] The OFCU 1 can include a watchdog safety circuit 30 (e.g., as illustrated in FIG. IB). The watchdog safety circuit 30 can be an electronic monitoring subsystem that provides a fail-safe mechanism to ensure uninterrupted oxygen delivery in the event of an electronic malfunction. The watchdog safety circuit 30 can operate independently of the main controller in the OFCU 1 (e.g., the microcontroller control unit 9 described below in more detail). The watchdog safety circuit 30 can be configured to continually expect periodic communication "heartbeats" from the controller. If the controller fails to communicate with the watchdog safety circuit 30 within a defined interval (e.g., due to a processor crash, firmware fault, power anomaly, or other system error) the watchdog safety circuit 30 can interpret this as a fault condition. In response to determining the fault condition, the watchdog safety circuit 30 can automatically force the oxygen delivery system into a mechanical bypass mode. In this bypass mode, power to a normally-open bypass valve can be removed, allowing the valve to return to its default open state and thereby route oxygen to the subject through a non-electronic continuous-flow pathway. The bypass valve may be mechanically biased to open when power can be lost, ensuring that oxygen delivery continues even if all electronics fail. In this manner, the watchdog safety circuit 30 can ensure that safety-critical oxygen flow can be maintained regardless of the operational state of the controller. If the mechanical bypass path can be active, the unified control knob 29 can display an indication that the mechanical bypass path can be active.Supplemental Oxygen Sources

[0046] In one aspect the oxygen source comprises a small, compressed oxygen bottle sufficiently small and light to be readily portable. This bottle may be attached by a quick release pressure regulator such that the regulator can be easily and quickly removed from the tank and placed on a new tank once depleted. In some aspects the oxygen source may comprise a group of bottles with one or more pressure regulators arranged in such manner as to provide a low profile or to conform to the subject's back in a manner in which the subject can be spared injury should they fall or otherwise strike the oxygen source against a surface in the course of their activities. In other aspects the oxygen source may comprise a larger non-portable compressed oxygen source and regulator. In some aspects the oxygen source may comprise a liquid oxygen reservoir, which may be portable or non-portable and may further comprise a pressure regulator. In some aspects the oxygen source may be a large8121390.098910\4899-1153-2431.2institutional liquid oxygen source such as may be found in a hospital, neighborhood based Urgent Care facility or an extended care facility such as a hospice center.

[0047] In other aspects the oxygen source may comprise a portable oxygen concentrator which may be battery powered. In other aspects the oxygen source may comprise a larger non-portable oxygen concentrator. Other oxygen production or enrichment systems familiar to those of skill in the art may be incorporated into the disclosure on the basis that all such systems deliver oxygen to the subject through a standard tubing system into which the OFCU 1 and other components of the disclosed closed loop oxygen delivery system may be incorporated.Oxygen Flow Control Unit (OFCU 1)

[0048] The OFCU 1 comprises the heart of the disclosed closed-loop oxygen supply system. In some aspects the OFCU 1 comprises a proportional controlled valve. In one aspect, the proportionally-controlled valve provides an oxygen flow range of 0-15 LPM across pressures ranging from atmospheric pressure to 50 psi. In some aspects the proportionally controlled valve provides an oxygen flow resolution of about 0.7 standard liters per minute (SLPM) per step with a step position resolution of about 0.001 inches. In some aspects the proportional controlled valve can be a low power valve. In some aspects the power consumption can be less than 4 watts nominal during adjustment with zero power consumption to maintain position. In some aspects the proportional control valve has a duty cycle (fully open to fully closed) response time of less than one second. In some aspects a valve controller can control the proportional control valve. In some aspects a microprocessor can actuate the controller.

[0049] Various alternative aspects of the OFCU 1 device are indicated within the shaded box within FIG. 2. External devices and connections are indicated outside of the shaded region. Within the figure, oxygen flow is depicted by the heavy black lines, electrical signals are depicted with a light black line, and wireless signals are depicted with dots for Bluetooth and dashes for Wi-Fi. The oxygen derived from an external source 2 flows to the patient depicted in the lower right comer of the figure. The oxygen first enters the OFCU 1 through a port of device 4 configured to fit a standard oxygen delivery tube and enters the proportional controlled valve 8. The proportional controlled valve regulates the flow rate of the oxygen transiting the valve body in response to inputs derived from a microcontroller control unit 9. The microcontroller control unit 9 receives power from a power regulator 10.

[0050] In some aspects the power regulator 10 may be powered by an integral battery 13, for example a lithium battery delivering 6,000 mAh at 5.0V. In some aspects of9121390.098910\4899-1153-2431.2the disclosure the microcontroller control unit 9 can be powered by an external power source, such as an external charging device via USB-C (5V). The power regulator 10 can also be responsible for charging and regulating the discharge of the battery 13. The battery may also be charged via a USB-C data port 12, which also allows two-way data transfer between external devices connected to the USB-C data port 12 and the microcontroller control unit 9. The USB-C data port 12 may also be used to update or modify the microcontroller control unit 9 firmware.

[0051] The microcontroller control unit 9 can be in further communication with an integrated main display. In some embodiments, the display may be located off-device (e.g., on a device other than the OFCU 1). For example, the display may be located on the Version 1 device or a Smartphone 15 on Versions 2 device and the top display on Version 1 device or Smartwatches (e.g., concentrator 7) on Version 2 devices. In some embodiments, the microcontroller control unit 9 may be in communication with a display that is on device (e.g., on the OFCU 1 and / or the unified control knob 29).

[0052] In some embodiments, the microcontroller control unit 9 may be in communication with the Bluetooth module 14 to allow Bluetooth communication with external device(s) 5 and may also provide a dedicated port for wired sensor inputs (e.g., sensor 3) such as pulse / oximetry sensors. The microcontroller control unit 9 may also comprise additional random-access memory storage 11 for extended data logging and nonvolatile storage of such data. The microcontroller control unit 9 also may activate an audio, visual or haptic alarm 18 upon reaching certain system and user-defined alarm states.

[0053] As illustrated in FIG. 2, the microcontroller control unit 9 may be in communication with trim table function 19. In some embodiments, trim table function 19 may be a subcomponent of the microcontroller control unit 9. At FIG. 2, trim table function 19 may enable the microcontroller control unit 9 to determine a trim value to be applied to the stepper control proportional valve, thereby controlling the supply of oxygen to a subject. In some embodiments, trim table function 19 may be utilized by the microcontroller control unit 9 to determine a trim value based on physiological conditions of the subject, environmental conditions (e.g., altitude and temperature), an oxygen flow command, and a current trim value of the stepper control proportional valve.

[0054] In various other aspects the microcontroller control unit 9 can communicate with sensors to adjust the valve settings to regulate the flow of oxygen to the subject based on the sensor input in a programmed manner to provide a closed loop system. In some aspects the amount of oxygen delivered by the system to the subject can be determined by the 10121390.098910\4899-1153-2431.2physiological demand of the subject as detected by the sensors. In some aspects the amount of oxygen delivered by the system to the subject can be determined by the amount of available oxygen in the oxygen source. The microcontroller control unit 9 may be programmed to deliver oxygen at varying rates in response to any sensor input signal and can be programmed to evaluate different responses to various combinations of sensor input.

[0055] In some aspects the Signal Extraction Pulse Oximetry Platform unit comprises a 9 or 10 pin power and communications connector. In some aspects the Signal Extraction Pulse Oximetry Platform comprises a 20 or 25 pin sensor connector. In some aspects the Signal Extraction Pulse Oximetry Platform comprises a DC input voltage of about ±5% consuming 310 mW to 600 mW power. In some aspects the Signal Extraction Pulse Oximetry Platform comprises a serial interface with a variable baud rate of 9600 bps to 57,600 bps. In some aspects of the disclosure the Signal Extraction Pulse Oximetry Platform unit can be configured in a single stack or a double stack orientation. In some aspects of the disclosure the microcontroller control unit 9, Bluetooth module, Wi-Fi module and pulse / oximetry sensor control electronics are in a single stack configuration and in some aspects they are in a double stack configuration mounted on one side of a wafer and the battery on the opposite side (e.g., as depicted by user interface element 22 of FIG. 4A) to provide thermal and electrical separation.

[0056] In various aspects the microcontroller control unit 9 can execute programmed responses to sensor input data based on a plurality of executable modules. In some aspects the microprocessor can log sensor data. In some aspects the logged data may be retained in non-volatile memory. In some aspects the non-volatile memory may comprise a removable device. In some embodiments the logged data may be transferred wirelessly to an external computing device. In some aspects data can be transferred to a HIPAA certified external computing device via Bluetooth devices, Wi-Fi or LTE communication. In other aspects data may be transferred via a direct physical connection that may be established between the microprocessor and the external computing device via a cable or wire utilizing the USB-C connector. All data can be encrypted before being transmitted to any external destination.

[0057] The data captured can include pulse, saturation level, respiration rate, perfusion index, latitude, longitude, speed, altitude, humidity, barometric pressure, temperature and air quality index (AQI). The data can be sent to the patient's medical team using the device's " Phone Home" feature to facilitate Remote Patient Monitoring (RPM).11121390.098910\4899-1153-2431.2RPM is increasingly important attribute of modern medicine and a valuable tool during times of heightened concern about communicable diseases.

[0058] In one aspect the microprocessor communicates with output devices housed within the OFCU 1, which may comprise character or graphical LCD screens, light sources such as LEDs, as well as haptic or audible alarms. In some aspects the microprocessor communicates with external computing devices to receive instructions, display oxygen flow rates, pulse rates and perfusion index and to activate alarms or status indicators on such devices.Data Display and Control Elements of the OFCU 1

[0059] In one variation as depicted in FIG. 3, the OFCU 1 can be approximately as tall as 15 cm. and as wide 7.5 cm and approximately twice as thick 2 cm as a large cell phone.

[0060] In another variation depicted in FIG. 4B, the OFCU 1 can be approximately as tall as 12 cm, and as wide as 8 cm and as thick as 3 cm. In some variations the OFCU 1 may be mounted semi-permanently to the oxygen regulator with a stainless-steel braided Teflon (or similar) hose {24) connecting the output port of the oxygen regulator and the in put port of the OFCU 1 manifold 25. In one variations, the oxygen will enter the input port of a manifold which supports two valves. One valve controls the flow of oxygen in normal operation of the OFCU 1 (e.g., using valve 8). The other valve can be normally closed however in the event of a system malfunction, the automatic bypass valve 26 opens completely, allowing 100% of the flow rate in LPM the patient has set as the upper limit in flow rate on the oxygen regulator. In these variations, all control of the OFCU 1 can be managed through the use of an application on a Smartphone. In these variations, all important data can be displayed on the users’ Smartphone and their Smartwatch if they have authorized the connections.

[0061] With reference to FIG. 3, the side view of the OFCU 1 displays (e.g., using smartphone 15) the level of oxygen saturation reported by the sensor input and may also depict other programmed information such as remaining oxygen, pulse rate and battery’ status. In some examples, controller 9 may be configured to cause the OFCU 1 to display the remaining quantity of oxygen. The side panel may also provide a touch screen or other input device to allow the subject to input initial parameters to control the OFCU 1. On the back panel an on / off toggle button may be found, which may be lit to indicate status. The Emergency bypass function operates automatically in the unlikely event of any type of12121390.098910\4899-1153-2431.2malfunction of the OFCU 1 and provides the patient with 100% of the flow rate in LPM selected on the oxygen regulator.

[0062] In FIG. 3, the front of the OFCU 1 may also possess an audible piezo buzzer alarm (e.g., haptic alarm 18) and a reset button such as a soft button on the screen or Smartphone / Smartwatch to silence the alarm signals. The OFCU 1 may possess a clip 20 for attachment to a belt or other article of clothing.

[0063] In FIG. 3. the bottom of the OFCU 1 contains a connector for wired pulse / oximeter probe or similar sensor 3. The back of the OFCU 1 provides a slot for accepting an SD card (e.g., memory storage 11).

[0064] In FIG. 3, the top of the OFCU 1 may repeat the side display (e.g., of concentrator 7) of oxygen saturation reported by sensor input and may also depict other programmed information such as remaining oxygen and battery status. The background color of this display will change as the quantity of remaining oxygen decreases.

[0065] In FIG. 3, at the bottom the OFCU 1 houses the oxygen ports that provide oxygen into the OFCU 1 (e.g., device 4 for In, and outflow 16 for Out) from the unit to the subject and a port for accepting sensor input (e.g., sensor 3) such as a 20 or 25 pinjack commonly found on pulse / oximeter sensors as well as a USB-C data port 12 for charging, data exchange, and firmware updates.Regulated Oxygen Delivery

[0066] In some aspects of the disclosure the regulated oxygen delivery system comprises a cannula or trans-tracheal tube. Other methods of oxygen delivery are known to those of skill in the art. The range of regulated oxygen delivery systems can be constrained by the ability of such systems to attach to the ports on the OFCU 1. which may be similar, if not identical, to the tubing attached to the oxygen source and are industry standard.Sensors

[0067] In some aspects (e.g.. FIG. 2) the sensors comprise a digital tank pressure sensor 6 for sensing the remaining available oxygen reserve. In some aspects the sensors comprise a global navigation satellite system (GNSS) that tracks navigation signals from the satellite systems operated by options such as the US Global Position System, the EU’s Galileo navigation system, the Russian GLONASS navigation system and the Chinese BeiDou (Compass) navigation system, to provide location, altitude, and timing information. In some varations the sensor may measure the atmospheric pressure, temperature as well as humidity. In some aspects the sensors monitor the condition of the battery or other power source. In various aspects the sensors comprise physiological sensors capable of monitoring the13121390.098910\4899-1153-2431.2subject's blood oxygen level (SpO2), pulse rate, and perfusion index. In some aspects the physiological sensor can be a pulse / oximeter. In some aspects the physiological sensor comprises a blood oxygen sensor - typically attached to the wrist, ear or finger of the subject that uses light (multiple wavelengths) to accurately sense blood oxygen levels of the subject.

[0068] In one aspect the pulse / oxygen sensor may be connected to the OFCU 1 utilizing either a Bluetooth or wired connection. In some aspects the pulse / oximetry sensor may include an electronic control (that includes the microcontroller 9, for example) that senses low perfusion and adjusts the sensor to provide a consistent signal to the microprocessor. In various embodiments, one or more sensors are configured to detect subject data accurately. When a subject is in motion, data from the subject can become inaccurate if the sensor is dislodged or moves. In various aspects, one or more sensors provided herein provide information that is accurate and related to motion. Examples of such sensors include the RD SET El sensor, and include Rainbow SET Board - MX-5 circuit board by Masimo, or other sensors that are designed to be motion certified by the FDA to read in motion. In this aspect, the sensor input may be configured to receive data from sensors designated to be motion certified, along with other functions including recording a subject's actual blood oxygen level and calculating a difference between the subject's actual blood oxygen level and the subject's target blood oxygen level.

[0069] The modular nature and variation of the disclosed device can allow as many forms of physiological sensors to provide input to the OFCU 1 as possible. In some variations, the closed-loop system can include sensors keyed to detecting leading indicators of oxygen demand (such as pulse rate) and can provide a distinct advantage over sensors measuring lagging indicators of oxygen demand (such as respiratory or percent blood oxygen level).Failsafe design of the OFCU 1

[0070] In various aspects of disclosure, the oxygen flow control unit can automatically bypass the microprocessor control of the proportional controlled valve. The OFCU 1 verifies the health of the entire system every two seconds and if it determines there is any malfunction that would interrupt the flow of oxygen to the patient, the system automatically bypasses the closed loop oxygen delivery system and supplies oxygen to the subject directly from the oxygen regulator via the bypass valve 26 (e.g., as illustrated in FIG.4A). The bypass valve 26 is a normally open valve that is electrically kept in the closed position during routine operation, if an OFCU 1 problem is detected, the power can be removed from the valve and it moves to the fully open position.14121390.098910\4899-1153-2431.2

[0071] As shown in FIG. 4A, a crucial mechanical part of the OFCU 1 is the Control Proportionally Valve (e.g., valve 8) which may comprise two pressure sensors embedded into its design providing tank side and patient side pressure measurements. The Automatic Bypass capability provides a smooth step back to the current state of the art in oxygen regulator control. The Automatic Bypass provides the liters per minute flow the subject has set as their personal worst-case scenario (pulse or continuous flow), on their oxygen regulator or concentrator. As illustrated in FIG. 4 A, OFCU 1 may include mounting point 21 which may be used to mount OFCU 1 to other physical components. As illustrated in FIG. 4A, OFCU 1 may include USB-C data / charging port 23 which may be used to charge OFCU 1 or load data to OFCU 1. As illustrated in FIG. 4 A, OFCU 1 may include internal manifold fitting 28, which may be configured to connect proportionally controlled valve 8 to bypass valve 26. In this way, internal manifold fitting 28 may regulate oxygen flow during normal and fault operations.

[0072] In various aspects, as shown in FIG. 4B, the OFCU 1 may be mounted and secured to the oxygen regulator. The merit of this approach relocates the majority’ of the weight of the OFCU 1 to a pull cart location or to a backpack-carried oxygen tank. This aspect of disclosure also allows the users’ Smartphone to control the OFCU 1 receiving all important data / readouts and make any required changes to serve their needs using a proprietary software tool.

[0073] The example Oxygen Flow Control Unit (OFCU 1) can be a portable device configured to provide a closed loop oxygen delivery system by automatically adjusting the How of oxygen to the subject through a proportional controlled valve in response to the oximetry readings derived from a pulse / oximetry sensor attached either at the Cavum Conchae (central part of the outer ear), forehead, foot, finger or other part of the subject. The Cavum Conchae can be the preferred location as it can be more sensitive to changes in oximetry, providing as much as 2 minutes notice versus finger-located sensors, providing better predictive changes in flow rates. The second-best location can be the ear lobe, also part of the body's core circulation. Additionally, the ear location can be hidden more easily by subjects and can be utilized, for example, in freezing weather (low perfusion) and wearing gloves and / or hats or caps.

[0074] Based on sensor input provided to the microprocessor and in conjunction with pre-programmed responses encoded in the microprocessor firmware, the microprocessor modifies the proportionally controlled valve settings to increase or decrease the flow of oxygen through the valve.15121390.098910\4899-1153-2431.2

[0075] The OFCU 1 can be connected via standard oxygen tubing between the source of supplemental oxygen, either tank or concentrator (portable or home unit), and the person requiring supplemental oxygen. The OFCU 1 utilizes the oxygen regulator present on oxygen tanks to reduce the pressure of the oxygen source from 400-2000 psi to a safer, low pressure (5-35 psi), delivered to the subject. When utilizing oxygen concentrators, the OFCU 1 can be connected via oxygen tubing between the concentrator and the subject. Lower pressure operation and commonly available tubing were selected because supplemental oxygen patients and their caregivers are familiar with these durable medical equipment (DME) supplies, and they are readily available. The OFCU 1 may be connected to the subject utilizing either a cannula, a trans -tracheal device, or other devices known to those skilled in the art. The OFCU 1 provides the subject with precise amounts of supplemental oxygen, in real time, whether the subject is utilizing pulse or continuous flow operation and may be used in lieu of pulse regulators (pulse regulators are six times more expensive than continuous flow regulators) as the OFCU 1 automatically conserves oxygen while in operation.

[0076] The OFCU 1 measures, controls, communicates and reports key health and system parameters to the subject (and their medical staff if desired). The oxygen sensor may also detect SpO2, pulse rate and perfusion index (Pl). Pl provides an evaluation of the quality of the signal and automatically adjusts the sensing parameters to provide quality data to the microprocessor. In various aspects of disclosure some OFCU 1 models may also measure methemoglobin saturation, acoustical respiration rate, total hemoglobin, carboxyhemoglobin, Methemoglobin pulse co-oximetry, Pleth Variability Index, Respiration Rate, Desat 3D Alarm, Pi Delta Alarm, Plethysmograph Waveforms, Oxygen Reserve Index Signal Identification, Quality Indicator and oxygen content. These additional measurements are especially valuable for emergency transport situations or elite athletes and medical studies in a real-life setting (the subject "living their life," not an artificial test as in a clinical setting), especially when the subject can be active at higher altitudes or encountering contact with hazardous materials such as byproducts of uncontrolled fires.

[0077] When the subject first turns the unit on, the integrated input / output displays (e.g., using smartphone 15, as illustrated in FIG. 2) may query the subject if they are using an oxygen tank or a concentrator as their supplemental oxygen source. If the subject selects a tank, the tank digital pressure sensor 6 (e.g., as illustrated in FIG. 2) can be queried to determine the beginning pressure. This information will be used as part of the predictive supply management operation to keep the subject informed of remaining supplemental oxygen, a process that actively tracks initial tank pressure, current tank pressure and flow 16121390.098910\4899-1153-2431.2rate, and alerts the subject as the oxygen in the tank is depleted. The OFCU 1 may alert smart devices the subject has paired with the OFCU 1 (smart watches, fitness trackers and Bluetooth enabled hearing aids (e.g., as illustrated in FIG. 2), as the quantity decreases from full to ¾ to ½ to ¼ thereby helping the subject avoid running out of oxygen. There may also be an alarm 18 (e.g., as illustrated in FIG. 2) that can alert the subject audibly if they have not paired any smart devices with the OFCU 1 or a small vibrator to provide haptic notice. The display of concentrator 7 (e.g., as illustrated in FIG. 2) on the OFCU 1 can also change its background color as the available oxygen decreases from full to ¾ to ½ to ¼ remaining and may be equipped with a dedicated alarm light.

[0078] The user can also be prompted to decide if they desire to have emergency personnel called if a subject's selectable preset parameter is breached, such as 55% blood oxygen saturation level and no movement sensed by the Global Navigation Satellite System (GNSS) sensor.

[0079] The user will be asked what their desired minimum oxygen saturation level should be (88% can be the lowest that may be selected as that is the generally agreed minimum safe level except in the Elite Athlete mode). The user can also input their desired maximum heart rate during strenuous activity. The user can be alerted to excursions below (or above for heart rate) preset limits audibly and visually on smart devices and reports generated from data logged during the period. The user can be instructed as standard procedure to set their supplemental oxygen source regulator at the highest flow rate (in LPM) they have needed in the past. That can be a key safety feature of the OFCU 1. In the event of battery failure or a malfunction of the OFCU 1, the Automatic Bypass feature bypasses all internal OFCU 1 features and delivers 100% of the oxygen flow the user set before that day's activities, giving them time to reach the source of their oxygen supply and manually set the flow rate actually required in their current situation (not the worst case scenario).

[0080] After those few user-selected settings are established, the user can begin their daily activities without requiring constant system monitoring. The OFCU 1 will continually monitor their oxygen saturation, pulse and perfusion while automatically adjusting the flow of their supplemental oxygen, maintaining them at the desired level (+2%).

[0081] The OFCU 1 stores pertinent data related to the activities of the user if they desire. Such data may be stored locally and / or uploaded to other devices such as a cloudbased HIPAA storage system for review or sharing with medical personnel. Data stored and transmitted by the OFCU 1 may include, without limitation, in graphical and tabular formats: pulse readings, oxygen saturation levels, oxygen flow rates, as well as speed17121390.098910\4899-1153-2431.2traveled / elevation (above / below sea level) and the air quality index (AQI) for their location and time as determined by the GNSS module. The tabular format includes the number of events where the user exceeded the parameters they selected during setup (pulse-high or low and highest and lowest measurement, oxygen saturation-high or low and highest and lowest measurement and averages for both). All of the data can be graphically displayed with a data point gathered every few seconds. All reports can be produced in a format that can be printed or emailed to the user’s medical team. The 32 GB SD Card can be sufficient to store eight years of reports for comparison purposes and for later investigation / sharing with the user's medical team.Example 2

[0082] User blood oxygen saturation was maintained during the transition from sedentary’ to active states. A 70-year old male subject suffering from idiopathic pulmonary fibrosis (IPF) was fitted with an OFCU 1 connected to a high-pressure portable oxygen bottle via 4 feet of tubing to the OFCU 1, which was in turn connected to a 4-foot-long nasal cannula oxygen delivery system to form a closed-loop oxygen delivery system. The OFCU 1 included a MASI MO RD SET-E1 pulse / ox sensor attached to the user at the subjects' Cavum Conchae. The OFCU 1 was programmed to maintain the user's oxygen saturation level at 91% regardless of the level of activity, and the oxygen regulator was set to 6 LPM pulse. This flow rate would be very wasteful without using the OFCU 1, because it can be more than the user needs while sedentary; however, it can be the correct flow rate during times of increased / vigorous activity such as walking up an incline.

[0083] During a sedentary test period (sitting) the user's observed blood oxygen saturation level was held at a constant 91% ± 1%, while during the active test period (vigorous walking on an incline between 6800 feet and 7500 feet above sea level) blood oxygen saturation varied between 89% - 92%. Under similar conditions, while using supplemental oxygen set to the prescribed flow rate but without use of the OFCU 1, the subject experienced a sedentary blood oxygen saturation of 82% - 89% and an active blood oxygen saturation variance of 72% - 92%.Example 3

[0084] Oxygen duration charts (Table 2) show that the standard M9 / C oxygen tank (255 liters) lasts 2.5 hours when used at 4 LPM (pulse). Organizational testing has been completed on the Inogen One GS using a single battery at an altitude of 7500 above sea level. Additional data has been gathered from the GS's Technical Manual (See Table 1 below) related to the size of the bolus at different settings.18121390.098910\4899-1153-2431.2by the

[0085] Many of the portable oxygen concentrators do not in fact deliver oxygen in a relationship of 1 = 1 LPM, 2 = 2 LPM. The GS effectively provides 1 LPM continuous at a setting of 6 with 17 breaths / minute.

[0086] Added for comparison and to provide a level playing field for all sources of portable oxygen, the liquid oxygen portable reservoir from Puritan-Bennet, Companion 500, was added to Table 2 below. Testing has shown that using an M9 / C tank set to 4 Lpm (pulse) when used in connection with the OFCU 1 (normal daily activities) consistently lasts over 8 hours by conserving oxygen when the user is sedentary and providing sufficient oxygen when activities levels are higher, resulting in a user who is better served with proper oxygen to their brain and critical organs (heart, etc.). The device with the highest performance at each flow rate is shown with a black background and white numbers.1 i1 I _ 4 ’ _ _ _ _ _f s1 6^Example 4

[0087] The OFCU 1 may function in an emergency medicine support role, wherein multi-frequency blood sensors coupled to the OFCU 1 capable of real -time detection of circulating specific hemoglobin levels in a user can be used to monitor a patient for internal bleeding. In other cases, a detector capable of real -time monitoring of specific carbon monoxide levels can be coupled to the OFCU 1 to provide supplemental oxygen to first responders or patients exposed to combustion products that impact their ability to absorboxygen. In addition, real time detectors of methemoglobin (a side effect of many drugs used by hospitals and first responders) can be monitored through the OFCU 1 and supplemental oxygen can be data to patients as necessary. Respiratory rate can be monitored by real -time analysis of pressure in the oxygen delivery tube connected to the OFCU 1 to provide a log of a patient's respiration. The portable, low profile, data logging, closed-loop system represented by the OFCU 1 increases the amount of actionable data delivered to the medical team serving a patient and can be transmitted while in route to the receiving medical facility, so they have high quality, specific data for this particular patient.

[0088] The OFCU 1 may also find uti 1 i ty in military operations. In particular, the system provides vital remote data monitoring functions in Chemical, Biological, Nuclear and Radiological and Explosive (CBNRE) scenarios. Using telemetry to transmit the encrypted data to a command location while also ensuring reliable, appropriate oxygen saturation levels to the participant can be a capability multiplier. The future battlefield can include many combatants using exoskeletons to assist individuals carrying the extreme weight of communications and navigation equipment and munitions; however, such exoskeletons also inhibit the body's capability to dissipate heat, exacerbating the user's need for supplemental oxygen. The OFCU 1 reads and reacts to this user's needs and delivers precise amounts of supplemental oxygen to keep the military member functioning at their peak while providing high quality, specific medical data to secure radio transmitting equipment, giving commanders actionable information.

[0089] It will be recognized that any variation disclosed herein can be combined with any other variation, whether or not such variations are described together or separately.

[0090] The table below highlights various benefits that the OFCU 1 described herein provides over existing devices:Function Existing Devices OFCU 1Closed-Loop Delivered adequate Enhanced operational modes with Control with SpO₂ control activity and breath sensing SensingSpO₂ Monitoring Wired sensors (ear & Wireless SpO2 sensors with advanced hand) with motion Bluetooth connectivityartifactsOxygen Delivery Rapid flow increases; Tailored proportional valves for patients desired smoother precise flow (up to 6 LPM) deliveryPortability Battery limitations due to Improved electronics for longer design constraints battery life and smaller form factor20121390.098910\4899-1153-2431.2Function Existing Devices OFCU 1Patient Monitoring Swappable SD card with Cloud connectivity enabling near real-time monitoring and long-term remote monitoring

[0091] FIG. 5 illustrates the Device Operating Modes. The Device Operating Modes provide a sophisticated flow control system. Each mode caters to different scenarios, ensuring optimal oxygen delivery, oxygen saturation, and user comfort. The oxygen flow control unit (OFCU 1) device will be able to supply oxygen to COPD and ILD users to maintain their oxygen saturation levels during activity.

[0092] Each of these modes represents a different strategy for oxygen delivery, tailored to various users' needs and system performance levels. From maintaining a consistent flow to adapting to the breathing pattern and activity levels of the user, the system ensures that oxygen delivery is both efficient and responsive to changing requirements. This adaptability is crucial in providing user-centered care in diverse oxygen-demand situations from the user.

[0093] The device significantly aids in oxygen conservation through its autotitration modes of operation, adjusting oxygen flow to the user’s precise needs, and its autotitration-breath-sensing modes add breath-sensing technology to further reduce waste by providing oxygen mainly during the inhalation part of the breathing cycle. The modes of operation are:

[0094] Bypass Mode: This mode can be activated when battery power is not available, when a critical error is detected in the control system, or when the user manually activates this mode with the bypass button. A separate system will monitor the operation of the primary control systems and will force the system into this mode if any critical errors are detected. In this mode, a secondary flow path can be activated where the user mechanically sets the oxygen flow rate using a knob like those on standard oxygen regulators.

[0095] Constant Flow Mode: This mode can be activated when the system powers on, when there is a loss of the SpO2 signal, or when the system exits Bypass Mode. This can be the default starting mode of the system, which it uses until a SpO2 signal with adequate confidence is available. In this mode, the system maintains the default oxygen flow setting set by the user, caregiver, or therapist.

[0096] Titration Mode: This mode can be activated from constant flow mode when a SpO2 signal with adequate confidence is available. In this mode, oxygen flow can be21121390.098910\4899-1153-2431.2dynamically adjusted to maintain the SpO2 of the user at the configured setpoint. In this mode, only the SpO2 signal is used to determine the oxygen flow rate.

[0097] Titration and Breath Mode: This mode can be activated from Titration Mode when a breath signal with adequate confidence is available. The system will revert to Titration Mode if the breath signal confidence falls below an adequate level or if the SpO2 of the user falls below a specified threshold. In this mode only the SpO2 signal can be used for determining the oxygen flow rate, and flow can be introduced during detected inspirations and reduced between breaths.

[0098] Titration and Activity Mode: This mode can be activated from the Titration Mode when an activity signal with adequate confidence is available. The system will revert to Titration Mode if the activity signal confidence falls below an adequate level. In this mode both the SpO2 and activity signals are used for determining the oxygen flow rate.

[0099] Titration, Breath, and Activity Mode: This mode can be activated from either the Titration and Breath Mode or the Titration and Activity Mode when breath and activity signals with adequate confidence are both available. If confidence in either the breath or the activity signal falls below an adequate level, the system will revert to the appropriate oxygen flow mode. If the SpO2 of the user falls below a specified threshold, the system will revert to Titration and Activity' Mode. In Titration, Breath, and Activity' mode both the SpO2 and activity signals are used for determining the oxygen flow rate, and flow can be introduced during detected inspirations and reduced between breaths.

[0100] The control system in the OFCU 1 can function as a simple proportional control mechanism, where oxygen flow adjustments were based solely on real-time deviations from the desired SpCh setpoint. This non-iterative approach can result in direct adjustments to oxygen flow without predictive or historical correction factors. The simplified equation for this Titration Mode can be:O(t) = Obase+ Ks· e(t)Equation 1- Single Variable Controlwhere:- O(t) is the oxygen flow rate at time t.- Obase is the minimum oxygen flow rate for the device. Currently set at 0.5 LPM (Liter per minute)- e(t) is the error signal, defined as the difference between the desired SpO₂ setpoint and the measured SpO₂ value.22121390.098910\4899-1153-2431.2- Ks is a simple proportional gain constant that determines the direct response of the system based on e(t).Augment Control System

[0101] The OFCU 1, which has additional sensors and control, can alternatively utilize an expanded control equation to increase the quality of oxygen delivery and to maintain saturation levels. The additional sensors and algorithmic functionality enable the addition of algorithmic logic across different modes of regulation, further reducing oxygen consumption (estimated to be up to 60%). The algorithmic logic added three additional control modes to the titration mode - breath sensing mode, activity-based mode, and a combination of breath sensing and activity-based modes. Table 1 - Measure Quantities for System Control shows the Measured Variables in the system, where the variables in grey are part of the previous controlled system, and the variables in white are additional variables in the latest version of the device.02 BolusBreath EPOCH time msBreath Duration msInspiration durationNegative Inspiratory Pressure (Force) mmHgMax pressure exhale mmHgTidal Volume mlAccelerometer gMovement, XYZ min / max gSpO2 * HRBarometric pressure mmHgTemperature CHumidityTable 1- Measure Quantities for System Control

[0102] To ensure optimal closed-loop oxygen titration, a Proportional-Integral-Derivative (PID) Controller can be implemented. The oxygen flow rate can be dynamically adjusted based on the following control equation, which contains timed variables and proportional factor for greater tuning of the oxygen supply:dO(t) = Obase+ Kp· e(t) + Ki∫ e(t)dt + Kdd / dt e(t)Equation 2- Multiple Variable EquationWhere:23121390098910\4899- 1153-2431 2- 0(t) is the oxygen flow rate at time t.- Obase is the minimum oxygen flow rate for the device. Currently set at 0.5 LPM (Liter per minute)- e(t) is the error signal, defined as the difference between the desired SpO₂ setpoint and the measured SpO₂ value.- KPis the proportional gain, determining how much the system reacts to the error magnitude.- Ki is the integral gain, which accounts for accumulated past errors to reduce steadystate error.- Kdis the derivative gain, predicting future trends based on the rate of change of the error.

[0103] The PID controller can ensure smooth and responsive oxygen delivery adjustments, preventing sudden fluctuations while maintaining SpO₂ levels within the desired range.PID Control Equation General Terms

[0104] Proportional Term Immediate Response:Kpe(t)This term can be used to determine the immediate reaction of the system based on the current error e(t), which can be the difference between the desired SpO₂ setpoint and the measured SpO2value. A higher KPcan increase responsiveness, allowing the sy stem to correct deviations quickly. However, excessive KPcan lead to overshooting or oscillations.

[0105] Integral Term Accumulated Error Correction:Kie(t) dtThis component can sum past errors over time to correct for prolonged deviations. It can ensure that minor but persistent discrepancies do not accumulate and affect long-term SpO₂ stability. A well-tuned Kihelps eliminate steady-state error, but excessive integration can lead to lagging corrections.

[0106] Derivative Term Predictive Adjustment:de(t)ddt24121390.098910\4899-1153-2431.2This term considers the rate of change of the error signal, predicting trends to prevent sudden fluctuations. If SpO₂ can be rapidly increasing or decreasing, Kdmodulates oxygen delivery accordingly to prevent large overshoots or undershoots. Too much derivative action can introduce instability if not properly damped.

[0107] Proportional Gain and Dynamic Modulation. The error signal e(t) modulates the amplitude of oxygen delivery adjustments. The proportional gain Kp defines the system's baseline response and adjusts its direction dynamically to adapt to external factors. Instead of scaling the error directly, KPmodifies the standard reaction pattern, ensuring stable titration under vary ing conditions.

[0108] To enhance Kpadaptability, can be dynamically adjusted based on one or more of: user motion (based on measured accelerometer data), fluctuations in respiratory effort (based on measured Negative Inspiratory Pressure, measured Exhale Pressure, measured Breath Duration), breath rate changes (a secondary signal indicating variations in oxygen demand), heart rate variability (additional physiological insight into systemic response), and measured SpO₂ (primary determinant for oxygen titration adjustments). By continuously tuning its response, the system can ensure accurate and stable oxygen flow regulation tailored to the user’s needs.

[0109] To further refine system responsiveness, the updated framework can incorporate tuning tables based on experimental results from trials. These tuning tables enable dynamic adjustments in real-time, ensuring optimal system performance across varying physiological conditions.

[0110] Table 2 below represents an example of how KPadapts under different trial-based Conditions. " Condition" in the table represents the overall physiological state or external influences affecting the user’s oxygenation needs. It serves as a general classification based on multiple physiological parameters, helping to categorize different response scenarios for the PID controller. For example, in a “low condition,” the user can be stable with minimal movement, steady breathing effort, and normal vital signs, requiring only baseline oxygen adjustments. In a “moderate condition,” some noticeable movement or minor fluctuations in respiratory effort and vital signs can exist, indicating the need for a slightly increased oxygen response. In a “high condition,” there may be significant user motion, rapid breathing rate, erratic breathing effort, and elevated heart rate, requiring the system to actively modulate oxygen delivery while damping excessive fluctuations.25121390.098910\4899-1153-2431.2Condition Patient Breath Breathing Heart Measured Adjusted Motion Rate Effort Rate SpCh KpLow Minimal Stable Stable Normal Within Standard range KpModerate Noticeable Slight Slight Slight Slightly Increased Increase Fluctuation Elevation Low KPHigh Significant Rapid Erratic Elevated Low Adaptive KPwith dampingTable 2 - Proportional Gain Adaptations

[0111] This structured adaptation ensures the system responds efficiently to real-world conditions while maintaining stability and patient safety.Integral Gain and Cumulative Error Modulation

[0112] The function of Ki can be to modulate O(t) based on the accumulated error differential rate between the measured SpO₂ and the expected SpO₂. If the rate of SpO₂ change can be lagging behind the expected response in the patient, this additional factor attempts to modify the adjustment rate to reduce the difference between the observed and expected saturation levels. By integrating past errors, Ki helps to correct for deviations, ensuring that oxygen delivery more accurately aligns with the patient's physiological needs.

[0113] Table 3 below illustrates how Kiadapts based on the rate of change difference between observed and expected SpO₂:Rate of SpO₂ Change Difference Adjusted ResponseLow (<2% deviation) StandardModerate (2-5% deviation) IncreasedHigh (>5% deviation) Adaptive with dampingTable 3 - Integral Gain Adjusted Response

[0114] This adjustment ensures that cumulative error correction can be effectively applied to improve the alignment of oxygen flow regulation with user needs.

[0115] Derivative Gain and Dynamic Rate Control. The function of Ka can be to modulate O(t) based on the rate of change of the error signal e(t), which represents the difference between the desired and measured SpO2 values. This term predicts trends in the error signal and adjusts the oxygen supply accordingly to mitigate potential instability.26121390.098910\4899-1153-2431.2

[0116] The last variable Ka(d / dt*e(t)) can be a dynamic variable that commands the rate of oxygen supply to change based on the rate of change of the measured oxygen saturation. If the rate of change can be rapidly increasing or decreasing, as derived from sample-to-sample measurements, the system can command a more rapid change in the supply. Additionally, it can adjust the magnitude proportionally to the expected difference to prevent large overshoots or undershoots of the saturation level. This predictive capability enhances the responsiveness of the system, helping to maintain SpCh levels within the target range while avoiding unnecessary fluctuations.

[0117] Table 4 below illustrates how Kdadapts based on the observed rate of change of SpO₂:Rate of SpO2Change Adjusted ResponseLow (<1% per second) StandardModerate (1-3% per second) IncreasedHigh (>3% per second) Adaptive with dampingTable 4- Derivative Gain Adaptation

[0118] The last variable can be a dynamic variable that commands the rate of oxygen supply to change based on the rate of change of the measured oxygen saturation. If the rate of change can be rapidly increasing or decreasing, as derived from sample-to-sample measurements, the system can command a more rapid change in the supply. Additionally, it can adjust the magnitude proportionally to the expected difference to prevent large overshoots or undershoots of the saturation level. This predictive capability enhances the responsiveness of the system, helping to maintain SpC>2 levels within the target range while avoiding unnecessary' fluctuations.

[0119] Tables 5-8 show additional measured variables and data for use in the algorithmic logic, according to variations of the disclosure.27121390.098910\4899-1153-2431.202 Bolus LBreath EPOCH time msBreath Duration msinspiration duration msNegative Inspiratory Pressure (Force) mmHgMax pressure exhale mmHgTidal Volume mlAccelerometer g^Movement, XYZ min / max gSpO2 r- HRBarometric pressure mmHgTemperature CHumidity %Table 5 - Dynamic Patient DataDisease StateDefault Flow Rate LPMMin Flow rate LPMMax Flow rate LPMPatient Weight lbsTable 6 - Static Patient DataTarget Flow Rate LPMKnob setting LPMActual flow rate average LPMMin / max flow during inspiration LPMSystem StateTank pressure PSIBattery Voltage / Current mV, mA Reported Go Time minNetwork connectivity / statusEstimated tank sizeSidewalk location ApproxTable 7 - Dynamic Device DataBreathing Effort -Thoracic Circumference cmWearable Strain Sensor (piezoelectric} V, peWearable Strain Sensor (capacitive) FaradsAcoustic Sensor dB, HzOptical & Photoplethysmography AmplitudeTable 8 - Other Variables and Sensors

[0120] FIG. 6A depicts graphs for oxygen saturation of arterial blood (SpO₂) reported from a pulse oximeter and supply of oxygen flow from an oxygen flow control unit28121390098910\4899- 1153-2431 2(OFCU 1), according to some variations of the disclosure. FIG. 6B describes operational guidelines depicted in Figure 6A.

[0121] FIG. 7A depicts graphs for oxygen saturation of arterial blood (SpO₂), pulse rate (PR), respiratory rate (RR), and oxygen flow rate (FLOW) from running the algorithm logic via a simulation engine that mimics lung activity for one hour of typical usage. FIG. 7B depicts a graph for the SpO₂ range and the time spent in each SpO2 range for the data in Figure 7A.

[0122] FIG. 8A depicts graphs for oxygen saturation of arterial blood (SpO2), pulse rate (PR), respiratory' rate (RR), and oxygen flow rate (FLOW) from running the algorithm logic via a simulation engine that mimics lung activity for one hour of typical usage, with a desaturation event. FIG. 8B depicts a graph for the SpO2 range and the time spent in each SpO2 range for the data in FIG. 8A.

[0123] FIG. 9A depicts graphs for oxygen saturation of arterial blood (SpO2), pulse rate (PR), respiratory' rate (RR), and oxy gen flow rate (FLOW) from running the algorithm logic via a simulation engine that mimics four weeks of lung activity for Short-Term Oxygen Therapy (STOT). FIG. 9B depicts a graph for the SpO2 range and the time spent in each SpO2 range for the data in FIG. 9A. FIG. 9C depicts a graph for the number of desaturation events per day (with and without activity) for the data in FIG. 9A.

[0124] FIG. 10A depicts graphs for oxygen saturation of arterial blood (SpO₂), pulse rate (PR), respiratory rate (RR), and oxygen flow rate (FLOW) from running the algorithm logic via a simulation engine that mimics twelve weeks of lung activity for a person with Chronic Obstructive Pulmonary' Disease (COPD). FIG. 10B depicts a graph for the SpO2range and the time spent in each SpO2 range for the data in FIG. 10A. FIG. 10C depicts a graph for the number of desaturation events per day (with and without activity) for the data in FIG. 10A.

[0125] FIG. 11 is a schematic of the System Architecture that incorporates a cloud component for Artificial Intelligence (Al) and machine learning (ML), according to some variations of the disclosure. The cloud component enhances device management, dashboard display and overall patient monitoring and management. As data becomes available, the system uses real-time controls from derived digital technologies to adjust oxygen delivery to the patient. The system may also provide a differential notification on disease progression based on aggregate population curve data versus the individual patient's oxygen demand curve data. Deviation on slope, or delta between the two curves will notify interested parties29121390.098910\4899-1153-2431.2on the changes, which may lead costs microphone solves to changes in the patient’s routines and treatment.

[0126] FIG. 12 is a schematic showing tailored response for oxygen delivery based on integration of feedback response over time and algorithmic logic, according to some variations of the disclosure. As shown, the device collects and sends user data to the algorithmic logic, which may include a cloud component that uses Al and ML technologies to analyze the data and tailor oxygen delivery (Individually Tailored Response Table). In one aspect, the tailored oxygen delivery may be based only on user data. In another aspect, the cloud component may adjust oxygen delivery based on user data and other timely load factors, such as environmental factors like wildfires.

[0127] Oxygen Supply can be provided as part of an updated recipe for the Algorithmic Logic based on time activity or behavior timing. The proportional function for oxygen supply O(t) can be modeled as:O(t)= Obase + K • P(t) ' f(R(t))where:- O(t)is the oxygen flow rate at time t.- Obase is the minimum oxygen flow rate for the device. Currently set at 0.5 LPM (Liter per minute)- P(t) is Al-derived predicted oxygen demand at time t (normalized, e.g. 0 to 1). May be based on trends or immediate needs.- R(t) is recorded physiological response at time t (e.g., heart rate, SpO₂, or respiratory rate, other factors).- f(R(t)) is a function of recorded physiological responses to previous oxygen levels. This could be linear or nonlinear, depending on the nature of the response.- K can be a proportional constant (adjusts the system's sensitivity to the prediction).

[0128] Oxygen Supply can be provided as part of an updated recipe for the Algorithmic Logic - and current load factors that are observed just prior to inspiration.( R(t) — Rmin\ O3(t) = Obase + K • [P(t) + / M-(t)]\ "max 'Lnin / where:- O(t): Oxygen supply at time t (liters per minute, L / min).- Obase Baseline oxygen supply (minimum amount, L / min).30121390098910X4899-1153-2431 2- K: Proportional constant controlling sensitivity to predictions and responses.- P(t): AI-derived predicted oxygen demand, normalized (e.g., 0 ≤ P(t) ≤ 1).- β: Weight assigned to the load factor L(t)L(t)L(t).- L(t): External load factor at time t, accounting for conditions like exertion or environment.- R(t): Real-time physiological response (e.g., SpO₂, respiratory rate, heart rate). - Rmin, Rmax: Physiological range limits for R(t)R(t)R(t).- a: Blending coefficient for real-time and historical corrections (0≤α≤10).- H(t): Historical correction factor based on past responses. - Or differential predictive factor - Vierzba Factor™

[0129] Historical correction factor based on past responses vs. predicted by typical disease progression. The historical correction factor adjusts oxygen supply based on deviations between expected and actual responses in the past, and the input response from historical disease patterns H(t).H t =∑ⁿᵢ₌₁ wᵢ^expected (A) ^actualWhere:H(t): Historical correction factor at time t.wi: Weight assigned to prioritize recent data points.AR(ti): Difference between expected and actual physiological responses. The expected physiological response can be the predictive model based on machine learning from the samples and the physiological state

[0130] FIG. 13 illustrates an example process 1300 for providing oxygen to a user according to some variation of the disclosure. At FIG. 13, the steps of process 1300 may be performed by a controller (e.g., microcontroller control unit 9) that is part of an oxygen delivery system configured to provide a closed loop oxygen supply to a subject (e.g., a user of the oxygen delivery system).31121390.098910\4899-1153-2431.2

[0131] The oxygen delivery system can include an oxygen flow control unit (OFCU), such as the OFCU 1, an oxygen source, and one or more sensors configured to detect physiological conditions associated with a user (e.g., patient) and / or environmental conditions associated with the user. The OFCU can include an electronically controllable oxygen-flow valve, a memory storing a plurality of context-trim entries, and a controller. Each of the plurality of context-trim entries can associate a stored context with a corresponding trim value. Each of the stored contexts can include at least one physiological parameter and at least one environmental parameter. Each of the trim values can be a stored correction factor that tweaks the oxygen-flow command when the system recognizes a context it has seen before. The controller can include at least one processor. The processor(s) can be communicatively coupled to the one or more sensors, the memory, and the electronically controllable oxygen-flow valve.

[0132] At step 1302, the controller may receive, from the one or more sensors, data. The data received from the sensors can indicate physiological conditions and / or environmental conditions associated with the user. At step 1304, the controller may determine a current context based on the received data. The current context may indicate a snapshot of the user’s current physiological state and / or environmental state. At step 1306, the controller may determine, based on a plurality of context trim entries, a first trim value corresponding to a stored context that matches a current context. The first trim value can indicate an amount by which to adjust the oxygen-flow. In particular, the first trim value can specify a magnitude and direction (positive to increase, negative to decrease) applied to the baseline oxygen-flow command computed elsewhere (e.g., by proportional / PID control, etc.).

[0133] At step 1308, the controller may determine an oxygen flow command based on the received data and the first trim value. Determining the oxygen-flow command can comprise executing a proportional-integral-derivative (PID) control algorithm. The PID algorithm can compute proportional, integral, and derivative terms based at least on a deviation between measured and target blood-oxygen saturation and a rate of change of blood-oxygen saturation.

[0134] The controller can be further configured to operate in a plurality of modes (e.g., at least one of a constant flow mode, a titration mode, a titration and breath mode, and a titration-and-activity mode). In each of the plurality of modes, the controller can be configured to determine the oxygen-flow command based on a different combination of physiological conditions and environmental conditions, such that the oxygen-flow command is determined based on the received data, the first trim value, and the mode.32121390.098910\4899-1153-2431.2

[0135] At step 1310, the controller may actuate an electronically controllable oxygen-flow valve based on the oxygen flow command, thereby supplying oxygen to the subject. In particular embodiments, actuating the electronically controllable oxygen-flow valve can include generating one or more control signals (e.g., electrical drive currents, pulse-width-modulated commands, or valve-position setpoints) that cause the valve to open, close, or assume an intermediate proportional position corresponding to the computed oxygen-flow command. By issuing these control signals, the controller regulates both the magnitude and timing of oxygen flow delivered through the valve, ensuring that the commanded flow rate is achieved at the outlet of the oxygen flow control unit. In some variations, the controller may further shape the valve's transition profile, such as controlling how rapidly the valve opens during inspiration (e.g., inhalation) or how gradually it closes during exhalation, based on additional physiological information (e.g., inspiratory effort or breath-phase detection). For example, the controller can be further configured to control a rate at which the electronically controllable oxygen-flow valve is opened in response to a strength associated with at least one inhalation, wherein the rate is increased for stronger inhalations and decreased for weaker inhalations. Once actuated, the valve permits oxygen from the oxygen source to pass through the output port and into the patient interface (such as a cannula or transtracheal delivery device), thereby supplying oxygen to the subj ect in accordance with the desired therapeutic regimen.

[0136] In embodiments, the oxygen delivery system further comprises a dual-lumen cannula including a flow lumen, configured to deliver the oxygen to the subject, and a sense lumen. The oxygen delivery system can include a pressure sensor coupled to the sense lumen. The controller can be further configured to detect at least one inhalation associated with the user and at least one exhalation associated with the subject based on a pressure signal obtained via the sense lumen. The controller can be configured to modify the oxygen-flow command based on the detected at least one inhalation and the detected at least one exhalation.

[0137] In embodiments, the oxygen source can comprise a tank filled with oxygen. The oxygen delivery system can further comprise a flow sensor configured to measure oxygen flow delivered to the subject. The oxygen delivery system can further comprise a pressure sensor configured to measure pressure in the tank. The controller can be further configured to determine a size of the tank and a remaining quantity of oxygen in the tank based on correlating an outlet flow rate (measured by the flow sensor) with a rate of tank pressure change (determined based on readings of the pressure sensor) over time. The 33121390.098910\4899-1153-2431.2controller can be further configured to cause display, via at least one user interface (e.g., either on the OFCU, such as on the unified control knob, or via a peripheral device), of an indication of the remaining quantity of oxygen in the tank.

[0138] In embodiments, the OFCU can further comprise a unified control knob that is adjustable by a user to select either one of a plurality of continuous oxygen flow settings or an automatic electronic mode in which delivery of the oxygen is controlled by the oxygen delivery system. The OFCU can include a watchdog safety circuit configured to monitor the controller and force the oxygen delivery system to deliver the oxygen via a mechanical bypass path based on determining a failure associated with the controller. The unified control knob can be configured to display an indication that the mechanical bypass path is active.

[0139] FIG. 14 can be a block diagram illustrating an example computing environment 1400 according to implementations disclosed herein. Any or all of the components depicted in FIGS. 1-12 can be implemented via the example computing environment 1400. Computing environment 1400 may comprise all or a part of all systems disclosed herein or utilized herewith, and / or may implement some or all of methods disclosed herein or utilized herewith. Computing environment 1400 may comprise hardware or a combination of hardware and software. The functionality to facilitate telecommunications via a telecommunications network may reside in one or combination of network devices 1400. Computing environment 1400 depicted in FIG. 14 may represent or perform functionality of an appropriate computing environment 1400, or combination of network devices 1400, such as, for example, a component or various components of a cellular broadcast system wireless network, a processor, a server, a gateway, a node, a mobile switching center (MSC). a short message service center (SMSC), an ALFS, a gateway mobile location center (GMLC), a radio access network (RAN), a serving mobile location center (SMLC), or the like, or any appropriate combination thereof. It can be emphasized that the block diagram depicted in FIG. 14 can be an example and not intended to imply a limitation to a specific implementation or configuration. Thus, computing environment 1400 may be implemented in a single device or multiple devices (example, single server or multiple servers, single gateway or multiple gateways, single controller, or multiple controllers). Multiple network entities may be distributed or centrally located. Multiple network entities may communicate wirelessly, via hard wire, or any appropriate combination thereof.

[0140] Computing environment 1400 may comprise a processor 1402 and a memory 1404 coupled to processor 1402. Memory 1404 may contain executable instructions 34121390.098910\4899-1153-2431.2that, when executed by processor 1402, cause processor 1402 to effectuate operations associated with mapping w ireless signal strength. As evident from the description herein, computing environment 1400 can be not to be construed as software per se.

[0141] In addition to processor 1402 and memory 1404, computing environment 1400 may include an input / output system 1406. Processor 1402, memory 1404, and input / output system 1406 may be coupled together (coupling not shown in FIG. 14) to allow communications therebetween. Each portion of computing environment 1400 may comprise circuitry for performing functions associated with each respective portion. Thus, each portion may comprise hardware, or a combination of hardware and software. Accordingly, each portion of computing environment 1400 can be not to be construed as software per se. Input / output system 1406 may be capable of receiving or providing information from or to a communications device or other network entities configured for telecommunications. For example, input / output system 1406 may include a wireless communication (example, Wi-Fi / 2.6G / 3G / 4G / 5G / GPS) card. Input / output system 1406 may be capable of receiving or sending video information, audio information, control information, image information, data, or any combination thereof. Input / output system 1406 may be capable of transferring information with computing environment 1400. In various configurations, input / output system 1406 may receive or provide information via any appropriate means, such as, for example, optical means (example, infrared), electromagnetic means (example, radio frequency (RF), Wi-Fi, Bluetooth®. ZigBee®), acoustic means (example, speaker, microphone, ultrasonic receiver, ultrasonic transmitter), or a combination thereof. In an example configuration, input / output system 1406 may comprise a Wi-Fi finder, a two-way global positioning system (GPS) chipset or equivalent, or the like, or a combination thereof.

[0142] Input / output system 1406 of computing environment 1400 also may contain communication connection 1408 that allows computing environment 1400 to communicate with other devices, network entities, or the like. Communication connection 1408 may comprise communication media. Communication media typically embody computer-readable instructions, data structures, program modules or other data in a modulated data signal such as a carrier wave or other transport mechanism and includes any information delivery media. By way of example, and not limitation, communication media may include wired media such as a wired network or direct-wired connection, or wireless media such as acoustic, RF, infrared, or other wireless media. The term computer-readable media as used herein includes both storage media and communication media. Input / output system 1406 also may include an input device 1410 such as keyboard, mouse, pen, voice input device, or 35121390.098910\4899-1153-2431.2touch input device. Input / output system 1406 may also include an output device 1412, such as a display, speakers, or a printer.

[0143] Processor 1402 may be capable of performing functions associated with telecommunications, such as functions for processing broadcast messages, as described herein. For example, processor 1402 may be capable of, in conjunction with any other portion of computing environment 1400, determining a type of broadcast message and acting according to the broadcast message type or content, as described herein.

[0144] Memory 1404 of computing environment 1400 may comprise a storage medium having a concrete, tangible, physical structure. As can be known, a signal does not have a concrete, tangible, physical structure. Memory' 1404, as well as any computer-readable storage medium described herein, can be not to be construed as a signal. Memory 1404, as well as any computer-readable storage medium described herein, can be not to be construed as a transient signal. Memory' 1404, as well as any computer-readable storage medium described herein, can be not to be construed as a propagating signal. Memory 1404, as well as any computer-readable storage medium described herein, can be construed as an article of manufacture.

[0145] Memory 1404 may store any information utilized in conjunction with telecommunications. Depending upon the exact configuration or type of processor, memory 1404 may include a volatile storage 1414 (such as some types of random-access memory), a nonvolatile storage 1416 (such as ROM. flash memory), or a combination thereof. Memory 1404 may include additional storage (example, a removable storage 1418 or a nonremovable storage 1420) including, for example, tape, flash memory, smart cards, CD-ROM, DVD, or other optical storage, magnetic cassettes, magnetic tape, magnetic disk storage or other magnetic storage devices, USB-compatible memory, or any other medium that can be used to store information and that can be accessed by computing environment 1400. Memory 1404 may comprise executable instructions that, when executed by processor 1402, cause processor 1402 to effectuate operations to map signal strengths in an area of interest.

[0146] FIG. 15 illustrates a framework 1500 employed by the system of FIG. 11 for device management, dashboard display, and patient monitoring and management. In some examples, framework 1500 may be hosted remotely (e.g., within or in communication with the OFCU 1), as discussed herein. In other examples, the framework 1500 may reside within a device associated with an oxygen delivery system, as discussed herein. As illustrated in FIG. 6. the framework 1500 includes machine learning model 1510 and training data 1520.36121390.098910\4899-1153-2431.2

[0147] At FIG. 15, framework 1500 can be associated with machine learning and / or artificial intelligence (Al). At FIG. 6. the machine learning model 1510 may be operably coupled to the training data 1520 via data interface or data store. In some examples, training data 1520 may be stored training data in a training database (e.g., a data store that can be periodically synchronized with user device, and environmental data collected by the oxygen delivery system). In some examples, the machine learning model 1510 may be associated with other operations. The machine learning model 1510 may be one or more machine learning models configured to perform one or more operations disclosed herein.

[0148] In some examples, the training data 1520 may include, without limitation: (i) physiological time-series associated with a user (e.g., blood oxygen saturation (SpO₂), pulse rate, respiratory rate, inspiratory effort metrics, and / or perfusion index); (ii) device telemetry (e.g., commanded and measured oxygen flow rate, tank or line pressure, battery state-of-charge, bypass events, error / fault codes, and firmware version); (iii) context data, including environmental attributes (e.g., barometric pressure / altitude, temperature, humidity, and air quality index (AQI)) and activity indicators (e g., accelerometer-derived motion states); and (iv) clinical labels or outcomes (e.g., desaturation events, time-in-range, hypoxic episodes, exacerbation / stability indications, adherence summaries). Training data 1520 may be fixed, updated periodically, or updated in near real-time based on evaluations performed by the machine learning model 1510 during a non-training or inference mode, as illustrated by the double-headed arrow connecting the model 1510 and training data 1520.

[0149] In some examples, training data 1520 may include data collected from interactions between one or more oxygen delivery devices (e.g., OFCU 1-equipped systems with sensors such as SpOa, flow, pressure, and accelerometry) and a server configured to provide a computing environment to facilitate device management and patient monitoring (e.g., the cloud / Al environment of FIG. 11). In other examples, a server may simulate operating scenarios (e.g., workload or environmental changes) to generate synthetic or augmented records for model development. Training data 1520 may include interaction records for dashboard usage (e.g., clinician or caregiver navigation events), contextual attributes (e.g., device type, configuration, and firmware), and origination information (e.g., coarse geolocation for altitude inference). In some examples, training data 1520 may use labels that indicate one or more of: device health status (normal, warning, fault, watchdog-bypass), adherence status, and patient state (e.g., at-rest, active, deteriorating).

[0150] In some examples, a server may process training data 1520 by normalizing features (e.g., resampling and unit scaling across devices), encoding categorical attributes 37121390.098910\4899-1153-2431.2(e.g., device configurations), balancing classes (e.g., rare fault events and infrequent desaturations), and enforcing privacy controls (e.g., de-identification or on-device aggregation prior to upload). In some examples, the server may maintain a secondary dataset to evaluate model drift or to perform offline A / B testing of proposed control policies; such a dataset may include synthetic or simulated sequences to safely stress-test oxygen-flow recommendations before clinical deployment. The machine learning model 1510 may, based on training data 1520, generate outputs that include, for example: (i) device-management actions (e.g., initiate firmware update, adjust alarm thresholds, pre-emptively schedule service based on fault likelihood); (ii) dashboard signals (e.g., time-in-SpO₂ target range, predicted time-to-tank-depletion, adherence summaries, and alerts ranked by urgency); and (iii) patient-management recommendations (e.g., anticipatory oxygen-flow adjustments during activity modes, or notifications to a care team for early intervention).

[0151] Typically, such determinations by some existing systems may require a large quantity of manual annotation(s) and / or brute force computer-based annotation to obtain the training data in a supervised training framework. However, example aspects of the present disclosure may deploy a machine learning model(s) (e.g., machine learning model 1510) that may be flexible, adaptive, automated, temporal, learns quickly and trainable. Manual operations or brute force device operations may be unnecessary for the examples of the present disclosure due to the learning framework aspects of the present disclosure that are implementable by the machine learning model 1510. As such, this enables one or more user inputs, notices, or other aspects of the examples of the present disclosure to be flexible and scalable to billions of users, and their associated communication devices, on a network device. Machine Learning (ML), Neural Network (NN), Al, and LLM are generally used interchangeably herein.38121390.098910\4899-1153-2431.2

Claims

CLAIMS1. An oxygen delivery system configured to provide a closed-loop oxygen supply to a user, the system comprising:an oxygen source;an electronically controllable oxygen-flow valve;one or more sensors configured to detect physiological conditions associated with the subject and environmental conditions associated with the user;a memory storing a plurality of context-trim entries, wherein each of the plurality of context-trim entries associates a stored context with a corresponding trim value, wherein each of the stored contexts includes at least one physiological parameter and at least one environmental parameter; anda controller comprising at least one processor communicatively coupled to the one or more sensors, the memory, and the electronically controllable oxygen-flow valve, the controller being configured to:receive, from the one or more sensors, data indicating the physiological conditions and the environmental conditions associated with the user;determine a current context based on the received data;determine, based on the plurality of context-trim entries, a first trim value corresponding to a stored context that matches the current context;determine an oxygen-flow command based on the received data and the first trim value; andactuate the electronically controllable oxygen-flow valve based on the oxygen-flow command, thereby supplying oxygen to the user.

2. The oxygen delivery system of claim 1, wherein determining the oxygen-flow command comprises executing a proportional-integral-derivative (PID) control algorithm that computes proportional, integral, and derivative terms based at least on a deviation between measured and target blood-oxygen saturation and a rate of change of blood-oxygen saturation.

3. The oxygen delivery system of claim 1, wherein the controller is further configured to operate in a plurality of modes, wherein the plurality modes comprise at least one of a constant flow mode, a titration mode, a titration and breath mode, a titration-and-acti vity39121390.098910\4899-1153-2431.2mode, and a titration-breath-activity mode, and wherein, in each of the plurality of modes, the controller is configured to determine the oxygen-flow command based on a different combination of physiological conditions and environmental conditions.

4. The oxygen delivery system of claim 1, further comprising:a dual-lumen cannula including a flow lumen, configured to deliver the oxygen to the subject, and a sense lumen; anda pressure sensor coupled to the sense lumen, wherein the controller is further configured to detect at least one inhalation associated with the subject and at least one exhalation associated with the user based on a pressure signal obtained via the sense lumen and to modify the oxygen-flow command based on the detected at least one inhalation and the detected at least one exhalation.

5. The oxygen delivery system of claim 4, wherein the controller is further configured to control a rate at which the electronically controllable oxygen-flow valve is opened in response to a strength associated with the at least one inhalation, wherein the rate is increased for stronger inhalations and decreased for weaker inhalations.

6. The oxygen delivery system of claim 1, wherein the oxygen source comprises a tank filled with oxygen, the oxygen delivery system further comprising:a flow sensor configured to measure oxygen flow delivered to the user; and a pressure sensor configured to measure pressure in the tank, wherein the controller is configured to determine a size of the tank and a remaining quantity of oxygen in the tank based on correlating an outlet flow rate, measured by the flow sensor, with a rate of tank-pressure change, determined based on readings of the pressure sensor, over time.

7. The oxygen delivery system of claim 6, wherein the controller is further configured to cause display, via at least one user interface, of an indication of the remaining quantity of oxygen in the tank.

8. The oxygen delivery system of claim 1, further comprising:a unified control knob that is adjustable by a user to select either one of a plurality of continuous oxygen flow settings or an automatic electronic mode in which delivery of the oxygen is controlled by the oxygen delivery system.40121390.098910\4899-1153-2431.

29. The oxygen delivery system of claim 8, further comprising:a watchdog safety circuit configured to monitor the controller and force the oxygen delivery system to deliver the oxygen via a mechanical bypass path based on determining a failure associated with the controller, wherein the unified control knob is configured to display an indication that the mechanical bypass path is active.

10. An oxygen flow control unit (OFCU 1) configured to provide a closed-loop oxygen supply to a subject, the OFCU 1 comprising:an electronically controllable oxygen-flow valve;a memory storing a plurality of context-trim entries, wherein each of the plurality of context-trim entries associates a stored context with a corresponding trim value, wherein each of the stored contexts includes a at least one physiological parameter and at least one environmental parameter; anda controller comprising at least one processor communicatively coupled to one or more sensors, the memory, and the electronically controllable oxygen-flow valve, the controller being configured to:receive, from the one or more sensors, data indicating physiological conditions associated with the user and environmental conditions associated with the user; determine a current context based on the received data;determine, based on the plurality of context-trim entries, a first trim value corresponding to a stored context that matches the current context;determine an oxygen-flow command based on the received data and the first trim value; andactuate the electronically controllable oxygen-flow valve based on the oxygen-flow command, thereby causing oxygen to be supplied to the user.

11. The OFCU 1 of claim 10, wherein determining the oxygen-flow command comprises executing a proportional-integral-derivative (PID) control algorithm that computes proportional, integral, and derivative terms based at least on a deviation between measured and target blood-oxygen saturation and a rate of change of blood-oxygen saturation.

12. The OFCU 1 of claim 10, wherein the controller is further configured to operate in a plurality of modes, wherein the plurality of modes comprise at least one of a constant flow 41121390.098910\4899-1153-2431.2mode, a titration mode, a titration and breath mode, a titration-and-activity mode, and a titration-breath-activity mode, and wherein, in each of the plurality of modes, the controller is configured to determine the oxygen-flow command based on a different combination of physiological conditions and environmental conditions.

13. The OFCU 1 of claim 10, -wherein the controller is further configured to:detect at least one inhalation associated with the user and at least one exhalation associated with the user based on a pressure signal obtained via a sense lumen of a dual-lumen cannula; andmodify the oxygen-flow command based on the detected at least one inhalation and the detected at least one exhalation.

14. The OFCU 1 of claim 13, wherein the controller is further configured to control a rate at which the electronically controllable oxygen-flow valve is opened in response to a strength associated with the at least one inhalation, wherein the rate is increased for stronger inhalations and decreased for weaker inhalations.

15. The OFCU 1 of claim 10, further comprising:a flow sensor configured to measure oxygen flow delivered to the user, wherein the controller is configured to determine a size of a tank filled with oxygen and a remaining quantity of oxygen in the tank based on correlating an outlet flow rate, measured by the flow sensor, with a rate of tank-pressure change, determined based on readings of a pressure sensor, over time.

16. The OFCU 1 of claim 15, wherein the controller is further configured to cause display, via at least one user interface, of an indication of the remaining quantity of oxygen in the tank.

17. The OFCU 1 of claim 10, further comprising:a unified control knob that is adjustable by a user to select either one of a plurality of continuous oxygen flow settings or an automatic electronic mode in which delivery of the oxygen is controlled by the OFCU 1.

18. The OFCU 1 of claim 17, further comprising:42121390.098910\4899-1153-2431.2a watchdog safety circuit configured to monitor the controller and force the oxygen to be delivered via a mechanical bypass path based on determining a failure associated with the controller, wherein the unified control knob is configured to display an indication that the mechanical bypass path is active.

19. A computer-readable medium storing instructions that, when executed, cause:receiving, from one or more sensors, data indicating physiological conditions and environmental conditions associated with a user;determining, based on the received data, a current context;determining, based on a plurality of context-trim entries, a first trim value corresponding to a stored context that matches the current context, wherein each of the plurality of context-trim entries associates a stored context with a corresponding trim value, wherein each of the stored contexts includes at least one physiological parameter and at least one environmental parameter;determining an oxygen-flow command based on the received data and the first trim value; andactuating an electronically controllable oxygen-flow valve based on the oxygen-flow command, thereby supplying oxygen to the user.

20. The computer-readable medium of claim 19, wherein determining the oxygen-flow command comprises executing a proportional-integral-derivative (PID) control algorithm that computes proportional, integral, and derivative terms based at least on a deviation between measured and target blood-oxygen saturation and a rate of change of blood-oxygen saturation.43121390.098910\4899-1153-2431.2

Images