Stable time detection method and apparatus for wearable physiological monitors

Abstract

A physiological monitor has an auxiliary sensor and a processor, the processor configured to read the auxiliary sensor and determine when the monitor is in a stable position on a wearer. Alternatively, analog circuitry is provided to determine when the monitor is in a stable position on the wearer. When in an unstable position on the wearer, the monitor enters a low power state. Once in a stable position on the wearer, the processor goes to a full power state and detects physiological events with a primary sensor and logs these events with timestamps. A method is disclosed of determining when the monitor is in a stable position on a subject by reading signals from an auxiliary sensor and classifying the signals from the auxiliary sensor with a classifier selected from the group consisting of a decision tree classifier, a KNN classifier, and a neural network classifier.

Description

Stable Time Detection Method and Apparatus for Wearable Physiological MonitorsCross-Reference to Related Applications

[0001] This application claims the benefit of U.S. Provisional Application No. 63 / 528,319, filed on July 21, 2023, the disclosure of which is incorporated herein by reference in its entirety.Background

[0002] Many people suffer from breathing disorders such as asthma, chronic obstructive lung disease, congestive heart failure, pneumonia, and others. These diseases often have symptoms that are variable, changing with diet, exercise or other activities, allergen exposure, and time of day.

[0003] Many breathing disorders are treated with a combination of short-acting medications such as bronchodilators and longer-acting medications like anti-inflammatory steroids or anti-IgE antibodies. Since patients differ and symptoms can be highly variable, management of those with breathing disorders often requires adjustment of medication types and dosage according to each individual patient’s symptomology to minimize episodes of severe symptoms without unduly interfering with other bodily functions like growth, body weight, or immune function. It is also desirable to identify when exacerbations (episodes of more severe symptoms and possible superinfection) begin because exacerbations may require extra dosing and additional medication of different types.

[0004] In order to properly manage patients with breathing disorders and some other physiological problems, physicians need to know details of how each patient’s symptoms vary over time. For example, an asthma patient who has recently had many severe episodes of difficulty breathing or “asthma attacks” may need more or different longer-acting inhaled steroids, oral steroids, or injected antibodies, to improve their experience — or may need to reduce exposure to triggering antigens or other stimuli; similarly, one who has had few or “attacks” may be able to reduce dosages to reduce side effects of their medications.

[0005] Unfortunately, patient interviews are often inadequate for medication adjustment because they are based on subjective experience and memory. While objective data gathering is often performed in a physician’s office, this is also inadequate for medication adjustment because it gives a single snapshot in time — while breathing issues areoften variable. Accuracy of medication adjustment can be improved by monitoring symptoms over extended times and during ordinary activities of each patient.

[0006] There are physiological monitor devices that can be worn by patients that will record their symptoms throughout a day so physicians can determine whether they need adjustments to medications and dosages both for breathing and some other physiological issues like cardiac arrhythmias. Most such devices must, however, be worn tightly strapped or taped to the patient since distance between device and patient corrupts readings; some physiological monitors like those that sense blood sugar even require an intradermal or subcutaneous sensor. Further, even when devices remain in contact with a patient, signals those devices record may vary with their exact position on the patient, so devices loosely worn as pendants may record acceptable data when they happen to be over a lung but fail to do so when shifted a few inches to be positioned over a subject’s heart.Summary

[0007] We introduce a wearable physiological monitor device with a stable-time detection unit (see stable-time detection unit 132 or 142 of FIG. 1), that recognizes when the physiological monitor is in a stable position on a patient and positioned properly to record good monitor signals, and which may also identify when the monitor is in an unstable position on or near the patient that provides erratic or unacceptable data. In a particular embodiment the physiological monitor is a breathing monitor device that detects wheezes and respiratory rate, and in some variations heart sounds, when in a stable position, while detecting coughs through a microphone in both stable and unstable positions. In another embodiment the physiological monitor device is a cardiac monitor.

[0008] A device using this stable-time detection unit may be worn, for example, loosely as a pendant instead of being tightly strapped or taped to a patient to allow for better patient comfort and encourage good patient compliance while still providing physicians with useful objective data regarding the patient’s symptoms.

[0009] In another embodiment, a physiological monitor has an auxiliary sensor and a processor, the processor configured to read the auxiliary sensor and determine when the monitor is in a stable position on a wearer. The processor detects physiological events with a primary sensor and logs these events with timestamps. A method is disclosed of determining when the monitor is in a stable position on a subject by reading signals from an auxiliary and classifying the signals from the auxiliary sensor with a classifier selected from the groupconsisting of a decision tree classifier, a k-nearest neighbors (KNN) classifier, and a neural network classifier.Brief Description of the Drawings

[0010] Fig. 1 is a block diagram of a physiological monitor, in an embodiment adapted to monitor breathing, according to the present design.

[0011] Fig. 2 is a flow diagram of detecting stable-time periods for the physiological monitor of Fig. 1.

[0012] Fig. 3 is a flowchart of a method of data acquisition performed by firmware of the physiological monitor of Fig. 1

[0013] Fig. 4 is a set of waveforms captured in an experiment with a breathing monitor having a stable-time detection circuit.

[0014] Fig. 5 is a block diagram of an analog stable-time detection unit having multiple bandpass filters, each feeding through magnitude detectors (not shown) to comparators.

[0015] Fig. 6 is a block diagram of an analog alternative stable-time detection unit having multiple bandpass filters each feeding through magnitude detectors (not shown) to comparators.

[0016] Fig. 7 is a block diagram of another analog alternative stable-time detection unit having a peak detector and a timer.Detailed Description of the Embodiments

[0017] For purposes of this document, a low-power state of a processor is either a reduced-clock-state that reduces power consumption below that of a full-speed, full-power, state of the same processor, or a standby or sleep state where the processor essentially is turned off and draws less power than when the processor is in a full-speed, full-power, state. Further, analog circuitry is circuitry where signal values are represented by a voltage, current, or charge quantity and not as a number in a processor; analog filters are typically implemented with inductors, capacitors, amplifiers, analog multipliers, and resistors and may include switched-capacitor circuitry; while digital filters are typically implemented in a processor using multiply and add instructions on a sequence of numbers representing signals.

[0018] An exemplary physiological monitor device 100, particularly tailored for monitoring breathing, has acoustic sensors 101, such as microphones, configured to pick up breathing and cardiac sounds from a chest of a wearer; the microphones are also adapted topick up some breathing sounds, like coughs, from the environment. Signals from the acoustic sensors are read by a processor 102, also known as a microcontroller unit (MCU) executing a breathing monitor firmware 104 in memory 106 that is configured to identify breath sounds of interest, such as coughs and wheezes, in breath sounds as they are read from the acoustic sensors 101. When coughs are heard by the acoustic sensors, these are classified, and the event recorded with an event time in an event log 110 in memory 106. When the stable- time detection unit 132 or 142 determines the physiological monitor device 100 is positioned on the wearer in a position that will give good data and wheezes are heard by the acoustic sensors, the wheezes are classified, and the event of wheezing is recorded in the event log 110. When detecting coughs and wheezes, most embodiments also monitor respiratory rate from the breath sounds. When the stable-time detection unit 132 or 142 determines the physiological monitor device 100 is positioned on the wearer in a position that provides good data and heart sounds of interest, such as erratic, tachycardic, or bradycardic heartbeat sounds, are detected, processor 102 classifies the heart sounds and records occurrence of the heart sounds of interest in the event log 110 in memory 106 together with an event time read from a real-time clock 108.

[0019] The physiological monitor device may be adapted to monitor other physiological attributes of a subject instead of or in addition to breathing when the stabletime detection unit 132 or 142 determines that the physiological monitor device is in a stable position on the subject.

[0020] The physiological monitor device 100 may also include a power supply 112 that may include a rechargeable battery with a charger 114, in some embodiments the charger draws power from a USB -compatible serial port and in other embodiments the charger receives power inductively from a nearby electromagnetic charging station.

[0021] In some, but not all, embodiments, the stable-time detection unit 132 or 142 may also use a photoplethysmographic sensor 120 that includes a plurality of light emitting diodes and at least one photodiode (not shown in Fig. 1) for determining stable periods suitable for accurate measurements to be taken and, in an embodiment, data from the photoplethysmographic sensor 120 may be processed to record physiological data.

[0022] In order to make physiological data gathered by the physiological monitor device 100 available to a physician or other post-event analysis, monitor device 100 also has either or both of a digital radio 122 or a USB-compatible serial communications port 124. In embodiments 100 with a digital radio 122, the short range digital radio is compliant with Bluetooth, Zigbee, or another standard suitable for communicating with a base station, cellphone, body area network hub, or other device adaptable for compiling data recorded by monitor device 100 and forwarding that data over the Internet to a server.

[0023] The physiological monitor device 100 is configured for wear as a pendant. As such, monitor device 100 may flop around when a wearer is moving about and may generate good data at some times, and bad data at other times, the stable-time detection unit 132 or 142 is adapted to determine when the physiological monitor device is in a stable position on the wearer. To help the stable-time detection unit determine when the physiological monitor is in a stable position on the wearer, the monitor device 100 may have one or more auxiliary sensors. Auxiliary sensors that may be included in the physiological monitor include accelerometers 126, which may or may not include gyros, and other auxiliary sensors 128 that may include one or more of an infrared optical proximity sensor, a capacitive skin-contact sensor, a temperature sensor, a photoplethysmographic sensor with optical range capability, an electromyographic sensor, an electrocardiographic sensor, and a microphone adapted to record slapping sounds of impact of the monitor device 100 with the wearer’s skin. The accelerometers 126 and / or other auxiliary sensors 128 are coupled to be read over one or more intra-device communications busses 130 to the processor 102, which processes data from these sensors according to a stable-time detection firmware to serve as a stable-time detection unit 132 or 142.

[0024] In some embodiments, stable- time detection unit 132 or 142 is implemented as stable-time detection firmware 132 that executes on processor 102, in these embodiments processor 102 may conserve power by slowing its processor clock 140 while the monitor device 100 is not in a stable position on the wearer while still executing the stable-time detection firmware because the stable-time detection firmware is much simpler than the full event-detection and analysis routines executed when the monitor device is in a stable position on the wearer. In alternative embodiments, stable-time detection unit is an analog stable-time detection unit 142 that requires no intervention by the processor to detect when the monitor device is in a stable position on a wearer, and which may control processor clock 140 to place the monitor device in a low-power state or wake the processor from the low-power state.

[0025] Stable-time detection firmware operates in real time whenever monitor device 100 is being worn according to the method 200 of the flow diagram of Fig. 2.

[0026] If accelerometers 126 with or without gyros are present in monitor device 100 as a first auxiliary sensor, they are read to acquire 202 their signal, the signal is processed to extract 204 parameters that may include a rate-of-change of acceleration, X-Y - and Z- axiscomponents of acceleration where accelerometers 126 are three-axis accelerometers, and magnitude of acceleration. If accelerometers 126 include gyros, extracted parameters may include rates of rotation of monitor device 100.

[0027] If acoustic sensors like micromachined microphones are provided in auxiliary sensors 128 in monitor device 100, their signals are acquired 206 by processor 102 and parameters extracted 208 such as presence of impulse noises due to bumping against skin. In embodiments having such sensors, absence of bumping against skin may indicate “stable-time” other indications of skin contact are present, but presence of impulse noises due to bumping against skin implies the monitor device is not in “stable-time”. A micromachined microphone may do double-duty as both an auxiliary sensor for stable-time determination and as a microphone for recording breath sounds that can be analyzed for wheezes, coughs, and similar events.

[0028] If a capacitive skin-contact sensor is included in auxiliary sensors 128 of monitor device 100, its signal is acquired 210 by processor 102 and parameters 212 extracted such as time since last contact or break from contact with skin. Extended periods of skin contact may indicate “stable-time”, but frequent intervals of contact with skin interspersed with intervals of no skin contact may indicate the monitor device is not in “stable-time”.

[0029] If an infrared proximity sensor is included in auxiliary sensors 128 of monitor device 100, its signal is acquired 214 and parameters extracted 216 such as distance to skin and rate of change of distance to skin. Extended periods with close proximity to skin and few changes in distance to skin are indicative of “stable-time”, but frequent changes in distance to or contact with skin may indicate the monitor device is not in “stable-time”. We note that the LEDs and photodiodes of a photoplethysmographic sensor may do double duty as the LED and photodiode of a proximity sensor.

[0030] If a temperature sensor is included as an auxiliary sensor, widely varying temperatures or temperatures far from a normal body temperature may indicate the monitor device is not in “stable-time”. Similarly, if an electromyographic sensor is included as an auxiliary sensor, high muscle activity detected by the sensor, or frequent breaks in electromyographic signals as its electrodes bounce away from a subject, may indicate the monitor device is not in “stable-time”. Similarly, if an electrocardiographic sensor is included as an auxiliary sensor, frequent weak signals or breaks in signals may indicate the monitor device is not in “stable-time”.

[0031] The extracted parameters from the accelerometers and all other auxiliary sensors equipped in a particular embodiment are examined to determine 218 a compositescore or likelihood that the monitor device 100 is stable in position adjacent skin of the wearer, and that acoustic and / or photoplethysmographic data may be useful in determining heart rate and detecting arrhythmias. In embodiments, more than one auxiliary sensor are considered while determining stable-time. In a particular embodiment, the composite score is determined by comparing extracted parameters to limits; in another embodiment the extracted parameters are fed to a classifier selected from the group consisting of a support vector machines (SVM) classifier, a K-nearest-neighbors (KNN) classifier, neural network classifiers that may be recurrent neural networks, or a decision tree classifier. The score is used to set a stable / not stable position flag for example.

[0032] Meanwhile, when the monitor device 100 is being worn, processor 102, performs method 300 (Fig. 3) by continually (or at intervals) checking the stable / not stable flag. 302. If 304 the flag indicates the device 100 is not in a stable position on the wearer the processor in some embodiments sets its clock to a low speed 306 to conserve power; it then samples audio 308 from the primary acoustic sensors 101, which may be contact microphones or micromachined microphones, at a low rate and determines presence of loud signals. These loud signals are processed to recognize 310 coughs. If 312 coughs are present, coughs are classified and recorded 314 with a timestamp in event log 110, and the processor repeats 316 checking the stable / not stable flag. If 312, coughs are not present, the processor repeats 316 checking the stable / not stable flag.

[0033] In some embodiments, instead of setting the processor clock to a low speed, the processor enters a hibernation or sleep state when the flag indicates the monitor device 100 is not in a stable position on the wearer.

[0034] In the embodiments of Figs. 5, 6, and 7, analog circuitry is provided to detect stable-times, with the processor entering a low power state when the monitor is not in a stable position on the wearer; the low power state may be a low-clock-rate state or a “sleep” or “hibernate” state where processor status is saved and the processor is shut down.

[0035] During times the processor is in the low-power state, in the embodiment of Fig. 5 and Fig. 6 raw sensor data is passed through a series of analog filters, each tuned to a specific frequency band,. These filters isolate different frequency components of the sensor signals by breaking the signal down into its frequency bands, allowing for better characterization of physiological signs.

[0036] In the embodiment of Fig. 5, a circuit implementation of a classification model — such as a series of comparators designed to simulate a decision tree or an SVM classifier — is then used to analyze the energy levels within these frequency bins. This helpsdetect periods of stable-time, which may be indicated by particular patterns of magnitudes in the frequency bands, thereby setting the stable- time flag and triggering the processor to wake up. In this embodiment, both the filter bank and the classifier are implemented in a dedicated hardware with an FPGA and / or ASIC, and thus, the stable-time detection is managed without waking the main processor, allowing it to remain in hibernation to conserve battery life.

[0037] In some embodiments, the raw sensor data passed into the analog filers is from the primary sensor of the physiological monitor, in other embodiments the raw sensor data is from an auxiliary sensor.

[0038] In the embodiment of Fig. 6, a plurality of analog filters each feed through magnitude detectors to a single set of comparators, a single pattern of frequencies in the raw sensor data is detected and used to set the stable-time flag, thereby triggering the processor to wake up or transition to full-speed mode.

[0039] In another implementation, the stable-time detection unit (STDU) includes a tunable analog peak detector in conjunction with a timing circuit and a comparator as illustrated in Fig. 7. The peak detector is dynamically tuned to detect characteristic peaks of physiological signals, such as from heartbeats or breathing cycles. Then the timing circuitry measures the intervals between peaks to estimate the frequency of the signal, be it heart rate or respiratory rate. When the real-time estimated frequency falls within the nominal range, that is indicative of stable-time and therefore a positive stable-time flag is triggered to wake up the main device processor.

[0040] In both examples, the STDU is implemented as a dedicated hardware and the main processor is only woken from hibernation when a stable-time is detected, ensuring efficient power management, and reducing processor computation.

[0041] If, when checking the stable / not stable flag 302, the flag indicates the device 100 is in a stable position on the wearer, the processor 102 in some embodiments sets 320 its clock to a high speed to perform the digital filtering required to detect coughs and wheezes and in some embodiments heartbeat sounds, which involves considerably more computation to use microphone data sampled at a higher rate as compared to cough detection alone. The processor then samples 322 audio from the primary acoustic sensors 101 at a high rate; in a particular embodiment the high rate is 8 kilosamples per second although other embodiments may use sampling rates of 12, 16, 24, 32, or 44.1 kilosamples per second. The high sample rate audio is subsampled and cough detection 324 is performed. The high sample rate audio is digitally filtered in a plurality of frequency bands and wheezes are detected 326 by a classifier as previously described. Similarly, a respiratory rate is determined from thesampled audio. The high sample rate audio is also digitally filtered to detect and recognize 328 sounds associated with a particular phase, such as Tl, of the heartbeat; times of detecting these sounds are recorded in a temporary memory in memory 106. Times of detected sounds are then processed to measure 330 heartbeat periods and heart rate. Determined heart rates are processed to recognize 332 periods of arrhythmia, bradycardia, and tachycardia. If 334, no wheeze, arrhythmia, bradycardia, or tachycardia was detected, the processor 102 checks 312 to see if coughs were present, recording 314 any coughs in the event log 110, and repeats 316 checking the stable / not stable flag. If 334 wheezes, arrhythmia, bradycardia, or tachycardia is present, the processor attaches a timestamp and records 336 these events in event log 110, then checks 312 to see if coughs were present, recording 314 any coughs in the event log 110, and repeats 316 checking the stable / not stable flag.

[0042] In embodiments having as auxiliary sensors an infrared proximity detector as a subset of a photoplethysmographic sensor, an electrocardiographic sensor, or a temperature sensor, data obtained from such sensors may also be added to the event log. Similarly, accelerometers with or without gyros may optionally provide activity data, such as pedometer data, that may be processed to provide a step count that may also be added to the event log.

[0043] The physiological monitor device 100 is just one embodiment of a device that may use the stable-time detection unit 132 or 142. We anticipate a physiological monitor device 100 is a breathing monitor in some embodiments, and a heart arrhythmia monitor, an atrial fibrillation monitor in another embodiment, an exercise monitor in other embodiments, or a sleep apnea monitor in yet another embodiment; additional embodiments may become practical in future such as noninvasive glucose monitor devices. In embodiments having a photoplethysmographic sensor, the stable-time detection unit 132 or 142 may use this sensor when the stable / not stable flag indicates the device 100 is in a stable position on the wearer to determine times of peak blood oxygenation, which are then processed as for Tl times detected from the acoustic sensor processing described above by detecting 330 periods and rates. In these embodiments, episodes of low, excessive, or erratic heart rates, or low blood oxygenation as derived from the photoplethysmographic sensor, such as may occur with sleep apnea, are recorded with timestamps in event log 110.

[0044] Once data is gathered in event log 110, the processor 110 may receive a request for that data through either the digital radio 122 or USB port 124, whereupon the processor sends data from the event log 110 to a host machine through the digital radio orUSB port as appropriate. This data is then usable by medical personnel or as input data to health monitor devices software.

[0045] When physiological monitor device 100 is a breathing monitor device, while heart function data and wheeze data may have gaps, although cough data may be continuous, these gaps may not be significant for treating many wearers particularly because asthma “attacks,” may cause a wearer to cease activities long enough for the physiological monitor to be detected as in a stable position on the wearer. Asthma attacks are often marked by coughs and wheezing, and may be accompanied by a decrease in oxygen saturation as measured by a photoplethysmographic sensor. Similarly, if a physiological monitor is a heart monitor, episodes of atrial fibrillation may cause a wearer to cease or reduce activities long enough for the reduced activities to be detected, the stable / not stable flag be set to stable, and the events to be detected and logged. Further, many such events last long enough that the stable / not stable flag will enter the stable state for at least a portion of the event even if the wearer continues moderate activity.

[0046] In some embodiments, activity of the wearer is also monitored by recording changes in the stable / not stable flag with an associated timestamp in event log 110. This data may be useful in counseling inactive wearers — as can be observed from excessive time with the stable flag set — to become more active.

[0047] Fig. 4 illustrates data recorded from a particular experimental embodiment. Raw respirations 402 can be derived from breath sounds recorded from a pri mary sensor. A stable-time detection flag 404 is derived from, in this example, signals from auxiliary sensors including accelerometers. Stable-time respiration rates 406 are recorded and available for upload to a physician. Other parameters such as wheezes may also be recognized during stable-times and uploaded appropriately, but are not illustrated in Fig. 4.Combinations

[0048] The various features herein described can be combined into a variety of physiological monitors. Among combinations of these features anticipated by the inventors are:

[0049] A physiological monitor designated A includes at least one primary sensor; a processor with a memory having a low power state and a full performance state; and at least one auxiliary sensor selected from the group consisting of an accelerometer, an optical proximity sensor, a capacitive skin-contact sensor, a temperature sensor, anelectromyographic sensor, an electrocardiographic sensor, and a contact microphone. The processor is equipped with stable-time detection unit selected from group consisting of the processor being configured with firmware in the memory to read the at least one auxiliary sensor and determine when the physiological monitor is in a stable position on a wearer, and analog circuitry configured to read the at least one auxiliary sensor and determine when the physiological monitor is in a stable position on the wearer. The processor is configured to enter the low power state when the stable-time detection unit indicates the processor is not in a stable position on the wearer, and to enter the full performance state when the physiological monitor is in a stable position on the wearer; and the processor is configured by firmware in the memory to read the primary sensor when the physiological monitor is in the full performance state and to detect events of interest, to record a timestamp and classification of events of interest in an event log.

[0050] A physiological monitor designated A2 having at least one primary sensor; a processor with a memory; and at least one auxiliary sensor selected from an accelerometer, an optical proximity sensor, a capacitive skin-contact sensor, a temperature sensor, an electromyographic sensor, an electrocardiographic sensor, and a contact microphone. The processor is configured with firmware in the memory to read the at least one auxiliary sensor and determine when the physiological monitor is in a stable position on a wearer. In an embodiment the processor is configured by firmware in the memory to read the primary sensor when the physiological monitor is in a stable position on the wearer and to detect events of interest, to record a timestamp and classification of events of interest in the event log, and to not record events of interest in the event log when the physiological monitor is not in a stable position on the wearer.

[0051] A physiological monitor designated AA include the physiological monitor designated A wherein the processor is configured by firmware in the memory to observe heart sounds when the physiological monitor is in a stable position on the wearer and detect arrhythmia from those heart sounds, and to record a timestamp and classification of the arrhythmia in the event log.

[0052] A physiological monitor designated AB including the physiological monitor designated A or AA wherein the processor is configured by firmware in the memory to detect bradycardia or tachycardia from the heart sounds and to record a timestamp and classification of the bradycardia or tachycardia in the event log.

[0053] A physiological monitor designated AC including the physiological monitor designated A, AB or AA where the auxiliary sensor includes an accelerometer with or without gyros.

[0054] A physiological monitor designated AD including the physiological monitor designated A, AB, ACor AA where the auxiliary sensor includes a capacitive skincontact sensor.

[0055] A physiological monitor designated AE including the physiological monitor designated A, AA, AB, AC, or AD where the auxiliary sensor includes a contact microphone.

[0056] A physiological monitor designated AF including the physiological monitor designated A, AA, AB, AC, AD, or AE where the auxiliary sensor includes an optical proximity sensor.

[0057] A physiological monitor designated AG including the physiological monitor designated AF where the optical proximity sensor is a portion of a photoplethysmographic sensor.

[0058] A physiological monitor designated AH including the physiological monitor designated A, AA, AB, AC, AD, or AE where the auxiliary sensor includes an electrocardiographic sensor.

[0059] A physiological monitor designated AJ including the physiological monitor designated A, AA, AB, AC, AD, AE, AF, AG, or AH where the auxiliary sensors includes a temperature sensor.

[0060] A physiological monitor designated AK including the physiological monitor designated A, AA, AB, AC, AD, AE, AF, AG, AH, or AJ where the processor is configured by firmware to operate at a low speed or enter a sleep or hibernation state when the physiological monitor is not in a stable position on the wearer, and to operate at a high speed when the physiological monitor is in a stable position on the wearer.

[0061] A physiological monitor designated AL including the physiological monitor designated A, AA, AB, AC, AD, AE, AF, AG, AH, AJ, or AK wherein the processor is further configured by firmware in the memory to observe heart sounds when the physiological monitor is in a stable position on the wearer and detect arrhythmia from those heart sounds, and to record a timestamp and classification of the arrhythmia in the event log.

[0062] A physiological monitor designated AM including the physiological monitor designated A, AA, AB, AC, AD, AE, AF, AG, AH, AJ, AK, or AL wherein the processor is further configured by firmware in the memory to detect bradycardia ortachycardia from the heart sounds and to record a timestamp and classification of the bradycardia or tachycardia in the event log.

[0063] A physiological monitor designated AN including the physiological monitor designated A, AA, AB, AC, AD, AE, AF, AG, AH, AJ, AK, AL, or AM where the primary sensor is a microphone and the physiological monitor is a breathing monitor, and the events of interest include episodes of coughing and episodes of wheezing.

[0064] A physiological monitor designated AO including the physiological monitor designated A, AA, AB, AC, AD, AE, AF, AG, AH, AJ, AK, AL, AM, or AN where the primary sensor comprises a photoplethysmographic sensor and where the events of interest include episodes of low blood oxygenation.

[0065] A physiological monitor designated AP including the physiological monitor designated A, AA, AB, AC, AD, AE, AF, AG, AH, AJ, AK, AL, AM, AN, or AO where the primary sensor comprises both a photoplethysmographic sensor and a microphone.

[0066] A physiological monitor designated AQ including the physiological monitor designated A, A2, AA, AB, AC, AD, AE, AF, AG, AH, AJ, AK, AL, AM, AN, AO, or AP where the processor is configured by firmware to determine when the physiological monitor is in a stable position on a wearer by passing signals from the at least one auxiliary sensor through a classifier selected from the group consisting of a decision tree classifier, a K-nearest neighbors classifier, a support-vector-machines (SVM) classifier, and a neural network classifier.

[0067] A physiological monitor designated AR including the physiological monitor designated A, AA, AB, AC, AD, AE, AF, AG, AH, AJ, AK, AL, AM, AN, AO, AP, or AQ wherein the processor is configured with firmware in the memory to read the at least one auxiliary sensor and determine when the physiological monitor is in a stable position on a wearer.

[0068] A physiological monitor designated AS including the physiological monitor designated A, AA, AB, AC, AD, AE, AF, AG, AH, AJ, AK, AL, AM, AN, AO, AP, or AQ, wherein the stable-time detection unit includes analog circuitry configured to read the at least one auxiliary sensor and determine when the physiological monitor is in a stable position on the wearer.

[0069] A physiological monitor designated AT including the physiological monitor designated AS, wherein the stable-time detection unit includes a plurality of analog filters and a plurality of comparators configured to determine a pattern of frequencies in the auxiliary sensor data.

[0070] A method designated B of determining when a physiological monitor device is in a stable position on a subject comprising reading signals from an auxiliary sensor of the physiological monitor device and classifying the signals from the auxiliary sensor with a classifier selected from the group consisting of a decision tree classifier, a K-nearest- neighbors classifier, a support-vector-machines (SVM) classifier, and a neural network classifier.

[0071] A method designated BA including the method designated B further comprising reading primary sensor data from a primary sensor when the physiological monitor device is in the stable position on the subject, determining if an event of interest occurs from the primary sensor data, and recording the event of interest in an event log with a timestamp.

[0072] Changes may be made in the above methods and systems without departing from the scope hereof. It should thus be noted that the matter contained in the above description or shown in the accompanying drawings should be interpreted as illustrative and not in a limiting sense. The following claims are intended to cover all generic and specific features described herein, as well as all statements of the scope of the present method and system, which, as a matter of language, might be said to fall therebetween.

Claims

ClaimsWhat is claimed is:

1. A physiological monitor comprising : at least one primary sensor; a processor with a memory, the processor having a low power state and a full performance state; at least one auxiliary sensor selected from the group consisting of an accelerometer, an optical proximity sensor, a capacitive skin-contact sensor, a temperature sensor, an electromyographic sensor, an electrocardiographic sensor, and a contact microphone; a stable-time detection unit selected from group consisting of the processor configured with firmware in the memory to read the at least one auxiliary sensor and determine when the physiological monitor is in a stable position on a wearer, and analog circuitry configured to read the at least one auxiliary sensor and determine when the physiological monitor is in a stable position on the wearer; the processor configured to enter the low power state when the stable-time detection unit indicates the processor is not in a stable position on the wearer, and to enter the full performance state when the physiological monitor is in a stable position on the wearer; and the processor configured by firmware in the memory to read the primary sensor when the physiological monitor is in the full performance state and to detect events of interest, to record a timestamp and classification of events of interest in an event log.

2. The physiological monitor of claim 1 wherein the processor is further configured by firmware in the memory to observe heart sounds when the physiological monitor is in a stable position on the wearer and detect arrhythmia from those heart sounds, and to record a timestamp and classification of the detected arrhythmia in the event log.

3. The physiological monitor of claim 2 wherein the processor is further configured by firmware in the memory to detect bradycardia or tachycardia from the heart soundsand to record a timestamp and classification of the bradycardia or tachycardia in the event log.

4. The physiological monitor of claim 1 where the auxiliary sensor comprises an accelerometer.

5. The physiological monitor of claim 4 where the auxiliary sensor further comprises a capacitive skin-contact sensor.

6. The physiological monitor of claim 4 where the auxiliary sensor further comprises a contact microphone.

7. The physiological monitor of claim 4 where the auxiliary sensor further comprises an optical proximity sensor.

8. The physiological monitor of claim 4 where the auxiliary sensor or primary sensors comprises an electrocardiographic sensor.

9. The physiological monitor of claim 4 where the auxiliary sensors includes a temperature sensor.

10. The physiological monitor of claim 1 where the processor is configured by firmware to operate at a low speed or enter a sleep or hibernation state when the physiological monitor is not in a stable position on the wearer, and to operate at a high speed when the physiological monitor is in a stable position on the wearer.

11. The physiological monitor of claim 10, wherein the processor is further configured by firmware in the memory to observe heart sounds when the physiological monitor is in a stable position on the wearer and detect arrhythmia from those heart sounds, and to record a timestamp and classification of the detected arrhythmia in the event log.

12. The physiological monitor of claim 11, wherein the processor is further configured by firmware in the memory to detect bradycardia or tachycardia from the heart sounds and to record a timestamp and classification of the bradycardia or tachycardia in the event log.

13. The physiological monitor of claim 12, where the auxiliary sensor comprises an accelerometer.

14. The physiological monitor of claim 1, where the primary sensor is a microphone and the physiological monitor is a breathing monitor, and the events of interest include episodes of coughing and episodes of wheezing.

15. The physiological monitor of claim 2, where the primary sensor comprises a photoplethysmographic sensor and where the events of interest include episodes of low blood oxygenation.

16. The physiological monitor of claim 1, where the primary sensor comprises both a photoplethysmographic sensor and a microphone.

17. A physiological monitor of claim 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15, or 16 wherein the processor is configured with firmware in the memory to read the at least one auxiliary sensor and determine when the physiological monitor is in a stable position on a wearer.

18. A physiological monitor of claim 1, 2, 3, 4, 5, 6, 7, 8, 9, 10, 11, 12, 13, 14, 15 or 16, wherein the stable-time detection unit is analog circuitry configured to read the at least one auxiliary sensor and determine when the physiological monitor is in a stable position on the wearer.

19. A physiological monitor of claim 18 wherein the stable-time detection unit comprises a plurality of analog filters and a plurality of comparators configured to determine a pattern of frequencies in auxiliary sensor data from the at least one auxiliary sensor.

20. A physiological monitor of claim 18 wherein the stable-time detection unit comprises a peak detector and an interval timing circuit.

21. The physiological monitor of claim 17 where the stable-time detection unit is implemented in firmware on the processor and the processor is configured by firmware to determine when the physiological monitor is in a stable position on a wearer by passing signals from the at least one auxiliary sensor through a classifier selected from the group consisting of a decision tree classifier, a K-nearest neighbors classifier, a support- vector-machines (SVM) classifier, and a neural network classifier.

22. A method of determining when a physiological monitor device is in a stable position on a subject comprising reading signals from an auxiliary sensor of the physiological monitor device and classifying the signals from the auxiliary sensor with a classifier selected from the group consisting of a decision tree classifier, a K-nearest-neighbors classifier, a support- vector-machines (SVM) classifier, and a neural network classifier.

23. The method of claim 22, further comprising reading primary sensor data from a primary sensor when the physiological monitor device is in the stable position on the subject, determining if an event of interest occurs from the primary sensor data, and recording the event of interest in an event log with a timestamp.

24. A method of determining when a physiological monitor device is in a stable position on a subject comprising reading signals from an auxiliary sensor of the physiological monitor device comprises passing the signals through a plurality of analog filters and comparing magnitudes of outputs of the analog filters to values.

Images