Patent Application
20210330209
This disclosure relates to the field of wearable devices, and particularly to techniques for using photoplethysmography (PPG) sensors to generate heart rate (HR) and other physiological metrics.
A PPG sensor may be utilized to detect the volumetric change in blood vessels. A PPG sensor usually includes a light source, typically a light-emitting diode (LED), and a light-sensitive sensor, typically a photodiode. Blood passing through the vasculature between the light source and the sensor will modulate the light path between the two, resulting in a deviation in the current produced by the photodiode. By applying various algorithms to the signal sensed by the photodiode, an HR estimate can be determined.
Further, by looking at signals corresponding to two or more wavelengths (e.g., red and infrared), a pulsatile blood oxygenation estimate (SpO2) can be obtained. SpO2 refers to a fraction of oxygen-saturated hemoglobin relative to total hemoglobin in the blood. Decreased SpO2 in the blood can lead to impaired mental function, or loss of consciousness, and may serve to indicate other serious health conditions, such as sleep apnea or cardiovascular disease. Therefore, accurately measuring SpO2 is important in certain kinds of health monitoring.
Typical PPG technologies rely on emitting wavelengths of green, red, and/or infrared (IR) light from an LED. Many wearable PPG devices use green light, as the hemoglobin absorption of light is up to 20 times greater at green wavelengths than at IR wavelengths. Additionally, in some cases, green LED light sources may provide superior results in terms of cost, form factor, and power efficiency.
The systems, methods, and devices of this disclosure each have several innovative aspects, no single one of which is solely responsible for the desirable attributes disclosed herein.
One aspect of the disclosure provides a method for generating physiological metrics. The method may include obtaining one or more first PPG signals based on light received by a first light detector from a set of light sources, each light source in the set (i) having a spatial location that is different from that of another light source in the set and (ii) configured to emit light according to an emission schedule that is different from that of another light source in the set; obtaining one or more second PPG signals based on light received by a second light detector from the set of light sources; generating a physiological metric based on one or more of the first and second PPG signals; and causing the physiological metric to be displayed via a user interface on a user device.
The method of the preceding paragraph can have any sub-combination of the following features: where the set of light sources comprises at least one green light source, at least one red light source, and at least one infrared light source; where the set of light sources are provided across two or more light source packages, each light source package including at least one light source from the set and located at least 1 mm away from another one of the two or more light source packages; where the set of light sources comprises two or more light sources included in a single light source package, the two or more light sources in the light source package each configured to emit light having a wavelength that is different from another light source in the light source package; where the two or more light sources in the light source package are located within less than 1 mm from each other; where the two or more light sources comprise at least a red light source and an infrared light source; where the method further includes: obtaining a first signal of the first PPG signals based on the light received by the first light detector from a first light source in the set during a first temporal window, obtaining a second signal of the first PPG signals based on the light received by the first light detector from a second light source in the set that is different from the first light source during a second temporal window that is different from the first temporal window, and generating the physiological metric based at least in part on the first signal and the second signal; where the method further includes removing a motion component from at least one of the first and second signals based on location information associated with one or more of the first light source, the second light source, and the first light detector; where the method further includes determining a first confidence metric associated with the first signal, determining a second confidence metric associated with the second signal, and based on a determination that the first confidence metric is higher than the second confidence metric, generating the physiological metric based on the first signal; where the method further includes obtaining a first signal of the first PPG signals based on the light received by the first light detector from a first light source in the set, obtaining a second signal of the second PPG signals based on the light received by the second light detector from the first light source in the set, and generating the physiological metric based at least in part on the first signal and the second signal; where the first and second light sources are spatially located between the first and second light detectors; where the first and second light detectors are spatially located between the first and second light sources; and where the first and second light sources and the first and second light detectors are arranged such that each of the first and second light sources is equidistant from each of the first and second light detectors.
One aspect of the disclosure provides a wearable device for generating physiological metrics. The wearable device may include a plurality of light sources configured to emit light, a plurality of light detectors configured to detect the light emitted by the plurality of light sources, a display, one or more processors, and a memory. The memory may store instructions that that, when executed by the one or more processors, cause the one or more processors to: obtain one or more first PPG signals based on light received by a first light detector of the plurality of light detectors from a set of light sources of the plurality of light sources, each light source in the set (i) having a spatial location that is different from that of another light source in the set and (ii) configured to emit light according to an emission schedule that is different from that of another light source in the set; obtain one or more second PPG signals based on light received by a second light detector of the plurality of light detectors from the set of light sources; generate a physiological metric based on one or more of the first and second PPG signals; and cause the physiological metric to be displayed on the display.
The wearable device of the preceding paragraph can have any sub-combination of the following features: where the set of light sources comprises at least one green light source, at least one red light source, and at least one infrared light source; where the set of light sources are provided across two or more light source packages, each light source package including at least one light source from the set and located at least 1 mm away from another one of the two or more light source packages; where the set of light sources comprises two or more light sources included in a single light source package, the two or more light sources in the light source package each configured to emit light having a wavelength that is different from another light source in the light source package; where the two or more light sources in the light source package are located within less than 1 mm from each other; where the two or more light sources comprise at least a red light source and an infrared light source; where the instructions further cause the one or more processors to obtain a first signal of the first PPG signals based on the light received by the first light detector from a first light source in the set during a first temporal window, obtain a second signal of the first PPG signals based on the light received by the first light detector from a second light source in the set that is different from the first light source during a second temporal window that is different from the first temporal window, and generate the physiological metric based at least in part on the first signal and the second signal; where the instructions further cause the one or more processors to remove a motion component from at least one of the first and second signals based on location information associated with one or more of the first light source, the second light source, and the first light detector; where the instructions further cause the one or more processors to determine a first confidence metric associated with the first signal, determine a second confidence metric associated with the second signal, and based on a determination that the first confidence metric is higher than the second confidence metric, generate the physiological metric based on the first signal; where the instructions further cause the one or more processors to obtain a first signal of the first PPG signals based on the light received by the first light detector from a first light source in the set, obtain a second signal of the second PPG signals based on the light received by the second light detector from the first light source in the set, and generate the physiological metric based at least in part on the first signal and the second signal; where the first and second light sources are spatially located between the first and second light detectors; where the first and second light detectors are spatially located between the first and second light sources; and where the first and second light sources and the first and second light detectors are arranged such that each of the first and second light sources is equidistant from each of the first and second light detectors.
Existing PPG sensors can provide an effective method for measuring a user's HR. As noted above, a PPG sensor usually includes a light source, typically an LED, and a light-sensitive sensor, typically a photodiode. Using the photodiode, a PPG device including such a PPG sensor can obtain PPG signals indicative of how the blood is passing through the vessels and generate HR estimates based on the PPG signals.
Some PPG technologies rely on detecting light at a single spatial location, or adding signals taken from two or more spatial locations. Both of these approaches result in a single spatial measurement from which the HR estimate (or other physiological metrics) can be determined. In some embodiments, a PPG device employs a single light source coupled to a single detector (i.e., a single light path). Alternatively, a PPG device may employ multiple light sources coupled to a single detector or multiple detectors (i.e., two or more light paths). In other embodiments, a PPG device employs multiple detectors coupled to a single light source or multiple light sources (i.e., two or more light paths). In some cases, the light source(s) may be configured to emit one or more of green, red, and/or infrared light. For example, a PPG device may employ a single collimated light source and two or more light detectors each configured to detect a specific wavelength or wavelength range. In some cases, each detector is configured to detect a different wavelength or wavelength range from one another. In other cases, two or more detectors configured to detect the same wavelength or wavelength range. In yet another case, one or more detectors configured to detect a specific wavelength or wavelength range different from one or more other detectors). In embodiments employing multiple light paths, the PPG device may determine an average of the signals resulting from the multiple light paths before determining an HR estimate or other physiological metrics. Such a PPG device may not be able to resolve individual light paths or separately utilize the individual signals resulting from the multiple light paths.
In some cases, if the user wearing the PPG device is performing an activity involving motion (or contorting of the wrist, for example, for a wrist-worn PPG device, thereby affecting the dynamics of the blood flow within the wrist), the accuracy of the HR estimate provided by the PPG device may be reduced or compromised. The light intensity received by the light detectors may be modulated by these movements typically at an order of magnitude or greater than the desired cardiac signal. Therefore, a preprocessing step where the signal effect of these movements is removed would be desirable and improve HR estimation accuracy during motion.
In addition to the deleterious effects of motion, another cause of reduced signal quality in PPG devices may be the characteristics of the local area being sensed. For instance, signal quality can vary dramatically if a wrist-worn PPG sensor is moved only a few millimeters up or down the wrist. In addition, during motion, certain portions of the wrist-worn PPG devices may be subject to more motion depending on their location, position, and/or orientation, and PPG sensors placed on such portions may therefore result in greater degradation of the PPG signal due to motion.
Various embodiments of the present disclosure allow a PPG device to utilize signals based on two or more independently addressable source-detector combinations such that the signal quality of the PPG device is improved, especially during activities involving motion. In some embodiments, PPG signals can be acquired via multiple light paths involving one or more sources and one or more detectors placed at different spatial locations. These multiple PPG signals are then processed to isolate the cardiac component (e.g., by removing the motion component) from the PPG signals. For example, the motion component may be removed based on inputs from the accelerometer, unsupervised learning and/or previously done supervised learning. Additionally, or alternatively, the PPG signals corresponding to these multiple light paths are compared using a quality metric such that the highest-quality PPG signal can be selected to be used for estimating HR or other physiological metrics.
To utilize two or more source-detector pairs for motion signal rejection, a PPG device can use a computer program to identify the motion component of a given signal and remove the motion component from the composite signal, leaving only the cardiac signal as a remainder. In some implementations, the temporal phase of the cardiac waveform is assumed to stay constant between different light paths, while the phase of the motion signal is expected to vary between light paths due to how the PPG sensor interacts with the skin surface during activities involving motion (e.g., pressure at the PPG/skin interface may vary depending on the spatial location of the light source and the light detector of the light path). Using this concept, PPG devices can fit mathematical models to the spatial light path signals to identify the cardiac and motion components. First, PPG signals are extracted by each source-detector combination. For example, two light sources and two light detectors would result in four source-detector combinations. A mathematical model is then fit to the different spatial points, from which characteristic signals are extracted related to the cardiac and motion components of the PPG signals. Such techniques are described in greater detail below with reference to
Although some embodiments are described with reference to HR or cardiac components of PPG signals, the techniques described herein may be extended to other types of physiological metrics described herein (e.g., SpO2) or other types of signals that can be extracted from the PPG signals to determine such physiological metrics. For example, in some embodiments, a method for determining an SpO2 value comprises receiving a first set of one or more PPG signals from one or more PPG sensors, which may include analog signals or digital data sampled from analog components and stored in computer memory. The first set of PPG signals may correspond to red and/or infrared light previously emitted by one or more light sources 102 after the emitted light has interacted with the user's skin, when the monitoring device is worn by the user. The first set of PPG signals may include a noise component. The method for determining the SpO2 value may further comprise receiving a second set of one or more PPG signals from the one or more PPG sensors, which may include analog signals or digital data sampled from analog components and stored in computer memory. For example, the second set of PPG signals may be obtained from different ranges of wavelengths emitted from the light source 102 than the first set of PPG signals. For example, the second set of PPG signals may be obtained from one or more green light sources 102. In some cases, the second set of PPG signals is obtained from a system within the device used for tracking a user's HR. In other cases, the second set of PPG signals is received from a system separate from HR detection. The method for determining the SpO2 value may further comprise filtering the first set of PPG signals based on a feature of the second set of PPG signals to generate a filtered set of PPG signals. Various filtering techniques may be used, depending on embodiments, to remove noise or other features from the first set of PPG signals based on a feature of the second set of PPG signals. As one example, HR may be the feature of the second set of PPG signals. In the case of HR, the device may create a filter based the detected frequency of the HR signal. Examples of filters include a low-pass filter, a high-pass filter, and a narrow band filter that excludes frequencies that are inconsistent with the frequency of the HR signal. The method for determining the SpO2 value may further comprise using one range of wavelengths to better measure an underlying signal on which the wavelengths of the first set of PPG signals operates. Based on this underlying signal (or features derived therefrom), the device can improve the first set of PPG signals based on filtering noise from the first set of PPG signals. Further, the filtered set of PPG signals can be used to create and store a SpO2 value. As an example, the filtered set of PPG signals may have a reduced or eliminated noise component and therefore may serve as a more accurate basis for creating and storing the SpO2 value.
According to some implementations, an intermediate HR estimation is performed based on PPG signals from two or more light paths. For each of the acquired PPG signals, the PPG device may determine an estimate of the HR in beats-per-minute (BPM) and compute a confidence metric associated with the PPG signal, which is indicative of the signal quality for the particular light path associated with the PPG signal. It may also be possible to compute a confidence metric without an intermediate HR estimation, for example by characterizing characteristics (e.g., statistics) of the PPG signal or filtered versions of the PPG signal. In some embodiments, each confidence metric corresponds to a single PPG signal. In other cases, each confidence metric corresponds to multiple PPG signals. For example, a confidence metric may be computed for each way of combining the PPG signals (e.g., signals A+B, signals A+C, signals B+C, signals A+B+C, etc.), as well as for various combinations of PPG signals (e.g., selecting at least two of signals A, B, and C). In other cases, one confidence metric corresponds to a single PPG signal and another confidence metric corresponds to a combination of multiple PPG signals. The PPG device can select an HR estimate from the multiple HR estimates corresponding to the multiple light paths (e.g., by selecting the HR estimate of the PPG signal having the highest confidence metric). Alternatively, the PPG device may assign different weight values to the multiple HR estimates based on the confidence metric values associated with the individual and/or multiple PPG signals and compute a final HR estimate based on the weight values. The confidence values and/or the weight values may be updated or optimized using unsupervised machine learning. The PPG device may implement hysteresis logic which prevents jumping between light paths in a short time window if the confidence metric values corresponding to the two light paths are within a threshold value. The PPG device may also implement logic configured to bias the selection of HR estimates based on user data, activity data, movement data, or other data accessible by the PPG device. The PPG device may apply a smoothing filter on the HR estimates, for example, to improve accuracy and provide a better user experience. Such techniques are described in greater detail below with reference to
One advantage in some of the embodiments described herein is that the spatial information associated with the light sources and/or light detectors can be used by different algorithms to improve HR or other physiological metric estimation accuracy of the PPG sensing device, especially when the user of the device is exercising or performing activities involving motion. Existing implementations typically rely on algorithms to improve the HR or other physiological metric estimation performance, but do not have the benefit of the extra sensor data generated based on multiple light paths.
Embodiments of the present disclosure provide PPG-based devices that utilize multiple source-detector pairs to provide more accurate HR and other physiological metrics. As shown in
For purposes of this disclosure, the term “light path,” in addition to having its ordinary meaning, refers to the probabilistic path of photons from one location to another, typically from the light source (or emitter) to the light detector (or sensor). Photons emitted by the light emitter will follow many different paths to each detector. For simplicity and clarity, the path that results from the optical power-weighted average of all the possible paths is described simply as the “light path” in some embodiments. In some alternative embodiments, “light path” refers to the path along which most of the photons travel. In yet other embodiments, “light path” refers to an approximated vector having an origin at a center of a light source and terminating anywhere in the surface area of a detector, and representing an approximate path of light from the source to the detector.
As described above, a light path represents an approximate path of light from a given source to a given detector. Thus, for example, if there are multiple sources 102 and multiple detectors 106, then a distinct light path exist between each of the multiple sources and each of the multiple detectors. Thus, consistent with the embodiments described herein, PPG signals associated with any of the aforementioned light paths may be selectively obtained and utilized for estimating HR and/or other physiological metrics. For example, the PPG signals corresponding to any of multiple paths may be compared using a quality/confidence metric such as a signal-to-noise ratio (SNR), and the PPG signal having the highest quality can be selected to be used for estimating the HR and/or other physiological metrics.
For example,
With reference to
I/O devices 120 may include, for example, motion sensors, vibration devices, lights, loudspeakers or sound devices, microphones, or other analog or digital input or output devices. For example, in addition to the elements shown in
Memory 112 may comprise RAM, ROM, FLASH memory, or other digital data storage, and may include a control program 118 comprising sequences of instructions which, when loaded from the memory and executed using the processor 110, cause the processor 110 to perform the functions that are described herein. Light sources 102 and light detectors 106 may be coupled to bus 116 directly or indirectly using driver circuitry 122 by which the processor 110 may drive the light sources 102 and obtain signals from the light detectors 106.
The host computer 130 may be coupled to the wireless networking interface 124 via one or more networks 128, which may include one or more local area networks, wide area networks, and/or internetworks using any of terrestrial or satellite links. In some embodiments, host computer 130 executes control programs and/or application programs that are configured to perform some of the functions described herein including but not limited to the processes described herein with respect to
In some embodiments, each light source 102 (e.g., light sources A, B, C, D) can be individually controlled, or each light detector 106 (e.g., light detectors E, F, G, H) can be individually read out when multiple detectors are used, and in such embodiments, PPG sensor data along several different light paths can be collected. For example, in the arrangement shown in
In related aspects, the processor 110 and other component(s) of the PPG device 100 may be implemented as a System-on-Chip (SoC) that may include one or more central processing unit (CPU) cores that use one or more reduced instruction set computing (RISC) instruction sets, and/or other software and hardware to support the PPG device 100.
The PPG device 100 may collect one or more types of physiological and/or environmental data from one or more sensor(s) and/or external devices and communicate or relay such information to other devices (e.g., host computer 130 or another server), thus permitting the collected data to be viewed, for example, using a web browser or network-based application. For example, while being worn by the user, the PPG device 100 may perform biometric monitoring via calculating and storing the user's step count using one or more sensor(s). The PPG device 100 may transmit data representative of the user's step count to an account on a web service (e.g., www.fitbit.com), computer, mobile phone, and/or health station where the data may be stored, processed, and/or visualized by the user. The PPG device 100 may measure or calculate other physiological metric(s) in addition to, or in place of, the user's step count. Such physiological metric(s) may include, but are not limited to: energy expenditure, e.g., calorie burn; floors climbed and/or descended; HR; heartbeat waveform; HR variability; HR recovery; respiration, SpO2, blood volume, blood glucose, skin moisture and skin pigmentation level, location and/or heading (e.g., via a GPS, global navigation satellite system (GLONASS), or a similar system); elevation; ambulatory speed and/or distance traveled; swimming lap count; swimming stroke type and count detected; bicycle distance and/or speed; blood glucose; skin conduction; skin and/or body temperature; muscle state measured via electromyography; brain activity as measured by electroencephalography; weight; body fat; caloric intake; nutritional intake from food; medication intake; sleep periods (e.g., clock time, sleep phases, sleep quality and/or duration); pH levels; hydration levels; respiration rate; and/or other physiological metrics.
The PPG device 100 may also measure or calculate metrics related to the environment around the user (e.g., with one or more environmental sensor(s)), such as, for example, barometric pressure, weather conditions (e.g., temperature, humidity, pollen count, air quality, rain/snow conditions, wind speed), light exposure (e.g., ambient light, ultra-violet (UV) light exposure, time and/or duration spent in darkness), noise exposure, radiation exposure, and/or magnetic field. Furthermore, the PPG device 100 (and/or the host computer 130 and/or another server) may collect data from one or more sensors of the PPG device 100, and may calculate metrics derived from such data. For example, the PPG device 100 (and/or the host computer 130 and/or another server) may calculate the user's stress or relaxation levels based on a combination of HR variability, skin conduction, noise pollution, and/or sleep quality. In another example, the PPG device 100 (and/or the host computer 130 and/or another server) may determine the efficacy of a medical intervention, for example, medication, based on a combination of data relating to medication intake, sleep, and/or activity. In yet another example, the PPG device 100 (and/or the host computer 130 and/or another server) may determine the efficacy of an allergy medication based on a combination of data relating to pollen levels, medication intake, sleep and/or activity. These examples are provided for illustration only and are not intended to be limiting or exhaustive.
As discussed above, PPG devices according to the embodiments of the present disclosure includes multiple light sources and/or light detectors.
In some embodiments, the light sources and detectors are arranged such that the distance between a green light source and a green light detector is shorter than the distance between a red or infrared light source and a red or infrared light detector. For example, as shown in
Although not shown in
In an embodiment, the PPG device 100 may comprise a processor, memory, user interface, wireless transceiver, one or more environmental sensors, and one or more biometric sensors other than the detector 106. For example, embodiments may be implemented using a monitoring device of the type shown in U.S. Pat. No. 8,948,832 of Fitbit, Inc., San Francisco, Calif., the entire contents of which are hereby incorporated by reference for all purposes as if fully set forth herein. In other words, the monitoring device of the type shown in U.S. Pat. No. 8,948,832 could be modified based upon the additional disclosure herein to result in a working activity monitoring apparatus capable of performing the functions that are described herein. Therefore, the present disclosure presumes that the reader knows and understands U.S. Pat. No. 8,948,832, and this disclosure is directed to persons having a level of skill sufficient to modify or adapt the monitoring device of the type shown in U.S. Pat. No. 8,948,832 based upon the additional disclosure herein to result in a working PPG device capable of performing the functions that are described herein.
In various embodiments, the light sources 102 comprise electronic semiconductor light sources, such as LEDs, or produce light using any of filaments, phosphors, or laser. In some implementations, each of the light sources 102 emits light having the same center wavelength or within the same wavelength range. In other cases, at least one light source 102 may emit light having a center wavelength that is different from another one of the light sources 102. The center wavelengths of the light emitted by the light sources 102 may be in the range of 495 nm to 570 nm. For example, a particular green light source 102 may emit light with a center wavelength of 528 nm. In other embodiments, one or more of the light sources 102 may emit red light (e.g., 660 nm center wavelength) or IR light (e.g., 940 nm center wavelength). In some embodiments, one or more of the light sources 102 may emit light with peak wavelengths typically in the range of 650 nm to 940 nm. For example, in various embodiments, a particular red light source may emit light with a peak wavelength of 660 nm, and an infrared light source may emit light with peak wavelengths in the range of 750 nm to 1700 nm. By way of example and not limitation, a particular infrared light source may emit light with a peak wavelength of 730 nm, 760 nm, 850 nm, 870 nm, or 940 nm. In some cases, commercial light sources such as LEDs may provide output at about 20 nm intervals with a center wavelength tolerance of +/−10 nm from the manufacturer's specified wavelength and thus one possible range of useful peak wavelengths for the light sources is 650 nm to 950 nm. The green light sources may be configured to emit light with wavelengths in the range of 495 nm to 570 nm. For example, a particular green light source may emit light with a wavelength of 528 nm. The green light sources may be equally spaced from light detectors 106 as the pairs of red and infrared light sources. For example, if the distance between light detectors 106 and a center of a first red light source is 2 mm, the distance between light detectors 106 and a green light source may also be 2 mm (e.g., equidistant). In some other cases, the distance between the light detectors and one or more light sources is not equidistant. Further, in some embodiments, one or more of the light sources 102 may comprise a single LED package that emits multiple wavelengths, such as green, red and infrared wavelengths, at the same or substantially the same (e.g., less than 1 mm difference) location with respect to multiple detectors. Such LEDs may include multiple semiconductor elements co-located using a single die in a single package.
The spacing of the light sources 102 may be measured from the side of the light source or the center of the light source. For example, the light sources may be configured such that the center of each light source is at a first distance from the edge of the closest one of the light detectors 106. In some embodiments, the first distance may be 2 mm. In some implementations, each light source is located at a second distance from the closest one of the light sources 102, and each light detector is located at a third distance from the closest one of the light detectors 106. In some embodiments, the second and third distances are identical to the first distance. In other embodiments, each of the second and third distances is different from the first distance. The second distance may be identical to or different from the third distance. The particular magnitude of the spacing may depend on a number of factors and this disclosure does not limit the embodiments to any particular spacing. For example, spacing in a range of 1 mm (or less) to 10 mm would be workable in various embodiments.
In some embodiments, independent control of all light sources is provided. In other embodiments, several light sources are controlled together as a gang or bank. A benefit of independent control of each light source, or independent readout from each of multiple detectors (e.g., obtaining independent signals based on the same or different light wavelengths from each of multiple detectors), is that a multiple light path approach may be used to improve the estimation of HR and/or other physiological metrics, as discussed further herein.
Light detectors 106 comprise one or more sensors that is/are adapted to detect wavelengths of light emitted from light sources 102. A particular light source 102 combined with a particular detector 106 may comprise a sensor such as a PPG sensor. A first PPG sensor and a second PPG sensor can share components, such as the same light sources and/or detectors, or have different components and thus the term “PPG sensor,” in addition to having its ordinary meaning, may refer to any of such arrangements although actual embodiments may use multiple components in implementing a PPG sensor. The term “PPG device,” in addition to having its ordinary meaning, may refer to a device including a PPG sensor. A light detector 106, in an embodiment, may comprise one or more detectors for detecting each different wavelength of light that is used by the light sources. For example, a first detector may be configured to detect light with a wavelength of 560 nm, a second detector may be configured to detect light with a wavelength of 940 nm, and a third detector may be configured to detect light with a wavelength of 528 nm. Examples include photodiodes fabricated from semiconductor materials and having optical filters that admit only light of a particular wavelength or range of wavelengths. The light detectors 106 may comprise any of a photodiode, phototransistor, charge-coupled device (CCD), thermopile detector, or complementary metal-oxide-semiconductor (CMOS) sensor. The light detectors 106 may comprise multiple detector elements, as further described herein. One or more of the detectors may comprise a bandpass filter circuit.
In other embodiments, detector 106 comprises one or more detectors configured to detect multiple wavelengths of light. For example, a single detector may be configured to tune to different frequencies based on data received from an electrical digital microprocessor coupled to detectors 106. In another way, the single detector may include multiple active areas where each active area is sensitive to a given range of wavelengths. In an embodiment, a single detector is configured to detect light with wavelengths in the red and IR frequencies and a second detector is configured to detect light with wavelengths in the green frequencies. Further, each of the light sources 102 may use any of one or more different wavelengths of light as previously described.
In an embodiment, light detectors 106 are mounted in a housing with one or more filters that are configured to filter out wavelengths of light other than wavelengths emitted by light sources 102. For example, a portion of the housing may be covered with a filter which removes ambient light other than light in wavelengths emitted by light sources 102. For example, signals from light sources 102 may be received at the light detectors 106 through an ambient light filter that filters out an ambient light source that generates an ambient light with a wavelength that is different from the wavelength that is detected by the detector. Although LEDs and photodiodes are used as examples of the light sources 102 and the light detectors 106, respectively, the techniques described herein may be extended to other types of light sources. For example, the PPG device 100 may include (i) single or multiple LEDs and a multi-element photodetector (e.g., a camera sensor), (ii) an LED array and single or multiple photodiodes, (iii) spatial light modulator (SLM) (e.g., a digital micromirror device [DMD] or a liquid crystal on silicon [LCoS] device) and single or multiple LEDs, other combinations thereof, or other configurations of light sources and detectors.
Certain flow diagrams are presented herein to illustrate various methods that may be performed by example embodiments. The flow diagrams illustrate example algorithms that may be programmed, using any suitable programming environment or language, to create machine code capable of execution by a CPU or microcontroller of the PPG device 100. In other words, the flow diagrams, together with the written description in this document, are disclosures of algorithms for aspects of the claimed subject matter, presented at the same level of detail that is normally used for communication of this subject matter among skilled persons in the art to which the disclosure pertains. Various embodiments may be coded using assembly, C, OBJECTIVE-C, C++, JAVA, or other human-readable languages and then compiled, assembled, or otherwise transformed into machine code that can be loaded into ROM, EPROM, or other recordable memory of the activity monitoring apparatus that is coupled to the CPU or microcontroller and then then executed by the CPU or microcontroller.
In an embodiment, PPG signals obtained from multiple light paths may be processed to filter or reject signal components that are associated with motion of the user, using a computer program to identify the motion component of the signal and remove the identified motion component from the composite signal, leaving the cardiac component as a remainder or final signal.
In various embodiments, the methods of
Generating Estimate for Physiological Metric based on Multiple Source-Detector Pairs
At block 305, the processor 110 obtains one or more first PPG signals based on light received by a first light detector from a set of light sources. Each light source in the set may have a spatial location that is different from that of another light source in the set with respect to the first light detector. Further, each light source in the set may be configured to emit light according to an emission schedule that is different from that of another light source in the set. For example, a first light source in the set may emit light for a given time period at even seconds (e.g., 2, 4, 6, etc. 2/4, 4/4, 6/4, etc., or 2/8, 4/8, 6/8, etc., to name a few examples), and a second light source in the set may emit light for a given time period at odd seconds (e.g., 1, 3, 5, etc. ¼, ¾, 5/4, etc., or ⅛, ⅜, ⅝, etc., to name a few examples). The first PPG signals may be obtained from the one or more light detectors 106 based upon light sources operating at any of the frequencies that have been described above. For example, by executing instructions stored in memory 112 of
At block 310, the processor 110 obtains one or more second PPG signals based on light received by a second light detector from the set of light sources, each light source in the set having a spatial location that is different from another light source in the set with respect to the second light detector. For example, in a manner similar to that described above for block 302, the processor 110 drives the set of light sources 102 and obtains second PPG signals from the second light detector different from the first light detector discussed at block 305. Again, the second set of PPG signals may be represented in any suitable form capable of being processed by a processor, such as the analog signals or digital data sampled from the analog components and stored in computer memory. In an example, the second PPG signals may correspond to green light previously emitted by a light source (or light sources) after the emitted light has interacted with a user's skin, when the PPG device 100 is worn. The second PPG signals may include a motion component and a cardiac component. In another example, the first PPG signals may include a motion component and another physiological component. In some embodiments, the light sources that are used to obtain the first and second PPG signals may use light having the same wavelength.
At block 315, the processor 110 generates a physiological metric based on one or more of the first and second PPG signals. Various techniques for generating the HR, SpO2, and/or other physiological metric value may be used. For example, template matching based on a template previously created and stored from a high-quality PPG signal can be used to provide filter coefficients for a matched filter to maximize the signal-to-noise ratio, or an adaptive filter may be used that tunes a band-pass in real-time based upon heartbeat data derived from the PPG datasets, usually based upon green light sources. Once an HR estimate or other metrics has/have been generated, signals, data values, or data points that are available or apparent in the second PPG signals may be used to modify or improve the first PPG signals so that the resulting modified or improved first PPG signals can be transformed into a more accurate HR estimate. For example, if the intensity of the motion artifact signal varies in space between two or more light paths (i.e. source-detector pairs), a mathematical model can be applied to this spatial intensity difference to remove the motion artifact signal, thus isolating the signal of interest (e.g. cardiac or HR signal).
At block 320, the processor 110 causes the physiological metric to be displayed via a user interface on a user device. For example, the processor 110 may drive the display 114 to display the generated HR, SpO2, and/or other physiological metric value. In other embodiments, the method 300 may be programmed to cause uploading the estimated HR, SpO2, and/or other physiological metric value to host computer 130 for further processing, display, or reporting.
In the method 300, one or more of the blocks shown in
In some embodiments, a single light detector 106 receives reflected light originating from multiple light sources 102. The light detector 106 may distinguish one PPG signal based on light emitted by light source A from another PPG signal based on light emitted by light source B by using time multiplexing (e.g., at block 305 to distinguish among the first PPG signals, and at block 310 to distinguish among the second PPG signals). In some embodiments, light source A emits light according to a first emission schedule that is different from a second emission schedule of light source B. The first emission schedule may specify that light source A is to emit light for X seconds every Y seconds, and the second emission schedule may specify that light source B is to emit light for P seconds every Q seconds. In some cases, X may equal A, and Y may equal B. In other cases, each of X, Y, P, and Q may have the same value. In some cases, the period during which light source A emits light may differ from the period during which light source B emits light by an offset. For example, the PPG device 100 may activate (e.g., emit light via) light source A during a temporal window T1 (without emitting any light via light source B during T1) and activate light source B during a temporal window T2 (without emitting any light via light source A during T2). The light emitted by the light source A during T1 may be detected by the light detector during a temporal window T1′ and the light emitted by the light source B during T2 may be detected by the light detector during a temporal window T2′. Thus, the PPG device 100 can determine first PPG signals originating from light source A based on light detected during T1′ and determine second PPG signals originating from light source B based on light detected during T2′. In some embodiments, there is no temporal gap between T1 and T2 (and/or T1′ and T2′). In other embodiments, there is a temporal gap between T1 and T2 (and/or T1′ and T2′). In some embodiments, there is a partial overlap between T1 and T2.
In some embodiments, at most one light source is emitting light at any given moment. In some embodiments, as long as any light source is emitting light, all of the light detectors are sensing reflected light from the activated light source(s). In some implementations, an internal clock is maintained, and the processor 110 may identify the specific light path (e.g., the light source and the light detector) based on the internal clock and a light source activation schedule indicative of which light sources are emitting during which temporal windows.
In some implementations, a single amplifier is used to amplify the signals generated by multiple light detectors 106, and the use of the multiple light detectors 106 may also be time-multiplexed. For example, the amplifier may amplify signals from light detector A during T0-T1, amplify signals from light detector B during T1-T2 (or T2-T3 with the amplifier not amplifying any signals during T1-T2), and so on. In other cases, each light detector 106 may be associated with a separate amplifier and multiple amplifiers may amplify signals generated by multiple light detectors 106 during a given time period. Although time multiplexing is described herein as a method of distinguishing between different PPG signals, other multiplexing techniques such as frequency multiplexing and wavelength multiplexing may be used to distinguish the PPG signals.
The PPG device 100 may utilize the spatial location associated with each of the light sources 102 and the light detector 106 to further improve the estimate for HR, SpO2, and/or other physiological metric. In some embodiments, the PPG device 100 compares the PPG signals determined based on two (or more) unique light paths and identifies any discrepancy among the PPG signals. For example, if the phases of the two PPG signals are different, the PPG device 100 may discard one of the signals (e.g., based on the signal quality). In some cases, the PPG device 100 may determine the cause of the discrepancy and take different actions based on the cause. For example, if the PPG signals indicate that the discrepancy may be caused by the movement of the PPG device 100, the PPG device 100 may average two or more of the PPG signals and generate the estimate for HR, SpO2, and/or other physiological metric based on the averaged signal(s). On the other hand, if the PPG signals indicate that the discrepancy may be caused by the placement of the sensors or a foreign object in the light path (e.g., hair or other objects), the PPG device 100 may discard the affected PPG signals and generate the estimate for HR, SpO2, and/or other physiological metric based on the remaining PPG signal(s). For example, if the PPG device 100 determines that light paths involving light sources located on one side of the PPG device 100 are underperforming, the PPG device 100 may generate the estimate for HR, SpO2, and/or other physiological metric based on light paths involving light sources on the opposite side of the PPG device 100.
In some embodiments, the estimation for HR, SpO2, and/or other physiological metric may be agnostic to the spatial arrangement of the light sensors and detectors. For example, the PPG device 100 may calculate a spatially-agnostic metric for each PPG signal and generate the estimate for HR, SpO2, and/or other physiological metric based on the PPG signal having the most desirable metric value.
Generating Estimate for Physiological Metric based on Multiple Source-Detector Pairs
At block 405, the processor 110 obtains one or more first PPG signals based on light received by a first light detector from a set of light sources, each light source in the set having a spatial location that is different from another light source in the set with respect to the first light detector. The processor 110 may obtain the first PPG signals in a manner similar to that described with reference to block 305.
At block 410, the processor 110 obtains one or more second PPG signals based on light received by a second light detector from the set of light sources, each light source in the set having a spatial location that is different from another light source in the set with respect to the second light detector. The processor 110 may obtain the second PPG signals in a manner similar to that described with reference to block 310.
At block 415, the processor 110 extracts a physiological component from at least one of the first and second PPG signals by comparing the first and second PPG signals. For example, the physiological component may be a cardiac component. In another example, the physiological component may be a component other than a cardiac component and usable to generate a physiological metric. As discussed herein, the processor 110 may compare the first and second PPG signals to identify the motion component. For example, light paths involving light sources having different spatial locations may have different motion characteristics. For example, signals from a first light path may be known to be more sensitive to blood flow (or a factor other than motion) or otherwise have a relatively higher cardiac component (or another physiological component) in comparison to a motion component, while signals from a second light path may be more sensitive to motion or otherwise have a relatively higher motion component in comparison to a cardiac component (or another physiological component). With such data, the method may use the first light path to sense the cardiac component (or another physiological component) of a signal and the second, different light path to sense the motion component. As a non-limiting example, the second light path may be one where the light detector is further away from the light source in comparison to the first light path.
In some embodiments, if the processor 110 determines that the majority of a PPG signal (e.g., one of the first and second PPG signals) is a cardiac component (or another physiological component), the processor 110 may determine the estimate for HR, SpO2, and/or other physiological metric based on the PPG signal. If the processor 110 determines that the majority of a PPG signal (e.g., one of the first and second PPG signals) is a motion component, the processor 110 may subtract the motion component from the signal (e.g., motion component determined based on accelerometer readings) (or otherwise modify the PPG signal to reduce the effect of the motion component) and determine the estimate for HR, SpO2, and/or other physiological metric based on the PPG signal without the motion component. In some embodiments, for the purpose of determining an HR or a cardiac component, any PPG signal outside the range of a cardiac signal may be discarded. For example, if the PPG device 100 determines that a PPG signal would yield an HR (or BPM) estimate of 300, the PPG device 100 may discard the PPG signal. In some implementations, if the PPG device 100 that the PPG device 100 is undergoing a periodic motion (e.g., based on activity detection or based on sensor data), the PPG device 100 may identify the motion component corresponding to the periodic motion and remove the motion component from the PPG signals. Techniques for motion component removal (e.g., motion artifact removal) are discussed in U.S. Pat. No. 9,005,129 of Fitbit, Inc., San Francisco, Calif., the entire contents of which are hereby incorporated by reference for all purposes as if fully set forth herein. Other techniques that can be used include, but are not limited to, ICA and other forms of blind source separation.
At block 420, the processor 110 generates a physiological metric based on the physiological component. For example, the processor 110 may drive the display 114 to display the generated value for HR, SpO2, and/or other physiological metric. In other embodiments, the method 400 may be programmed to cause uploading the estimated value for HR, SpO2, and/or other physiological metric to host computer 130 for further processing, display, or reporting.
In the method 400, one or more of the blocks shown in
Measuring Heart Rate based on PPG and Motion Data
The blocks illustrated in
Under certain operating conditions, the HR of the user may be measured by counting the number of signal peaks within a time window or by utilizing the fundamental frequency or harmonic frequency components of the signal (e.g., via an FFT). In other cases, such as HR data acquired while the user is in motion, FFTs may be performed on the signal and spectral peaks extracted, which may then be subsequently processed by a multiple-target tracker which starts, continues, merges, and/or deletes tracks of the spectra.
In some embodiments, a similar set of operations may be performed on the motion signature and the output may be used to perform activity discrimination which may be used to assist the multi-spectra tracker 503. For instance, it may be determined that the user was stationary and has begun to move. This information may be used to by the multi-spectra tracker 503 to bias the track continuation toward increasing frequencies. Similarly, the activity identifier 504 may determine that the user has stopped running or is running slower and this information may be used to preferentially bias the track continuation toward decreasing frequencies.
Tracking may be performed by the multi-spectra tracker 503 with single-scan or multi-scan, multiple-target tracker topologies such as joint probabilistic data association trackers, multiple-hypothesis tracking, nearest neighbor, etc. Estimation and prediction in the tracker may be done through Kalman filters, spline regression, particle filters, interacting multiple model filters, etc.
The track selector 505 may use the output tracks from the multiple-spectra tracker 503 and estimate the user's HR based on the output tracks. The track selector 505 may estimate a probability for each of the tracks that the corresponding track is representative of the user's HR. The estimate may be taken as the track having the maximum probability of being representative of the user's HR, a sum of the tracks respectively weighted by their probabilities of being representative of the user's the HR, etc. The activity identifier 504 may determine a current activity being performed by the user which may be used by the track selector 505 in estimating the user's HR. For instance, when the user is sleeping, sitting, lying down, or sedentary, the user's estimated HR may be skewed toward HRs in the 40-80 BPM range. When the user is running, jogging, or doing other vigorous exercise, the user's estimated HR may be skewed toward elevated HRs, such as, for example, in the 90-180 BPM range. The activity identifier 504 may determine the user's current activity (e.g., a current exercise) based at least in part on the speed of the user. The user's estimated HR may be shifted toward (or wholly obtained by) the fundamental frequency of the selected output track when the user is not moving. The output track that corresponds to the user's HR may be selected by the track selector 505 based on criteria that are indicative of changes in activity. For instance, when the user begins to walk from being stationary, the track selector 505 may select the output track that illustrates a shift toward higher frequency based on output received from the activity discriminator 504.
Although some embodiments of the present disclosure are described with respect to HR, the techniques described herein may be extended to other metrics. For example, sensor data generated by the one or more sensors described herein may be used to determine respiration, SpO2, blood volume, blood glucose, skin moisture, and skin pigmentation level and, for example, utilize such metrics for activity detection/identification. Further, the motion removal techniques described in the present disclosure may be used in conjunction with other motion removal techniques.
Removing Motion Component from PPG Signals
At block 605A, the processor 110 acquires signal A from path A. At block 605B, the processor 110 acquires signal B from path B. In an embodiment having N unique paths, at block 605N, the processor 110 acquires signal N from path N, where N is a natural number. Although not shown in
At block 610, the processor 110 fits the data to a model. For example, the processor 110 may determine whether the PPG signals can be fitted to one of the mathematical models stored on the PPG device 100 to isolate the cardiac component (or any other physiological component) from the motion component.
At block 615, the processor 110 applies a linear transform. For example, the processor 110 may apply a linear transform to the motion component identified at block 610 before subtracting the motion component from the composite PPG signals (e.g., signals including both cardiac and motion components, or signals including the motion component and another physiological component). Other correction techniques may include ICA and other forms of blind source separation.
At block 620, the processor 110 estimates a BPM. For example, the processor 110 may estimate the BPM based on the resulting PPG signal from which the motion component has been removed at block 615.
In the method 600, one or more of the blocks shown in
At block 705A, the processor 110 acquires signal A from path A. At block 705B, the processor 110 acquires signal B from path B. In an embodiment having N unique paths, at block 705N, the processor 110 acquires signal N from path N, where N is a natural number. Although not shown in
At block 710A, the processor 110 estimates a BPM based on signal A. At block 710B, the processor 110 estimates a BPM based on signal B. In an embodiment having N unique paths, at block 710N, the processor 110 estimates a BPM based on signal N, where N is a natural number. Although not shown in
At block 715, the processor 110 selects the best BPM estimate using a confidence metric. For example, the processor 110 compares the PPG signals corresponding to multiple light paths using a confidence/quality metric such as SNR and selects the PPG signal having the highest confidence/quality to be used for estimating the HR of the user. In one embodiment, the processor 110 selects one estimate, among those obtained via blocks 710A-710N, that has the highest confidence metric (e.g., SNR). In some embodiments, the confidence/quality metric is based on the characteristics of the waveform (e.g., waveform feature, waveform fidelity, etc.). In an embodiment, the method 700 may use hysteresis logic that prevents jumping between PPG signals of two different light paths within a short time window, for example, if the confidence values of both are within a specified tolerance value.
In another embodiment, the processor 110 selects a BPM estimate based on an activity-specific confidence metric. For example, if the user is determined to be running, the processor 110 may calculate a running-specific confidence metric for each BPM estimate and select a BPM estimate based on the individual running-specific confidence metric values. As another example, if the user is determined to be sitting, the processor 110 may calculate a sitting-specific confidence metric for each BPM estimate and select a BPM estimate based on the individual sitting-specific confidence metric values. In yet another example, if the user is determined to be sleeping, the processor 110 may calculate a sleeping-specific confidence metric for each BPM estimate and select a BPM estimate based on the individual sleeping-specific confidence metric values. The processor 110 may perform activity detection according to any known activity detection techniques.
At block 720, the processor 110 applies a smoothing filter. For example, the processor 110 may apply the smoothing filter to smooth the estimated HR value. Filtering can be performed to improve accuracy, or to present a better user experience, or for both.
In the method 700, one or more of the blocks shown in
In some embodiments, the processor 110 processes multiple PPG signals. For example, each PPG signal of the multiple PPG signals may represent a unique source-detector pair (e.g., light source A to light detector X). The signal quality of the multiple PPG signals may vary depending on a variety of factors (time, location on the user's wrist, weather, orientation of the device, how tightly the user is wearing the device, magnitude and/or direction of movement experienced by the device, activity being performed by the user, and the like). Thus, the processor 110 may determine which one or ones of the multiple PPG signals may more accurately represent the various physiological metrics of the user such that the resulting value is more accurate than that resulting from relying on a single source-detector pair regardless of the changes in the various factors discussed above.
In some embodiments, machine learning can be used to update or optimize the way that the processor 110 processes the multiple PPG signals to generate a value for the physiological parameters. For example, the processor 110 may access a set of PPG signals and determine, using a signal analysis engine stored in one or more memories of the wearable device (e.g., memory 112), whether the accessed set of PPG signals requires updating weight values corresponding to the individual signals in the set. The signal analysis engine can be trained, using unsupervised learning (e.g., in some cases, further based on motion data from the accelerometer and/or previously done supervised learning), to update the weight values based on one or more metrics (e.g., confidence values) associated with the individual signals in the set or other environmental factors (e.g., time, location on the user's wrist, weather, orientation of the device, how tightly the user is wearing the device, magnitude and/or direction of movement experienced by the device, activity being performed by the user, and the like). For example, the processor 110 may determine the final value to be outputted on the graphical user interface based on the PPG signal (or an estimated value based on the PPG signal) with the highest confidence value. In some cases, some or all of the PPG signals (e.g., two or more PPG signals with the highest confidence values, or two or more estimates with the highest confidence values) may be averaged or weighted for the purpose of determining the final value.
The processor 110 may be notified (e.g., by a motion sensor of the wearable device) of movements undergone by the wearable device. For example, the motion sensor may be an accelerometer configured to generate motion data indicative of the movements undergone by the wearable device. The processor 110 may output a final value of the physiological metric tracked by the wearable device based at least on the updated weight values. The final value may be displayed on a display of the wearable device.
In some embodiments, the confidence values (associated with the individual PPG signals or estimates generated therefrom), the final values of the physiological metrics, the identity of the PPG signals (or source-detector paths) used for generating the final values, and/or the weights corresponding to the individual PPG signals (or source-detector paths), or any combination thereof, can serve as a training data set for updating or improving (e.g., by adjusting the individual weights) the way in which the processor 110 determines the final values of the physiological metrics. For example, if the confidence values associated with a given source-detector path exhibit a pattern of being lower than a threshold value when there is user motion along a given rotational axis, the weight value associated with given source-detector path may be reduced when the processor 110 detects user motion along the given rotational axis. As another example, if the confidence values associated with a given source-detector path exhibit a pattern of being higher than a threshold value regardless of environmental conditions, the weight value associated with given source-detector path may be increased for all environmental conditions.
In some embodiments, at least some of the PPG signals are averaged together to produce a single PPG signal. For example, two of the three PPG signals corresponding to the three unique light paths are averaged together to produce a single PPG signal. Then, the PPG device 100 may generate HR, SpO2, and/or other physiological metrics based on the averaged PPG signal and the remaining one of the three PPG signal. For example, in some cases, the light detector can be a camera sensor consisting of millions of pixels or individual detectors. Since performing a BPM estimation on millions of individual pixels may be difficult, the PPG device 100 may combine the signals received at some of the pixel locations by averaging such pixel signals together, and performing the BPM estimation on the fewer number of signals. In some implementations, even with the averaging of at least some of the PPG signals, the number of signals based on which the PPG device 100 generates the HR, SpO2, and/or other physiological metrics is greater than 1. In other embodiments, none of the PPG signals are averaged together (i.e., each PPG signal is separately considered by the PPG device 100).
Additionally or alternatively, the processes of
In some embodiments, one or more light paths are selectively activated based on the user activity. For example, if it is determined that the user is currently sleeping, only a low number of light paths may be activated (e.g., 1 or 2), if it is determined that the user is currently walking, a higher number of light paths may be activated (e.g., 3 or 4), and if it is determined that the user is currently engaging in a high intensity exercise (e.g. running), an even higher number of light paths may be activated (e.g., 5 or 6). Alternatively, in some embodiments, different activities may result in activation of different but the same number of light paths. For example, if it is determined that the user is currently sleeping, only light paths A and B may be activated, if it is determined that the user is currently walking, only light paths C and D may be activated, and if it is determined that the user is currently engaging in an exercise, only light paths E and F may be activated. In other embodiments, the number and the identity of light paths may both vary depending on the detected activity. For example, if it is determined that the user is currently sleeping, only light path A may be activated, if it is determined that the user is currently walking, only light paths B and C may be activated, and if it is determined that the user is currently engaging in an exercise, only light paths D, E, and F may be activated. Such embodiments allow activity-specific arrangement of light sources and detectors to be used, such that the signal-to-noise ratio is improved or optimized. Although described in the activity detection context, these techniques can be extended to movement detection to provide movement-state-specific (e.g., based on whether the device is undergoing no movement, movement less than a threshold, moving greater than a threshold value, etc.) arrangement of light sources and detectors.
As discussed above, one of the advantages in some of the embodiments described herein is that the spatial information associated with the light sources and/or light detectors can be used by different algorithms to improve estimation accuracy of the PPG sensing device for HR, SpO2, and/or other physiological metrics, especially when the user of the device is exercising or performing activities involving motion. Existing implementations typically rely on algorithms to improve the estimation performance for HR, SpO2, and/or other physiological metrics, but do not have the benefit of the extra sensor data generated based on multiple light paths.
EE 1. A method for generating physiological metrics, the method comprising: obtaining one or more first PPG signals based on light received by a first light detector from a set of light sources, each light source in the set (i) having a spatial location that is different from that of another light source in the set and (ii) configured to emit light according to an emission schedule that is different from that of another light source in the set; obtaining one or more second PPG signals based on light received by a second light detector from the set of light sources; generating a physiological metric based on one or more of the first and second PPG signals; and causing the physiological metric to be displayed via a user interface on a user device.
EE 2. The method of EE 1, wherein the set of light sources comprises at least one green light source, at least one red light source, and at least one infrared light source.
EE 3. The method of EE 1, wherein the set of light sources are provided across two or more light source packages, each light source package including at least one light source from the set and located at least 1 mm away from another one of the two or more light source packages.
EE 4. The method of EE 1, wherein the set of light sources comprises two or more light sources included in a single light source package, wherein the two or more light sources in the light source package are each configured to emit light having a wavelength that is different from another light source in the light source package.
EE 5. The method of EE 4, wherein the two or more light sources in the light source package are located within less than 1 mm from each other.
EE 6. The method of EE 5, wherein the two or more light sources comprise at least a red light source and an infrared light source.
EE 7. The method of EE 1, further comprising: obtaining a first signal of the one or more first PPG signals based on the light received by the first light detector from a first light source in the set during a first temporal window; obtaining a second signal of the one or more first PPG signals based on the light received by the first light detector from a second light source in the set that is different from the first light source during a second temporal window that is different from the first temporal window; and generating the physiological metric based at least on the first signal and the second signal.
EE 8. The method of EE 7, further comprising removing a motion component from at least one of the first and second signals based on location information associated with one or more of the first light source, the second light source, and the first light detector.
EE 9. The method of EE 7, further comprising: determining a first confidence metric associated with the first signal; determining a second confidence metric associated with the second signal; and based on a determination that the first confidence metric is higher than the second confidence metric, generating the physiological metric based on the first signal.
EE 10. The method of EE 1, further comprising: obtaining a first signal of the one or more first PPG signals based on the light received by the first light detector from a first light source in the set; obtaining a second signal of the one or more second PPG signals based on the light received by the second light detector from the first light source in the set; and generating the physiological metric based at least on the first signal and the second signal.
EE 11. The method of EE 10, further comprising removing a motion component from at least one of the first and second signals based on location information associated with one or more of the first light source, the first light detector, and the second light detector.
EE 12. The method of EE 10, further comprising: determining a first confidence metric associated with the first signal; determining a second confidence metric associated with the second signal; and based on a determination that the first confidence metric is higher than the second confidence metric, generating the physiological metric based on the first signal.
EE 13. The method of EE 1, wherein the first and second light sources are spatially located between the first and second light detectors.
EE 14. The method of EE 1, wherein the first and second light detectors are spatially located between the first and second light sources.
EE 15. The method of EE 1, wherein the first and second light sources and the first and second light detectors are arranged such that each of the set of light sources is equidistant from the first and second light detectors.
EE 16. A wearable device for tracking a physiological metric of a user, comprising: two or more light emitters that are each configured to emit light at a respective emission schedule; a first light sensor positioned at a first location on the wearable device and configured to receive light from the two or more light emitters; a second light sensor positioned at a second location on the wearable device different from the first location and configured to receive light from the two or more light emitters; and an element configured to receive signals generated by the first and second light sensors and determine a value of the physiological metric associated with the user of the wearable device based on the received signals.
EE 17. The wearable device of EE 16, wherein the first light sensor is configured to receive light along a first rotational axis of the wearable device, and the second light sensor is configured to receive light along a second rotational axis of the wearable device different from the first rotational axis.
EE 18. The wearable device of EE 16, wherein the element is further configured to: receive a first signal based on light received by the first light sensor from a first light emitter of the two or more light emitters during a first temporal window; receive a second signal based on light received by the first light sensor from a second light emitter of the two or more light emitters that is different from the first light emitter during a second temporal window that is different from the first temporal window; and generate the value of the physiological metric based at least on the first signal and the second signal.
EE 19. The wearable device of EE 16, wherein the element is further configured to: receive a first signal based on light received by the first light sensor from a first light emitter of the two or more light emitters; receive a second signal based on light received by the second light sensor from the first light emitter; and generate the value of the physiological metric based at least on the first signal and the second signal.
EE 20. A method comprising: accessing, at a wearable device for tracking a physiological metric, a set of PPG signals; determining, using a signal analysis engine stored in one or more memories of the wearable device, whether the accessed set of PPG signals requires updating weight values corresponding to the individual signals in the set, the signal analysis engine being trained, using unsupervised learning, to update the weight values based on one or more environmental factors and being notified, via a motion sensor of the wearable device, of movements undergone by the wearable device, the motion sensor configured to generate motion data indicative of the movements undergone by the wearable device; and outputting, based at least on the updated weight values, a value of the physiological metric associated with a user of the wearable device on a display of the wearable device.
Information and signals disclosed herein may be represented using any of a variety of different technologies and techniques. For example, data, instructions, commands, information, signals, bits, symbols, and chips that may be referenced throughout the above description may be represented by voltages, currents, electromagnetic waves, magnetic fields or particles, optical fields or particles, or any combination thereof.
The various illustrative logical blocks, and algorithm steps described in connection with the embodiments disclosed herein may be implemented as electronic hardware, computer software, or combinations of both. To clearly illustrate this interchangeability of hardware and software, various illustrative components, blocks, and steps have been described above generally in terms of their functionality. Whether such functionality is implemented as hardware or software depends upon the particular application and design constraints imposed on the overall system. Skilled artisans may implement the described functionality in varying ways for each particular application, but such implementation decisions should not be interpreted as causing a departure from the scope of the present disclosure.
The techniques described herein may be implemented in hardware, software, firmware, or any combination thereof. Such techniques may be implemented in any of a variety of devices, such as, for example, wearable devices, wireless communication device handsets, or integrated circuit devices for wearable devices, wireless communication device handsets, and other devices. Any features described as devices or components may be implemented together in an integrated logic device or separately as discrete but interoperable logic devices. If implemented in software, the techniques may be realized at least in part by a computer-readable data storage medium comprising program code including instructions that, when executed, performs one or more of the methods described above. The computer-readable data storage medium may form part of a computer program product, which may include packaging materials. The computer-readable medium may comprise memory or data storage media, such as random access memory (RAM) such as synchronous dynamic random access memory (SDRAM), read-only memory (ROM), non-volatile random access memory (NVRAM), electrically erasable programmable read-only memory (EEPROM), FLASH memory, magnetic or optical data storage media, and the like. The techniques additionally, or alternatively, may be realized at least in part by a computer-readable communication medium that carries or communicates program code in the form of instructions or data structures and that can be accessed, read, and/or executed by a computer, such as propagated signals or waves.
According to some embodiments, the techniques described herein are implemented by one or more special-purpose computing devices. The special-purpose computing devices may be hard-wired to perform the techniques, or may include digital electronic devices such as one or more application-specific integrated circuits (ASICs) or field programmable gate arrays (FPGAs) that are persistently programmed to perform the techniques, or may include one or more general purpose hardware processors programmed to perform the techniques pursuant to program instructions in firmware, memory, other storage, or a combination. Such special-purpose computing devices may also combine custom hard-wired logic, ASICs, or FPGAs with custom programming to accomplish the techniques. The special-purpose computing devices may be desktop computer systems, portable computer systems, handheld devices, networking devices, wearable devices, or any other device that incorporates hard-wired and/or program logic to implement the techniques.
Processor(s) in communication with (e.g., operating in collaboration with) the computer-readable medium (e.g., memory or other data storage device) may execute instructions of the program code, and may include one or more processors, such as one or more digital signal processors (DSPs), general purpose microprocessors, ASICs, FPGAs, or other equivalent integrated or discrete logic circuitry. Such a processor may be configured to perform any of the techniques described in this disclosure. A general purpose processor may be a microprocessor; but in the alternative, the processor may be any conventional processor, controller, microcontroller, or state machine. A processor may also be implemented as a combination of computing devices, for example, a combination of a DSP and a microprocessor, a plurality of microprocessors, one or more microprocessors in conjunction with a DSP core, or any other such configuration. Accordingly, the term “processor,” as used herein may refer to any of the foregoing structure, any combination of the foregoing structure, or any other structure or apparatus suitable for implementation of the techniques described herein. Also, the techniques could be fully implemented in one or more circuits or logic elements.
The techniques of this disclosure may be implemented in a wide variety of devices or apparatuses, including a wearable device, a wireless handset, an integrated circuit (IC) or a set of ICs (e.g., a chip set). Various components, or units are described in this disclosure to emphasize functional aspects of devices configured to perform the disclosed techniques, but do not necessarily require realization by different hardware units. Rather, as described above, various units may be combined in a hardware unit or provided by a collection of inter-operative hardware units, including one or more processors as described above, in conjunction with suitable software and/or firmware.
Although the foregoing has been described in connection with various different embodiments, features or elements from one embodiment may be combined with other embodiments without departing from the teachings of this disclosure.
This application is a nonprovisional of U.S. Provisional Patent Application No. 62/482,997, filed Apr. 7, 2017, titled “MULTIPLE SOURCE-DETECTOR PAIR PHOTOPLETHYSMOGRAPHY (PPG) SENSOR,” which is hereby incorporated by reference in its entirety.
| Number | Date | Country | |
|---|---|---|---|
| 62482997 | Apr 2017 | US |
| Number | Date | Country | |
|---|---|---|---|
| Parent | 15948970 | Apr 2018 | US |
| Child | 17366235 | US |
