Health monitor.

(54) Title: HEALTH MONITOR.

(54) Title: HEALTH MONITOR.

(57) Summary

Apparatus and methods are disclosed for monitoring and evaluating exercise-related activities performed by a subject; a health monitor comprising an accelerometer, at least one physiological sensor, and a digital processor is configured to be supported by a subject and to identify a type of activity (eg, running, cycling, swimming) from among a plurality of different activities ; the health monitor is further configured to determine, from movement data and / or heart data, levels of health benefits received from exercise; Health benefits can be based on standards established by recognized health entities, so that reported health benefits are less susceptible to errors associated with conventional step counting.

(57) Abstract

Apparatuses and methods are disclosed for monitoring and evaluating exercise-related activities performed by a subject. A health monitor comprising an accelerometer, at least one physiological sensor, and digital processor is configured to be supported by a subject and identify an activity type (eg, running, biking, swimming) from among a plurality of different activities. The health monitor is further configured to determine from motion data and / or cardiac data levels of health benefits received from the exercise. The health benefits may be based on standards established by recognizable health entities, so that the reported health benefits are less susceptible to errors associated with conventional step counting.

HEALTH MONITOR

RELATED REQUESTS

This application is a continuation in part of U.S. Application Serial No. 13 / 840,098, entitled Versatile Sensors with Data Fusion Functionality, which was filed on March 15, 2013, which is a continuation in part of the U.S. Application Serial No. 13 / 690,313, entitled Intelligent Activity Monitoring, filed on November 30, 2012, claiming the benefit of Provisional Application Serial No. No. 61 / 566,528, which is titled Intelligent Activity Monitoring, which was filed on December 2, 2011. The full descriptions of the above applications are incorporated herein by reference.

FIELD OF THE INVENTION

This description refers generally to apparatus and methods for monitoring a subject's physical activity and determining a level of health benefit from the detected activity.

BACKGROUND OF THE INVENTION

Currently there are small sensors that can be used by a user to monitor a physical activity performed by the user or two similar types of user activity. As an example, the FitLinxx® Act¡Ped + (available from FitLinxx, Shelton, CT, USA) is a small device that can be clipped to a shoe and used to monitor walking and running activities by the user. When a user walks or runs, an on-board accelerometer outputs data that is stored in the device for later transmission to a computer system. The computer system can analyze the data to determine the type of activity, and calculate various parameters of the activity (for example, duration of activity, total steps, distance traveled, and calories burned). The results of the data analysis can be presented on the computer screen, so that a user can review the details of their activity. Results can be stored so that the user can keep track of exercise regimens and monitor progress toward exercise goals or so that the data can be used by medical personnel to monitor recovery from illness or injury. . Other modern activity monitors perform similar functions with varying degrees of accuracy.

Today, pedometer and activity monitor sellers and users consider 10,000 steps per day a healthy amount of activity. This data has accumulated over time, but was based in part on incorrect step counts throughout a subject's day. The literature frequently reports errors in the counting of steps by various pedometers and algorithms used. Although manufacturers continually try to improve the accuracy of their devices, many devices can be very inaccurate depending on a user's activity patterns. For example, activity monitors detect and count steps differently. Some activity monitors collect steps from activities including driving, spinning, restless leg syndrome, eating (for some wrist-worn devices), and taking a train or bus. These are attached to a baseline of imprecise step counts for one day. Making an all-day activity monitor correctly count steps and ignore all other movements may require multiple inputs and may require more expensive, more sophisticated, and more power-consuming devices. Such devices, while they can be very accurate, can be very expensive for many consumers.

BRIEF DESCRIPTION OF THE INVENTION

The inventors have appreciated that although a record of steps, distance traveled, duration of activity, or calories burned is useful for many users to keep track of physical activities, the usefulness of the data can be further enhanced by calculating, of the data representative of the activity, the levels or units of health benefits received by the user as a result of the activity. For example, two people who walk the same number of steps in a day, walk the same distance, and take the same amount of time to do them can receive markedly different health benefits from their physical activities. For example, an overweight and mature person may receive more health benefits from activity than a young and fit individual. The inventors have recognized that in some settings (e.g. medical diagnostics and treatment), it may be more relevant to assess the quality, or level of health benefits, of the activity performed rather than an unprocessed number of steps, distance traveled, calories burned or duration of activity. In some implementations, health benefit units can be standardized to provide a more accurate analysis of fitness benefits across populations and across fitness devices.

The inventors have developed methods and apparatus that can be used to determine levels of the health benefits of activities detected by IOS portable health or activity monitors. Determining health benefits can be tied to recognizable health standards set by a health entity, the Center for Disease Control, or the World Health Organization, for example. According to some modalities, the health benefit levels of the activities performed can be determined from a combination of data that may include, but are not limited to, type of activity, continuous time in activity, dynamism in activity, and exercise guidelines established by the health entity. By matching health benefits to recognizable health standards and using certain criteria to establish notable health activity, the inventors have developed a system that more accurately reflects the effectiveness of exercise detected by portable health monitors.

In some implementations, the accuracy of a health monitor can be improved by using cardiac data from a subject in addition to movement data representative of an activity performed by the subject. For example, heart rate data in combination with movement data can be used to improve the accuracy of calorie burning for an activity over a time interval, thereby improving the accuracy of the intensity level for the activity. . The type of activity may or may not be identified by the health monitor.

According to some embodiments, a health monitor is configured to be supported by a subject and may comprise an accelerometer configured to generate motion data in response to a first activity performed by the subject, a heart sensor configured to detect at least a subject's heart rate during the first activity performance, an onboard power source configured to provide power to the accelerometer and heart sensor, and a processor configured to calculate a first calorie burn average for the first activity performed by the subject according to a first method based on data from both the accelerometer and the heart sensor.

In some aspects, the processor is further configured to determine at least one health monitor power saving mode based on data received from the heart sensor. The at least one energy saving mode may, for example, comprise a mode where energy is reduced to at least the cardiac sensor for an interval between the T wave and P wave portions of successive cardiac cycles.

In some implementations, the processor is configured to calculate the first average calorie burn using values ​​of the subject's heart rate and speed, and includes the subject's heart rate and foot contact time values. According to some aspects, the processor is set to calculate the first average calorie burn using heart rate values ​​and a maximum VO2 value.₂ calculated by the subject.

In some aspects, a health monitor processor can be further configured to calibrate a second method for determining calorie burn for an activity subject, where the calibration is based on the results obtained from the first method. The second method may, for example, comprise the determination of calorie burn based on heart rate and not based on accelerometer data.

According to some implementations, a health monitor may further comprise a timer, and can be configured to calculate a second average calorie burn for a second activity performed by the subject during a first time interval of a sequence of first time intervals and determine if the second average calorie burn falls within at least minus a recommended scale of values ​​banned by a health entity for the duration of the first time interval. The health entity can be the Center for Disease Control or the World Health Organization. In some respects, the health entity may be a doctor, qualified medical personnel, or professional trainer.

According to some aspects, a processor of a health monitor can be configured to verify the performance of the second activity of an energy spectrum of the motion data. Additionally or alternatively the processor may be configured to determine the intensity level of the second activity from a heart rate and / or respiratory rate of the subject.

In some aspects, The processor of a health monitor can also be configured to record a first credit value for every first time interval that the average calorie burn falls within a recommended first scale of at least one recommended scale for the duration of the first time interval and record a second credit value for each first time interval that the average calorie burn falls within a second recommended scale of at least one Recommended scale for the duration of the first time interval. In some implementations, the first recommended scale of at least one recommended scale can be between approximately 3.5 kilocalories per minute and approximately 7 kilocalories per minute and the second recommended scale of at least one recommended scale includes values ​​greater than approximately 7 kilocalories per minute . In some aspects, the processor may be further configured to identify one type of the second activity performed by the subject from among a plurality of types of activities for which the health monitor is configured to recognize, and may identify non-activity activities. humans.

According to some implementations, a processor of a health monitor may be further configured to determine, from the data received from the accelerometer, a period of increased activity performed by the subject, and to determine a heart rate recovery time after leaving the heart rate. exercise. In some aspects, the processor can be further configured to calculate heart rate variability or a LF / AF spectral energy ratio from data received from the heart sensor.

In some implementations, a health monitor includes two electrodes configured to electrically connect to the subject using a release adhesive. In some respects, a health monitor accelerometer comprises a three-axis accelerometer.

The above aspects, features and implementations can be used in any suitable combination in one or more modalities of a health monitor.

Methods for operating a health monitor are also contemplated. According to some embodiments, a method of determining fitness metrics for a subject by means of a health monitor configured to be supported by the subject may comprise acts to receive, by a processor, motion data that was generated by the fitness monitor. health in response to a first activity performed by the subject. A method may further include receiving, from a cardiac sensor in communication with the processor, cardiac data for the subject detected during the performance of the first activity, and calculating a first average calorie burn for the first activity performed by the subject according to with a first algorithm based on information from both movement data and cardiac data.

According to some implementations, a method may further comprise executing a health monitor power saving mode based on the data received from the heart sensor. In some aspects, a method may further comprise reducing energy to at least the cardiac sensor for an interval between the T wave portions and P wave portions of successive cardiac cycles.

In some aspects, the calculation of the first calorie burn average includes values ​​of the subject's heart rate and speed or values ​​of the subject's heart rate and foot contact time. In some implementations, the calculation of the first average calorie burn includes heart rate values ​​and a maximum VO2 value.₂ calculated by the subject.

According to some aspects, a method may further comprise calibrating a second algorithm for determining calorie burn for a subject performing an activity, wherein the calibration is based on the results obtained from the first algorithm. In some aspects, the second algorithm comprises determining calorie burn based on cardiac data and not based on movement data.

In some implementations, a method of operating a health monitor comprises calculating a second average calorie burn for a second activity performed by the subject during a first time interval of a sequence of first time intervals and determining whether the second average of calorie burn falls within at least a recommended scale of values ​​prohibited by a health entity for the duration of the first time interval. In some respects, the health entity is the Center for Disease Control or the World Health Organization.

According to some implementations, a method may further comprise determining an energy spectrum for the motion data and verifying the performance of the second activity of the energy spectrum. In some aspects, A method further comprises recording a first credit value for each first time interval that the second average calorie burn falls within a recommended first scale of at least one recommended scale for the duration of the first time interval and recording a second credit value for each first time interval that the second average calorie burn falls within a second recommended scale of at least one recommended scale for the duration of the first time interval. In some respects, the first recommended scale of at least one recommended scale is between approximately 3.5 kilocalories per minute and approximately 7 kilocalories per minute and the second recommended scale of at least one recommended scale includes values ​​greater than approximately 7 kilocalories per minute. A method may further comprise identifying one type of the second activity performed by the subject from among a plurality of types of activities for which the health monitor is configured to recognize.

According to some implementations, a method for operating a health monitor may further comprise determining, from the motion data, a period of increased activity performed by the subject, and determining a heart rate recovery time after stopping activity. In some implementations, a method further comprises calculating heart rate variability or a BF / AF spectral energy ratio from the cardiac data.

The above aspects and implementations of various acts and features can be combined in any suitable way in one or more modalities to operate a health monitor.

Tangible storage devices or computer-readable media that include computer-readable instructions are also contemplated which, when run by at least one processor of a health monitor including a heart sensor and accelerometer, adapt the health monitor to run either or a combination of the above-mentioned acts to operate a health monitor. For example, a storage device may include computer-readable instructions that tailor the health monitor to receive, by the processor, motion data that was generated by the accelerometer in response to a first activity performed by a subject, to receive cardiac data to the subject detected during the performance of the first activity, and calculating a first average calorie burn for the first activity performed by the subject according to a first algorithm based on information from both the movement data and the heart data. In some respects, the instructions may further adapt the health monitor to calculate a second average calorie burn for a second activity performed by the subject during a first time interval of a sequence of first time intervals, and determine whether the second Average calorie burn falls within at least a recommended scale of values ​​prohibited by a health entity for the duration of the first time interval. In some implementations, The instructions may further adapt the health monitor to record a first credit value for each first time interval that the second average calorie burn falls within a recommended first scale of at least one recommended scale for the duration of the first time interval and record a second credit value for each first time interval that the second average calorie burn falls within a second recommended scale of by the minus a recommended scale for the duration of the first time interval.

The terms sensor or monitor as used herein refer to a small electronic device configured to detect at least one parameter of a subject to which the sensor or monitor can be attached. A sensor can have limited processing power. A sensor can be configured to transmit data wirelessly or via a wired link to a second device. The sensor or monitor can also be used as a key when referring to a health monitor or smart health monitor, the intended meaning is obvious from the context.

The terms smart sensor, smart activity monitor, activity monitor, or smart monitor as used herein can refer to a small electronic device configured to detect at least one parameter of a subject and representative processing data of an activity. detected. Such a device has a higher processing power than a sensor, and can include at least one digital processor. A smart sensor can be configured to transmit and / or receive data wirelessly or via a wired link to and / or from a second device. The health monitor can also be used as a key when referring to a smart health monitor, the intended meaning is obvious from the context.

The term digital processor or processor as used herein can refer to at least one microcontroller, microprocessor, digital signal processor (DSP), application specific integrated circuit (ASIC). or Field Programmable Gate Array (FPGA).

Digital Processor can also be used to refer to any combination of the above digital processing devices, including more than one particular device.

The terms versatile sensor or versatile monitor as used herein can refer to a small electronic device configured to detect by (or less than one parameter of an entity, to process received data, and to flexibly participate in a small area network. A versatile sensor can be configured to transmit data wirelessly or via a wired link to a plurality of devices.

The term "health monitor" as used herein can refer to a single smart sensor or a detection system that can include at least one smart sensor. A health monitor can include one or more versatile sensors in some modalities, and can include one or more sensors in some modalities.

The above aspects and other aspects, modalities and characteristics of the present teachings can be fully understood from the following description in conjunction with the accompanying drawings.

BRIEF DESCRIPTION OF THE DRAWINGS

Those skilled in the art will understand that the figures, described herein, are for purposes of illustration only. It should be understood that in some examples various aspects of the invention may be shown exaggerated or elongated to facilitate an understanding of the invention. In the drawings, like reference characters generally refer to similar features, functionally similar and / or structurally similar elements throughout the various figures. Drawings are not necessarily to scale, instead emphasis is placed after illustrating the principles of the teachings. The drawings are not intended to limit the scope of the present teachings in any way.

Figure 1A shows examples of components that can be included in a health monitor, according to some embodiments;

Figure IB is a block diagram illustrating examples of selected electrical components that can be included in a health monitor, in some implementations;

Figures 1C to 1F show various embodiments of health monitors that are configured to attach to a subject, and that include an LED display for communicating information to a user;

Figure 1G shows the elements of a patch-type health monitor including activity detection and heart rate detection, in some modalities;

Figure 2A is an illustrative example of a state diagram for low power operation of a health monitor, according to some embodiments;

Figure 2B is an illustrative example of a state diagram for low power operation of a health monitor, according to some embodiments;

Figures 2C to 2D show illustrative examples of energy saving methods for operating a health monitor, in some implementations;

Figure 2E illustrates a PQRST cardiac waveform, according to some embodiments;

Figure 2F illustrates types of data that can be detected with a health monitor and fitness metrics that can be calculated, according to some implementations;

Figure 3 shows an exemplary architecture of a health monitor data processing system according to some embodiments;

Figures 4A to 4B show illustrative examples of multi-axis accelerometer data and combined acceleration derivative data for various types of activities;

Figures 5A to 5C illustrate examples of the membership functions that can be used in fuzzy logic identification of different types of activities, according to some modalities;

Figure 6 represents acts of a method 600 for processing activity data, according to some embodiments;

Figure 7A illustrates acts related to a sub-process to determine the types of activity and intensity levels, according to some modalities;

Figure 7B illustrates acts related to a sub-process to determine health credits, according to some modalities;

Figure 7C illustrates acts related to a sub-process to determine health credits, according to some modalities;

Figure 8A depicts an example of a metabolic equivalents (MET) look-up table, in accordance with some implementations;

Figure 8B depicts an example of a health credit data stream that can be produced by a health monitor, according to some implementations;

Figure 8C depicts an example of compressed health credit data that can be stored on board a health monitor, according to some implementations;

Figure 9A depicts an example of health credit data that can be stored onboard a health monitor in a temporary activity buffer, according to some implementations;

Figure 9B depicts examples of compressed health credit data and activity-bearing data that can be produced and stored on board a health monitor, in accordance with some implementations; Y

Figures 10A to 10C depict illustrative examples of multi-axis accelerometer data and combined acceleration derivative data for an activity with the health monitor attached to a subject at various locations.

The features and advantages of the present invention will be more apparent from the detailed description set forth below when taken in conjunction with the drawings.

DETAILED DESCRIPTION OF THE INVENTION

I. Overview

As described above, the inventors have appreciated that conventional methods of using pedometers and activity monitors rely primarily on step counts and that many of such conventional devices inaccurately record step counts during a day of use. For example, the inventors have observed that pedometers manufactured by different commercial entities have widely varying degrees of accuracy in reporting steps. Two activity monitors fixed to the same subject can report a number of steps taken by the subject that differ by as much as 25%, and the number of calories burned that can differ by more than 50%.

With such variability between different activity monitors, the inventors have recognized that their use in a clinical setting as a diagnostic tool is compromised. For example, a doctor who prescribes an exercise regimen for an individual cannot know with reasonable certainty whether the prescription is being fulfilled by the individual, or it is being fulfilled in a way that the individual derives health benefits from the activity. Consequently, exercise regimens designed for physical conditioning or medical purposes (for example, treating illness or recovering health from an incident) cannot be reliably formulated and / or monitored based on step counts.

The inventors have recognized that activity monitors can be improved in terms of data acquisition and data analysis by calculating levels of health benefits or health units or credits for activities performed by a user and detected by the activity monitor. . The evolution of health credits can be linked to recognizable health standards and can be a more meaningful and universal assessment of an activity than raw numbers in relation to other parameters such as number of steps, distance traveled, calories burned and duration of an activity. Health credits can be based in part on the parameters detected by conventional activity monitors, but the data can be further processed to better assess a quality of the activity performed.

The inventors have developed apparatus and methods that can be used to determine the health credits of an activity performed by a subject. According to some modalities, a health monitor system is configured to detect an activity performed by a subject, identify a type of activity, and determine an intensity level of performance of that activity. A health monitor can be configured to determine, in fine time resolution, continuous duration intervals for which activity is performed, and calculate a health credit amount based on continuous duration intervals, intensity levels and the guidelines established by the health entity.

The inventors have also recognized that in some modalities, the accuracy of health credit calculations can be improved by using cardiac data (for example, heart rate data from a heart sensor attached to a subject) in combination with motion data. (eg motion data from an accelerometer supported by the subject) to calculate calorie burn from various activities performed by the subject. Consequently, the device modalities can capture and analyze the cardiac waveform in addition to the movement data. Additionally, the inventors have realized that a variety of different fitness metrics have been determined from cardiac data used alone or in combination with motion data.

Determining health benefit levels for activities performed and determining fitness metrics using the methods described herein can provide more accurate and useful diagnostic tools for evaluating exercise regimens and fitness levels. physics of the subjects. By evaluating physical activity in terms of health benefit, the individual or doctor may have a more accurate representation of monitored exercise than step counts. Additionally, the individual or doctor can easily determine if the individual receives any health benefits from exercise.

Various modalities of apparatus, methods and systems for a health monitor are described in more detail in the following sections of the specification.

II. Exemplary apparatus

With reference to Figure 1A, a health monitor 100 may comprise a small electronic device that can be attached to or supported by a subject, and that can be configured to identify one type of activity from among a plurality of different types of activities that are being performed. can be performed by the subject. Activity data {for example_F datQS «...« - * during the exercise performed by the subject can be processed on board, in some modes, to determine the health credits that quantitatively represent the health benefits received from the activity. Health credits can be based at least on the duration of the activity, level of intensity of the activity and health standards established by a health entity. In some modalities, health credits may also be based in part on the type of activity. In some implementations, the health monitor 100 can be configured to distinguish between different types of activities as well as detect non-human activities (for example, counterfeit activities such as strapping a health monitor to a wheel or fan) in such a way that that health credits are not assigned for said activities detected.

In some embodiments, a health monitor 100 may comprise a smart activity monitor, as depicted in the exploded view of Figure 1A. A health monitor 100 can, for example, comprise a housing that includes a first cover 170 and a second cover 172. The first and second covers can be formed from any of suitable materials including, but not limited to, metals and plastics. or combinations thereof. As an example, the first cover 170 may be a molded plastic and the second cover 172 a corrosion resistant metal. The first and second covers can be fastened together by any suitable means and can form a watertight seal that encloses a power source 105 and electronic circuitry 180 of the health monitor. A clasp or strap 174 can be placed on or attached to a surface of one of the covers so that the health monitor 100 can be attached to, or supported by, a subject or machine (for example, attached by a strap to a wrist, ankle, or end, attached to an article of clothing, strapped, or attached to a movable portion of a machine, such as an exercise machine or wheelchair.)

As shown, electronic circuitry 180, for example, may comprise a combination of circuit elements 182 placed on a printed circuit board. In various embodiments, circuit elements 182, for example, may include a selected combination of integrated circuit (IC) chips, application-specific integrated circuit (ASIC) chips, at least one digital processor, microelectric system devices- mechanical (MEMS), resistors, capacitors, inductors, diodes, light-emitting diodes, transistors, and / or traces of conductive circuits, etc. A microcontroller or microprocessor, for example, can coordinate and manage the operation of the electronic circuitry of the smart monitor and process the activity data, in some ways. In some embodiments, electronic circuitry 180 may further include at least one radio frequency (RF) antenna 185 for use in sending and receiving RF communication signals.

Figure IB shows in more detail an internal circuitry 102 that can be used in a health monitor 100, according to some embodiments. As shown, the health monitor circuitry may, for example, comprise a power source 105, for example, at least one battery or power pickup chip and a power management and activation circuit 150, which provides and manages power to an accelerometer 130, a digital processor 110, memory 120, and a transceiver 140. The processor 110 can be coupled to the drive circuit, the accelerometer, memory, and the transceiver. The processor can be configured to receive and process acceleration data from accelerometer 130, to read and write data to memory 120, and to send and receive data from a transceiver 140. Trigger circuitry 150 can be adapted to detect when the monitor health 100 is not in use, and in response, reducing the power consumption of internal circuitry 102, according to some modalities. The trigger circuit may further be adapted to detect when the health monitor 100 is put into use, and in response, triggers one or more elements of the internal circuitry 102.

In some embodiments, processor 110, for example, may comprise a low power 8-bit processor, which is configured to cause low power in idle mode operation, and capable of operating at multiple million instructions per second (MIPS ) when activated. An example of a suitable processor is the 8051F931 processor available from Silicon Laboratories Inc. of Austin, Texas. Another example of a processor is the nRF51822 processor available from Nordic Semiconductor of Oslo, Norway, although any other suitable processor or microprocessor may alternatively be employed in other embodiments. In some implementations, processor 110 can support radio frequency communications with other devices. A balun (eg BAL-NRF02D3 available from ST Microelectronics of Geneva, Switzerland) can be used to match RF signals between an antenna and the processor, according to some modalities.

Processor 110, for example, can include various types of on-board memory (e.g., flash memory, SRAM, and XRAM) to store machine-readable data and / or instructions, and can be clocked by an internal oscillator or external oscillator. . In some embodiments, the processor, for example, may be clocked by an internal high-frequency oscillator (for example, an oscillator operating at approximately 25 MHz or higher) when the processor is active and processing data and alternatively be clocked by a low-frequency oscillator. frequency (external or internal to the processor) when the processor is in a substantially idle or idle mode. Low frequency processor timing, for example, can reduce the power consumption by the processor during sleep mode. Low frequency timing can be at a frequency that is less than 50% of high frequency timing in some modalities, less than 20% of high frequency timing CD in some modalities, less than 10% of high frequency timing in some modalities , less than 5% of the high-frequency timing in some modalities, less than 2% of the high-frequency timing in some modalities, less than 1% of high frequency timing in some modalities, and even less than 0.1% in some modalities.

In various embodiments, processor 110 can be configured to receive acceleration data from accelerometer 130 and process the received data in accordance with pre-programmed machine-readable instructions that are loaded into and executed in the processor. Processor 110, for example, may be configured to receive analog and / or digital input data and may include on-board analog-to-digital and digital-to-analog converters and on-board timers or clocks. According to some embodiments, the processor can also be configured to receive and analyze cardiac waveform data. In some embodiments, the processor may further be configured to receive power through power management and drive circuitry 150. The processor, in some embodiments, may operate cooperatively with or comprise a portion or all of the power management circuitry 150, and facilitate the activation and deactivation of one or more circuit elements within the health monitor.

In some embodiments, processor 110 can be configured to operate at a number of different clock frequencies. When operating at a low clock frequency, the processor will normally consume less power than when operating at a high clock frequency. In some embodiments, the processor, for example, can be configured to be in sleep mode and operate at a low clock frequency when there is no movement of the health monitor 100, and to cycle through various states of operation when movement of the health monitor 100 is detected. An example of how a processor can be cycled in such a way will now be described with reference to Figure 2A. As an example, when in sleep mode, the processor can test data slower than 10 Hz and draw less than about 30 microamps.

In some embodiments, accelerometer 130, for example, may comprise a multi-axis accelerometer that is configured to detect acceleration along at least two substantially orthogonal spatial directions. Accelerometer 130, for example, may comprise a three-axis accelerometer based on micro-electromechanical systems (MEMS) technology. In some implementations, one or more single axis accelerometers can be used additionally or alternatively. In some embodiments, accelerometer 130 may be configured to provide one or more analog data stream outputs (eg, X, Y, Z data outputs corresponding to each axis of the accelerometer) each representing a magnitude and tare of aceteradon along a respective axis. An example of a suitable accelerometer is the Kionix model KXSC7 accelerometer available from Kionix Inc., Ithaca, New York. Another example of a suitable accelerometer is the LIS2DH accelerometer available from ST Microelectronics of Geneva, Switzerland. Accelerometer 130, for example, can provide analog output data, which can subsequently be converted to digital data, or it can provide digital output data representative of acceleration values.

The accelerometer 130 can be characterized by various parameters. Among these parameters, for example, there may be a sensitivity value and a sample rate value. As examples, the sensitivity of the accelerometer analog can be between about 100 millivolts (mV) per gravitational value (100 mV / G) and about 200 mV / G in some modes, between about 200 mV / G and about 400 mV / G in some modalities, between about 400 mV / G and about 800 mV / G in some modalities, and even between about 800 mV / G and about 1600 mV / G in some modalities. When configured to provide a digital output, the accelerometer sampling rate, for example, can be between approximately 10 samples per second per axis (10 S / sec-A) and approximately 20 S / sec-A in some modes, between approximately 20 S / sec-A and approximately 40 S / sec-A in some modes, between approximately 40 S / sec-A and approximately 80 S / sec-A in some modes, between approximately 80 S / sec-A and approximately 160 S / sec-A in some modes, between approximately 160 S / sec-A and approximately 320 S / sec-A in some modes, and even between approximately 320 S / sec-A and approximately 640 S / sec-A in some modes. It will be appreciated that in some embodiments higher sample rates may improve the quality of measured accelerations.

It will be appreciated that, in some embodiments, an accelerometer 130 can be combined with one or more analog-to-digital converters to provide digital output data representative of the acceleration values ​​at sample rates described above. When digital output data is provided by an accelerometer, the sensitivity of the accelerometer can be expressed in units of bits per gravitational constant (b / G). As examples, an accelerometer that provides digital output data may have a sensitivity of more than about 2 b / G in some modes, more than about 4 b / G in some modes, more than about 6 b / G in some modes, plus of approximately 8 b / G in some modalities, more than approximately 10 b / G in some modalities, more than approximately 12 b / G in some modalities, or even higher values ​​in some modalities.

According to some embodiments, a health monitor 100 may include one or more sensors in addition to motion sensor 152 and accelerometer 130. For example, a health monitor may include at least one physiological sensor 154 (eg, heart sensor, temperature sensor, blood glucose sensor, blood oxygenation sensor, etc.) that is configured to detect at least one physiological parameter of a subject. Examples of physiological parameters that can be detected in some modalities include, but are not limited to, cardiac waveform, skin temperature, core temperature, respiration rate, plethysmography waveform, EKG waveform. , blood oxygenation level, blood glucose level, hydration, blood pressure, etc. An illustrative example of a physiological sensor comprises the AD8232 ECG chip available from Analog Devices, Inc. of Norwood, Massachusetts. Said chip can be combined with electrodes arranged to make contact with the skin of a subject. A physiological sensor 154 can be packaged with a health monitor in some implementations or can be formed as a separate monitor to attach to the subject in a separate location and wirelessly, or via a wired link, transmit data to the health monitor according to a predetermined communication protocol.

In some embodiments, a health monitor 100 may include memory 120 that is external to and accessible to processor 110. Memory 120 may be any of or a combination of the following types of memory: RAM, SRAM, DRAM, ROM, flash memory. . Memory 120, for example, can be used to store and / or save raw data from accelerometer 130 and / or physiological sensor 154, machine-readable instructions for processor 110, program data used by the processor to process data from the accelerometer and / or physiological data and / or activity data representative of an activity. In some embodiments, memory 120 may be additionally or alternatively used to store diagnostic information about the health of the health monitor, eg, battery life, error status, etc., and / or physical parameters about the device, for example, memory size, gravitational sensitivity, weight, battery model, processor speed, operating software version, user interface requirements, etc. In some embodiments, memory can also be used to store information pertinent to a user, for example, user's weight, height, gender, age, training goals, specific work plans, user-specific activity data that can be used to identify an activity performed by the user or process data representative of an identified activity. In accordance with some embodiments, memory 120 can store metabolic equivalent tables (METs), calibration values, and health pattern data that are used to determine health benefit levels for various activities.

In some embodiments, memory 120 may be additionally or alternatively used to store data structures and / or code 5 received from an external device, eg, via a wired or wireless link. The data structures and / or codes, for example, can be used to update one or more data processing applications used by the health monitor. For example, a type of data structure can be data representative of an activity data pattern that can be used to identify a specific type of activity that is not previously recognized by the health monitor, for example, a new activity or a activity that is specific to an individual user of the health monitor. As another example, a data structure may comprise a membership function, described below, defined for a new activity or redefined for an identifiable activity. According to some embodiments, the data structure, for example, may include one or more sample accelerometer traces and physiological data obtained during performance of the activity and / or may comprise identification data (for example, membership functions. ) that result from the processing of accelerometer traces that can be used in an algorithm executed by the health monitor 100 to identify activity. Also, in some embodiments, memory 120 can be used to store updates and / or replacements for algorithms executed by the health monitor. The stored data structures and algorithms, for example, can be used to reprogram and / or expand the functionality of the health monitor 100 to identify new activities or activities not previously recognized by the health monitor and / or improve the accuracy or confidence of the data. results calculated for identified activities.

In some embodiments, memory 120 can also be used to store calibration and / or conversion data that is used by processor 110 to characterize detected activities. Calibration data, for example, can be used to improve the accuracy of detected activity parameters (eg, stride length, speed) and / or improve the accuracy of fitness metrics for detected activities. Conversion data, for example, can be used to convert a detected activity into an amount of human energy expended, for example, calories burned, metabolic equivalents, etc.

According to some embodiments, a health monitor 100 may include a transceiver 140 and / or one or more data communication ports (eg, a USB port, an RF communication port, a Bluetooth port) for communication of data. data between the health monitor and one or more external devices such as a computer, tablet, cell phone, portable communication device, data processor, sensor, other smart sensor, or versatile sensor, any of which can be configured to communicate with other similar devices on a network such as a world network or a local area network. A health monitor 100, for example, can be configured to communicate via transceiver 140 through a wired or wireless port to any device or combination or devices selected from the following list: a personal computer, laptop, tablet, PDA, watch, MP3 player, pod, mobile phone, medical device such as a blood glucose meter, blood pressure monitor, or InR meter, a electronic interactive gaming device, intelligent training equipment and a car system. Data retrieved from memory 120 or stored in memory 120, for example, can be communicated between the health monitor 100 and an external device via the transceiver 140. In some embodiments, the data transmitted from the health monitor 100 can be configured to route to a data services device adapted to process data received from a health monitor.

In some embodiments, power for the internal electronics of a health monitor 100 can be provided by at least one battery and can be managed by a power drive and management circuit 150. The battery can be small, for example, a battery. button cell type, and may, for example, comprise one or more lithium type batteries that can be recharged or replaced. As just one example, a 3-volt single lithium coin or button cell battery having a capacity of approximately 230 mAh can be used (model CR2032 available from Renata SA of Itingen, Switzerland). Another embodiment of a health monitor may include one or more Model CR1616 batteries, although any suitable type of battery can alternatively be used in various modes. In some embodiments, a health monitor may include power generation or power harvesting hardware (for example, a piezoelectric material or electrical generator configured to convert mechanical motion into electrical current, a solar cell, an RF, or thermal converter). The energy that is generated on board can be stored in a battery or charge storage component such as a super capacitor. In some implementations, the generated electrical current can be provided to a storage component via a diode bridge. An example of a suitable energy harvesting device is a MEC225 micro-energy cell available from Infinite Power Solutions, Inc. of Littleton, Colorado. In some embodiments, the power generation components can be used in combination with a rechargeable battery as a power source for a health monitor 100. A voltage regulator chip (for example, TPS78001 available from Texas Instruments of Dallas, Texas) can be used to determine power from at least one power source prior to supplying power to the components of a health monitor, according to with some modalities.

In some implementations, the power drive and management circuitry 150 may include a motion sensor 152 which, in combination with power drive and management circuitry 150, identifies when a health monitor 100 is moving in a certain way. which can be representative of an activity to be monitored. Power management and activation circuitry 150, for example, may comprise logic and control circuitry to enable, disable, reduce, and / or increase power to various circuit elements shown in Figure IB. The control and logic circuitry for the power management and activation circuitry, for example, may comprise machine-readable instructions and use hardware from processor 110, or it may comprise machine-readable instructions and use hardware from a application specific integrated circuit.

In some embodiments, motion sensor 152 may comprise one or more force-sensitive switches, for example, a piezo element configured to generate an electrical signal representative of an amount of acceleration that a health monitor experiences. In other embodiments, the motion sensor 152 may additionally or alternatively comprise one or more contact switches that close a circuit or open a circuit when the health monitor is subjected to acceleration, eg, a shock switch. Triggering, for example, can be initiated when a frequency of the switch closures exceeds a preset value. In other embodiments, sensor 152 may additionally or alternatively comprise one or more force-sensitive contact switches that close only when a health monitor is accelerated in excess of a preselected value.

Figures 1C through ID illustrate an exemplary embodiment of a health monitor 103 that includes light emitting diodes (LEDs) 184 for communicating information to a user. The illustrations show a packaged device in a plan view (Figure 1C) and an elevation view (Figure ID). A health monitor can be substantially round, rectangular, square, elliptical, or it can be in the shape of a multi-sided polygon. A health monitor may have a larger dimension D that is between about 5mm and about 40mm in some embodiments, and a thickness T that is between about 1mm and about 10mm in some embodiments. The size of the health monitor 103 can be largely determined by its power source, for example the size of its battery. A health monitor may have a shape fact that substantially matches a shape of the power source. For example, a health monitor may be shaped similar to the shape of a power source, and have a volume that is up to 20% larger than the volume of the power source in some modes, up to 50% larger than the volume of the power source in some modes, up to 100% larger than the volume of the power source in some modes, and even up to 200% larger than the volume of the power source in some modes. LEDs can be placed in a line, a circle, an ellipse, or any other shape. A first cover 170a and second cover 172a can be adjusted or screwed together, to form a seal with a polymer gasket 171a. In some embodiments, a clasp or strap may include the polymer gasket 171a, so that the device can be easily changed from a clasp to a strap fastener. In some implementations, a clasp or strap may be removably attached to the first or second covers. In still other embodiments, a clip or strap can hold the health monitor.

According to some embodiments, a health monitor 103 may be configured to recognize one or more sequences of keystrokes and / or movement gestures (eg, moving the device in a pattern of Figures 8A to 8C, a circle pattern, a linear pattern back and forth), and activate the LEDs to communicate information responsible for the sequence of keystrokes or gestures detected. A sequence of keystrokes or gestures can correspond to a particular information query, to which the health monitor can respond with the appropriate information. According to one embodiment, the health monitor can be exploited in a particular way, and in response activate a number of LEDs to indicate that a user has reached an approximate percentage of an activity goal (e.g., 8 out of 10 LEDs for indicate approximately 80%). Information on progress toward one or more activity goals can be communicated via the device (e.g., walk 30% of a 4.8-kilometer goal, run 60% of a 12.8-kilometer goal, swim 90% of a goal of 60 laps 60, achieved 70% of activity with creditable health benefits for one week, etc.) LEDs can also be used to communicate other responsible information for particular sequences of keystrokes or gestures, for example, battery life, pace comparison (before or after the best pace for an activity), heart rate, calories burned, etc. Although LEDs can be used to communicate information to a user, a small liquid crystal display can be used in some embodiments instead of, or in addition to, LEDs.

Figures 1E through 1G illustrate exemplary modalities of health monitors 104, 160 according to some modalities. A health monitor 104 can be formed in the form of a band having a central compartment 112 and clamping features 114. The central compartment 112, for example, can be attached to a flexible band 106 that can be strapped around a limb of a subject or attached to a subject in a suitable way (e.g. strap around a region of the ankle. , a wrist or an arm). The central compartment can store the electronic components 180, 185, 182 (see Figure 1A) of the health monitor. In some embodiments, health monitor 104 may include one or more LEDs 184 to convey information to a user, as described above. In some embodiments, the central compartment may be flexible (for example, formed of a polymer or elastomer) and the monitor electronics may be mounted on a nexibie substrate with flexible interconnections. In some embodiments, the central QMjHitento

112 and band 106 are semi-rigid and fastening features 114 attach to flexible bands that are used to strap the device to an extremity. The fastening features can be holes, snaps, buckles, snaps, or any other suitable fastening feature. The health monitor 104, for example, may have a length L between about 10mm and about 250mm or longer, a width W between about 5mm and about 30mm, and a thickness T between about 1mm, and about 10mm. mm. In some embodiments, the health monitor 104 may have a curved profile along its length so that it can better conform to the contour of a limb.

According to some embodiments, a health monitor 104 can include at least one light source 186 and at least one photodetector 187. At least one light source and the photodetector can be used, for example, to detect one or more plus physiological parameters of a subject, for example, blood oxygenation level, plethysmography waveforms, blood glucose level, etc. In some embodiments, light source 186 may comprise a high brightness infrared (IR) photodiode and a shorter wavelength photodiode. In recent years, progress in indium gallium nitride LED technology has produced devices with low junction voltage and increased radiated intensity. By using InGaN technology, and applying the energy management techniques described below, a health monitor capable of measuring heart rate can be provided that can run for multiple months on small silver oxide batteries, with a form factor such as that of a typical bandage or wristwatch. Photodetector 187 can be any suitable photodetector, and can be mounted to detect light from the light source that is scattered or reflected from the subject.

Figure 1G depicts a patch type health monitor 160, according to some embodiments. A patch-type health monitor can be flexible, thin, and in the form of an adhesive patch and may include some discarded parts in some implementations. For example, a patch-type health monitor can comprise a flexible cover 160-1, which can be formed of silicone. Health monitor 160 may include a first flexible printed circuit board (PCB) 160-3 to which at least one monitor (accelerometer and / or heart sensor) may be connected. Power can be provided by one or more batteries (for example coin button batteries such as CR1616 batteries). A patch type health monitor can include multi-component electrodes for electrically contacting a subject to detect cardiac waveforms. For example, the electrodes may comprise first conductive pads 160-4 (eg, copper pads) with a diameter between about 2mm and about 6mm and a thickness between about 0.1mm and about 1mm, according to some embodiments. . The first conductive pads can be attached to a second flexible 160-7 PCB, which can be less than 1mm thick. Conductive tapes 160-6 can be used to adhere the first conductive pads 160-4 to the second flexible PCB 160-7. In some embodiments, the second conductive pads 160-8 (for example, silver chloride pads) can be contacted with the second flexible PCB and further contacted with the adhesive hydrogel pads 160-10 (for example , adhesive material 9880 available from 3M Corporation of St. Paul, Minnesota) that are positioned to make contact with the subject in separate locations. In some implementations, a first adhesive layer 160-5 (for example, adhesive LSE 96042 available from 3M Corporation of St. Paul, Minnesota) can be used to adhere the first layer of flexible PCB to the second layer of flexible PCB 160-7 and an intermediate hydrocolloid layer 160-9 (eg, hydrocolloid 9943 available from 3M Corporation of St. Paul, Minnesota). In some embodiments, a second layer of biocompatible adhesive 160-11 (eg, adhesive 2475P available from 3M Corporation of St. Paul, Minnesota) can be used to adhere the patch-type health monitor 160 to a subject. Cardiac signals can be transported to an ECG chip that locates the first flexible PCB 160-3 by circuit pads extending through hydrogel pads 160-10, second conductive pads 160-8, second flexible PCB 160-7 , and conductive tape 160-6, and first conductive pads 160-4.

In some embodiments, a patch type health monitor 160 may have a thickness between about 1mm and about 4mm and a length between about 30mm and about 100mm. The width of the monitor can be between about 10mm and about 30mm. In accordance with some embodiments, the patch monitor can be configured to adhere to a subject's torso close to the heart, for example in the vicinity of a subject's second to fourth ribs. The device can be waterproof, and can be configured to adhere to a subject for a period of one or more days. In some implementations, the adhesive biocompatible with the device may allow adhesion for a period of time of more than four days or more. In some embodiments, the two hydrogel pads 160-10 can be separated a distance between about 15mm and about 50mm. In some implementations, the separation distance can be between 20mm and 30mm.

In accordance with some aspects, the hydrogel pads 160-10 and a second biocompatible adhesive layer 160-11 can be replaced when the health monitor is removed from a subject and reattached. For example, a subject can remove the health monitor 160 at any time and reattach it at a later time. Although the health monitor can be configured to be used continuously, a subject may wish to remove it daily in some THINGS, weekly in some cases, monthly in some cases, or at any suitable interval. After removal, the 160-10 hydrogel pads and the second biocompatible adhesive layer 160-11 can be removed from the health monitor and replaced with new pads and adhesive.

In some implementations, a heart monitor of a health monitor 160, as shown in FIG. 1G, may be configured to capture cardiac waveform data continuously in some modes of operation and intermittently in some modes of operation. The types of information that can be determined by a cardiac waveform health monitor 160 may include heart rate, beat interval (IBI), heart rate variability (HRV), PQRST waveform profile, and breathing rate.

Although health monitors are depicted as a single device packaged in Figures 1E through 1G, in some embodiments a health monitor may comprise two or more separate sensors attached to or supported on a subject that are configured to communicate over a small area network. or body local area network (body LAN). One or more of the sensors can be configured as a versatile sensor, as described in US Patent Application No. 13/840098 entitled Versatile Sensors with Data Fusion Functionality, which was released on March 15, 2103 and is incorporated herein by reference in its entirety. In some embodiments, the additional sensors can detect other physiological data and communicate representative data to a health monitor 104.

Referring now to Figures 2A through 2B, in some embodiments, the power management and activation circuitry 150 for a health monitor can be configured to cycle the device through a plurality of operating states, as depicted in the drawings. The operating state to which the health monitor cycles, for example, may depend on motion detected by the power management and trigger circuitry, or data received from another sensor on a body LAN. Some of the operating states, for example, may be the energy conservation, low energy, and no energy states.

According to some embodiments, there may be a low-energy or no-energy state 210 and one or more energized operating states 230, 250, 270. The energized operating states may, for example, include various energy conservation states. As illustrated, a health monitor can be moved from any state to any other state along paths 220, 240, 260, 280, 215. In some embodiments, there may be trajectories in addition to or in place of those shown in Figures 2A to 2B, for example, directly from the step detection state 270 to the activation qualification state 230.

In some modes, when a subject is inactive (for example, the Salid monitor is displayed with no movement or movement less than the first preset limit or threshold), the monitor can be operated in a sleep mode 210. In some modes, the mode resting may not consume energy. In other implementations, however, the sleep mode can draw low power, for example, drawing about 1 microamperes or less. Low energy may, for example, be provided to motion sensor 152 to detect movement of a health monitor 100. In some embodiments, no combination of or all of motion sensors 152, physiological sensor 154, the processor 110, memory 120, and transceiver 140 while in sleep mode. According to some embodiments, when sufficient movement has been detected by the motion sensor 152, the health monitor can be moved to an activation qualification state 230.

In some embodiments, sufficient movement to move the health monitor out of the idle state 210 may be detected by the motion sensor 152 in accordance with an amplitude and / or frequency of a signal from the motion sensor. For example, when the motion sensor comprises one or more piezoelectric elements, a signal greater than a predefined signal value can be used to identify sufficient movement of the health monitor and move the monitor to a wake-up rating state 230. In another example, motion sensor 152 may comprise one or more contact switches, and a predefined number of switch openings, or closures per predefined time interval may be used to identify sufficient movement of the health monitor and move the health monitor. health to activation qualification status 230.

In some embodiments, when in the wake-up rating state 230, low current levels (eg, less than about 30 microamps) can be drawn by a health monitor 100 in the system. Power, for example, can be provided to the systems accelerometer and processor so that accelerometer data can be processed while in the 230 wake-up qualification state. In some embodiments, power can also be provided to memory 120 while it is in the wake-up qualification state. In some implementations, energy can be provided to one or more physiological sensors in an activation qualification state.

In some embodiments, when in the wake-up rating state 230, the system may be clocked at a low speed to conserve power. For example, the processor 110 can be clocked at a low frequency and / or the accelerometer can be tested at a low frequency (eg, less than about 10 Hz) to obtain and process data from the accelerometer. The data, for example, can be processed to determine whether a second predefined threshold has been exceeded. In some embodiments, the second threshold may comprise a predefined amount of force or acceleration at which a health monitor 1QQ 56 50016 (6, and may include an additional parameter, for example, a frequency of events of measured values ​​that exceed the amount acceleration preset, a heart rate. When the second threshold is crossed, the health monitor, for example, may move to a 250 step or activity rating state. If the second threshold is not crossed within a predefined time period, the health monitor, for example, you can return to sleep mode 210.

In some embodiments, when in a step qualification state 250, more power is provided to the accelerometer 130 and / or processor 110, so that a greater amount of motion data data processing can occur. The accelerometer and / or processor, for example, can be operated at a higher timing frequency so that more data can be obtained from the accelerometer and processed by the processor during a given time interval, compared to the state activation qualification 230. The current drawn by a health monitor in step rating mode can be on the order of a few hundred microamps, for example between about 100 microamps and about 500 microamps in some modes. The data collection speed, for example, can be increased to normal operating speed in a 250 step qualification mode. Accelerometer data, for example, can be tested at several hundred Hertz (eg, 256 Hz) or higher values. In some implementations, a larger amount of physiological data can be obtained and processed in a step qualification state.

In some embodiments, when in step qualification mode 250, a health monitor 100 may analyze the detected acceleration data to determine if the data is representative of an activity which may be an activity recognizable to monitor 110, for example , walk, swim, jump rope, etc. In some modes, if the processor determines that activity may be discoverable, the Health Monitor may move to a 270 step detection state. In some modes, if it determines that there is insufficient activity in a predefined amount of time which can be recognized by the processor, then the health monitor can return to activation qualification mode 230.

According to some embodiments, when the step detection mode 270 is executed, a health monitor can be placed in full operation. In this mode, power, for example, can be provided to the transceiver 140, in addition to other operational components so that communications with an external device or additional sensors can be carried out. In some embodiments, normal operating timing frequencies and test speeds can be used to operate the accelerometer, one or more physiological sensors, and processor so that full data processing and activity detection can be carried out. Physiological sensors may or may not be fully energized in a step detection mode 270. The current draw in step detection mode 270, for example, can be several hundred microamps, eg, between about 200 microamps and about 600 microamps.

In some embodiments, each state of operation other than sleep mode 210 may include a provision to return a health monitor directly to sleep mode, for example, along state paths 215 as indicated in FIG. 2A. . For example, each operating state 230, 250, 270 can be configured to return a health monitor to sleep mode 210 if there is a termination of incoming data or of processed data parameters related to the incoming data. In some embodiments, a health monitor can be additionally or alternatively configured to return to sleep mode after a technical failure or system crash eg processor freeze. A return to sleep mode, for example, can be used to restart a health monitor.

According to some implementations, the energy conservation for the health monitor may additionally or alternatively be based on cardiac data, as shown in Figure 2B. Energy conservation methods based on cardiac data can be run in parallel with or in combination with energy conservation methods based on motion data. As just an illustrative example, the motion data can be analyzed by a system processor to determine that a subject is in an inactive state (eg, sitting, lying down, driving a vehicle, etc.). A health monitor can then determine that at least the energy to a physiological motion detector and sensor can be reduced. In some embodiments, a motion sensor may enter a sleep mode 210, and a heart sensor configured to detect a cardiac waveform from the subject may enter a sleep mode 212, as shown in Figures 2A through 2B. To extend battery life, the heart sensor can cycle through one or more low energy modes of operation from a resting state, as shown in Figure 2B while the subject is inactive.

For example and with reference to Figure 2B, in a first mode (mode A), a cardiac sensor can sleep (rest state 212) for a period of time between each heartbeat of a subject and wake up in time to capture a waveform (QRS detection status 232) of a heartbeat (for example, a QRS complex) for further analysis. The QRS waveform, for example, can be analyzed for arrhythmia, heart rate variability, and / or respiration rate, according to some implementations. In some modalities, respiration rate can be determined from an R wave envelope over multiple cardiac cycles. In a first mode, the entire waveform captured by the circuitry can be used to test γ analyze the cardiac waveform during the QRS complex alone.

In a second mode of operation (mode B), a heart monitor can sleep for a period of time between each heartbeat of a subject and wake up in time only to determine a point or time of an R wave (state of detection of R 252). In some implementations, the cardiac signal can be fed to a comparator configured to detect a threshold that crosses or changes in steepness of the R wave. The comparator may require less power to operate than the circuitry needed to capture and analyze a portion of the cardiac waveform.

In a third mode of operation (not shown in Figure 2B), a heart monitor can operate continuously to capture a complete cardiac waveform for multiple beats. A continuous mode of operation can be run periodically to time the P waves, in some modalities, and determine a sleep interval between heartbeats. In some implementations, a continuous mode of operation can be executed when a subject becomes active, or it can be executed when a subject's activity is moderate and / or vigorous. In some implementations, a user can command continuous monitoring of a cardiac waveform regardless of user activity by a sequence of pulses on the patch or motion sensor that can be detected by the motion sensor, processed, and recognized by the processor. of the health monitor.

In additional detail, Figures 2C to 2D are illustrative examples of methods for capturing and processing cardiac signals that in a health monitor can be configured to be implemented. With reference, an example of a cardiac waveform is depicted in Figure 2E, illustrating the P, Q, R, S, and T portions of the cardiac cycle.

According to some embodiments and with reference to FIG. 2C, an energy conservation method 204 based on the cardiac waveform implemented in a health monitor may comprise receiving 226 an ECG signal over a period of time and analyzing the results. data to determine 228 a PQRST waveform time for a subject. According to some modalities, a health monitor can wake up 232-1 a heart monitor prior to a P wave, record 232-3 a QRS complex, and process 232-5 the QRS complex to obtain cardiac data (for e.g. arrhythmia, heart rate variability, and / or breathing rate)). The health monitor can wake up the heart monitor while keeping the motion sensor in a sleep mode, according to some modalities. The heart monitor can be activated between approximately 1 ms and approximately 100 ms before the P wave, according to some modalities. In other modalities, the heart health monitor can be activated sooner or later. The health monitor can determine 232-7 if the recorded QRS waveform was valid (for example, not clipping or missing a portion due to an incorrect time gating technique). If the QRS wave is valid, the health monitor may enter a sleep mode 232-9 during which the energy to the heart monitor is reduced or closed until the next P wave. If the waveform is determined to QRS is invalid, the health monitor may turn on the heart monitor and related electronics to receive 226 a full ECG signal over a period of time so that an appropriate rest or alert interval can be determined.

In some implementations and with reference to FIG. 2D, an energy conservation method 206 based on the cardiac waveform implemented by a health monitor may comprise receiving 226 an ECG signal over a period of time and analyzing the data to determining 228 a PQRST waveform time for a subject. A health monitor can wake up the heart monitor 252-1 before an R wave (for example, between approximately 1 ms and approximately 100 ms before the R wave, according to some modalities, although it can be used in others at other times modalities). In this mode, a comparator can be put on alert and used to process an input cardiac signal instead of using A / D circuitry and digital processing of a QRS complex. By detecting a threshold crossover point or positive to negative tilt shift of the R wave, the health monitor processor can determine 252-3 an RR interval between heartbeats. These data can be used to assess an interval between heartbeats (heart rate) and / or heart rate variability. The health monitor can evaluate 252-5 if the determined R wave point is valid (eg, occurring within an expected time window). If the determined R wave point is determined to be valid, the heart monitor can enter a sleep mode 232-9. If the R wave point is determined to be invalid, the health monitor may turn on the heart monitor and related electronics to receive 226 a full ECG signal for a period of time.

Figure 2F represents the types of data that can be obtained from a motion sensor and cardiac waveform sensor, according to some modalities. For example, a motion sensor (eg, accelerometer) can be used for various activity measurements to obtain motion waveforms that correspond to a type of activity performed by a subject, as will be described in more detail below. Similarly, a cardiac sensor can be used to obtain cardiac waveforms for a subject during different levels of activity and during periods of inactivity. In some embodiments, data from different sensor types can be used separately in some examples and can be combined in other examples to determine different fitness metrics for an individual, some of which are represented in the drawing.

As just one example, motion waveforms can be analyzed to determine a body position of a subject to which a health monitor is attached. Motion waveforms can indicate that the subject is in a prone position. Analyzes of heart waveforms or RR interval data obtained during the same time period may indicate, for example, a lower than normal heart rate, which can be determined by the health monitor as a resting heart rate ( RHR, for its acronym in English). In some embodiments, the combined movement and cardiac data can indicate that the subject is inactive, and can be used, for example, to determine a basal metabolic rate (BMR) for the subject. Additionally, heart rate variability (HRV) data along with motion data from a motion sensor can be used to assess a subject's sleep quality. For example, low heart rate variability for a subject may indicate increased stress for the subject. In some modalities, low HRV can be determined from a ratio of the low frequency (LF) and high frequency (HF) spectral energies of a cardiac waveform. The spectral energies of a cardiac waveform can be determined from a Fourier transform or FFT of the cardiac cycles. The spectral energy of LF can be on a scale between about 0 and 150 Hz, and the spectral energy of HF can be on a scale between about 150 Hz and about 400 Hz. An increased ratio of LF / HF for a subject over a value Average during a resting state may indicate an increased level of stress for the subject and a reduced quality of sleep. Additionally or alternatively, excessive movement during sleep may indicate a reduced quality of sleep. Other fitness metrics that can be determined from movement and heart data are described below.

III. Data processing

This section provides an overview of data management and data processing paradigms that can be implemented with various modalities of a health monitor. In some implementations, activity data processing can be performed on one or more processors of a health monitor. With reference to the block diagram of FIG. 3, an illustrative example of a data management architecture 300 for a health monitor is shown, according to some embodiments. In some embodiments, the data handling architecture may be implemented at least in part in a processor 110 that is specially adapted with appropriate machine-readable instructions.

The data handling architecture 300 for a health monitor may comprise a data preprocessor 305, a characteristic generator 310, a buffer 325, an inference engine 320, and at least one activity engine 340-1, of according to some modalities. In some embodiments, the data management architecture may further include a multiplexer 330 and data service 360. In some implementations, any 305 preprocessor, 310 feature generator, 320 inference engine, 330 multiplexer, and 340-1-340-n activity engines may be represented in whole or in part as machine-readable instructions operable on the processor 110 which, when executed, adapts the processor to perform a respective functionality as described below, or as implemented in alternate embodiments. Additionally, any 305 preprocessor, 310 feature generator, 320 inference engine, 330 multiplexer, and 340-1-340-n activity engines may be additionally or alternatively represented in whole or in part as hardware configured to perform a respective functionality or a portion of the respective functionality.

It should be appreciated that other modalities may include few data processing elements, additional or different elements than those shown in Figure 3. For example, a data post processor can be added before or after data services 360. Furthermore, it should be appreciated that, in some embodiments, one or more depicted items may be combined into a single unit providing equivalent functionality of both separately depicted units. One or more components depicted in Figure 3 may be implemented in hardware (for example, as field programmable gate arrangements, digital signal processor, and / or an application specific integrated circuit) and / or a combination of hardware and readable instructions. by machine (for example, software) that can be run on a digital processor.

As illustrated in FIG. 3, in some embodiments, an accelerometer 130 may be configured to output a motion data stream 133 of representative values ​​of acceleration detected by the accelerometer along at least one direction of motion. Motion data stream 133, for example, may comprise acceleration values ​​representative of the acceleration measured along representative X-, Y-, and Z axes of motion defined on accelerometer 130. In some embodiments, the current Data 133 of the acceleration values ​​can be provided for both a feature generator 310 and a processor 305. In some implementations, supplemental data 302 (eg, physiological data such as any or a combination of cardiac data, respiratory data, blood oxygenation, or blood glucose data, etc.) received from one or more physiological sensors, are can be provided to processor 110, and can be provided to feature generator 310 and / or preprocessor 305. According to some embodiments, the preprocessor 305, for example, preprocesses at least the motion data stream 133 from the accelerometer to produce a stream of acceleration derivative values ​​that are provided to the feature generator 310. In some embodiments, Acceleration derivative values ​​may not be calculated. Any of the data can be filtered, for example to reduce noise components or select specific frequency components for analysis.

In some implementations, supplemental data 302 may be provided to one or more processor 110 devices. For example, supplemental data 302 may be provided to any one or a combination of preprocessors 305, feature generator 310, and inference engine 320. In some implementations, preprocessor 305, for example, may preprocess a plethysmography waveform or cardiac waveform to generate data characteristic of a subject's heartbeat, which can be provided to characteristic generator 310. Supplementary data 302 they can also be provided to buffer 325, in some embodiments. Supplemental data 302 can be analyzed to provide additional information about an activity (for example, processor 110 can be configured to analyze cardiac data to determine a heart rate, for example, from which an intensity level at which an activity is performed can be calculated) and / or to verify a type of activity performed (for example, distinguish between riding an exercise bike, riding a flat-terrain bike or riding a bike up a hill).

In certain embodiments, the feature generator 310 may process the received data stream 133 of acceleration values, received supplemental data, and received pre-processed data (eg, a stream of acceleration derivative values) to produce one or more features. features 312 that can be provided to an inference engine 320 and optionally to a buffer 325. Characteristic features, for example, can be used by the inference engine 320 to identify one type of activity detected by the sensors from a plurality of activity types. In some embodiments, upon identification of an activity type, the inference engine 320 may provide a control signal 322 to a multiplexer 330 to route the characteristic features 312 to an appropriate activity engine 340-1, 340-2,. .., 340-n for additional data processing. Each activity engine, for example, can be configured to process received characteristic features according to specific activity algorithms, (e.g. algorithms for running, walking, swimming or cycling, etc.) to calculate one or more descriptive parameters of the activity. Illustrative examples of one or more parameters include, but are not limited to, a measure of activity intensity or pace, an estimate of energy expended during the activity, an estimate of the distance traveled during the activity, and a duration of the activity. .

In some embodiments, buffer 325 can be used to temporarily hold data representative of an activity while inference engine 320 identifies an activity. For example, once activity has been identified, data can be routed from the buffer to the appropriate activity engine (s). Such temporary data buffering prevents loss of activity data during initial identification of an activity or transitions from one activity to another.

According to some embodiments, once the activity has been identified, the data from the feature generator 310, which may include raw data from the accelerometer, can be routed to an appropriate 340-m activity engine (m corresponds to the value of a selected motor 1, 2, 3, ... n). Each of the specific activity data processing engines, for example, can use any example or combination of acceleration derivative data D_n, one or more accelerometer trace data, physiological data, and characteristic feature data to determine a value of one or more activity-related parameters (eg, speed, distance, intensity, number of steps, etc.). In some embodiments, physiological data can be used in combination with movement data to determine an intensity of activity and / or other parameters related to an activity.

In some embodiments, a selected activity engine, inference engine, or post-processing element (not shown) may additionally or alternatively determine an energy expenditure (eg, calorie burning rate) for the activity. According to some embodiments, energy expenditure can, for example, be determined at least in part from a look-up table of metabolic equivalents (METs) for activity. The look-up table, for example, can comprise a list of metabolic equivalents where each entry can be related to one or more activity intensity parameters, for example, step speed, speed, heart rate, etc. A MET look-up table for each activity that can be performed by a human, for example, can be stored in the memory 120 of the health monitor. In some embodiments, the look-up tables for METs may be specific to a user, eg, specific to the sex, weight, age and / or height of the user. In some modes, during the course of an activity session, a health monitor can determine and record calorie burning rates as a function of time and can also calculate a total number of calories burned as a function of activity time. In some modalities, a health monitor can additionally or alternatively calculate and / or store other data, for example, type of activity, maximum speed, average speed, distance traveled, number of steps, maximum calorie burning rate, burning rate average calories, time of day etc.

As an example of motion data processing and referring to FIG. 4A, an inference engine 320 can identify an activity represented by the data in the figure as you go for a walk. Accordingly, the multiplexer, for example, can be configured to send acceleration derivative data D_n to a walking activity engine 3401. The walking activity engine 340-1, for example, can be configured to determine a distance T_c in the data between a large peak 410, which corresponds to a heel strike, and a successive smaller peak 420, which corresponds to a one step toe lift off. Distance

T_Q for example, it can represent the contact time of the foot with the ground, from which the speed of the walk can be determined. An example of a method for determining walking speed from one foot contact time is described in US Patent No. 4,578,769, which is incorporated herein by reference in its entirety.

In various embodiments the preprocessor 305 and feature generator 310 can preprocess the raw accelerometer data and / or received supplemental data 302 in any suitable way, for example, filtering the data, calculating the derivative values, compressing the data, generating sets of data, package data, etc. Some examples of data preprocessing by preprocessor 305 and feature generator 310 are described in US Patent Application No. 13/840098 entitled Versatile Sensors with Data Fusion Functionality, referenced and incorporated by reference above. In various embodiments, the feature generator preprocessor and process received raw data to reduce the processing load on the inference engine 320 and activity engines 340.

According to some implementations, the inference engine 320 may be additionally or alternatively configured to identify activities as unrecognizable (eg, activities that can be performed by a subject but for which the health monitor is not programmed to recognize) as well how to identify activities that are not performed by a human (eg, non-human activity that can be performed by a machine or animal). The inference engine 320 can identify activities based on data received from one or more of the preprocessor 305, feature generator 310, and activity engines 340. In some embodiments, the raw acceleration data may be additionally or alternatively provided to the engine. inference 320 for identification of detected activities. In some modalities, a health monitor may classify activities into two or more classifications, for example, (i) recognizable activities, (i) unrecognizable activities that are likely to be performed by the subject, and (Ni) non-human activities or non-animated (for example, activities performed by a machine or animal). A health monitor can be configured to later present an activity identified as unrecognizable to a user for identification by the user, in some modalities, and the health monitor can be modified for later recognition of the activity (for example, the health learns and builds one or more membership roles that encompass the characteristic traits for the activity).

In some embodiments, an inference engine may receive a set of characteristic features 312 (representative of the unrecognizable activity) that cannot be identified as corresponding to any of a plurality of activity types recognizable by the inference engine. In this case, the data handling architecture, for example, can be configured to provide the data from the feature set 312 directly to the data service 360 ​​for further analysis and determination of a type of activity related to the feature set ( for example, identifying activity by user assistance, or by comparing with online libraries of identified activities with similar sets of traits). Feature set data 312 can be routed by multiplexer 330 from buffer 325 to data service 360. In addition, the data handling architecture can be additionally or alternatively configured to receive machine-readable instructions and reconfiguration data 362 back from the appropriate data service 360 ​​to reconfigure the inference engine 320 and / or add or modify an activity engine 340 to subsequently identify a type of activity that corresponds to the previously unrecognizable set of traits. In this way, the recognition of activity by the health monitor can, for example, be improved or reconfigured at any time.

In some embodiments, an unrecognizable activity may be reported to a user for later identification by the user at a time shortly after the activity ends. For example, an unrecognized activity can be reported when the user reviews a log of monitored activities through a computer. The user, for example, may be notified that an unrecognizable activity occurred at a specific date and time and for a period of time, and then the user may be asked to identify the activity. In some embodiments, the user can then identify the activity, which in turn will relate a set of characteristic features 312 to the previously unrecognized activity. According to some embodiments, a list of possible activities with a similar calorie burning rate can be presented to a user, so that the user can select an activity from the list. The list may comprise the activities named in ”2011 Compendium of Physical Activities: A Second Update of Codes and MET Values, by Barbara E. Ainsworth et al., Published by the American College of Sports Medicine, the full contents of which are incorporated herein by reference. Improving the inference engine 320 and / or activity engine 340, for example, may comprise transmitting new data structures and / or code to processor 110 for use in recognizing new activities.

In some implementations, an inference engine 320 can be configured to recognize one or more activities that are not performed by a human. Such non-human activities can be related to mechanical movement of a machine, for example. An example of such activity may be a cyclical movement of a fan, where a health monitor can be strapped to a fan blade. Another example may be the movement of an automobile, motorized vehicle, or motorized machine (eg, riding a vehicle over rough terrain, a carnival ride, or getting on a train). Another example may be movement of a bicycle wheel, for example, a monitor fixed to the spokes of a bicycle. Other examples of activities not performed by a human can be a movement in a washing machine or drying clothes. Another example may include the walking or walking motion of a dog or horse. In some cases, activities not performed by a human, for example, may be recognized to avoid erroneous physical activity crediting for a subject, for example, in connection with health insurance incentive programs or prescribed physical exercise. The inference engine can recognize non-human activities as characteristics not capable of being achieved by a human, for example precise repetitive motion or excessive acceleration values.

According to some embodiments, the inference engine 320 is configured to further receive error signal indications 345 from each activity engine 340-1 ... 340-n. An error signal, for example, can represent a confidence level for an activity processed by an activity engine. For example, a first activity motor 340-1 may be a walking activity motor, and a second motor 340-2 a running motor. When the user walks, for example, a walking cadence may be below a threshold criterion for running, and an emission error signal from the Running Activity Engine 340-2 may be high, or a confidence level may be be short. The error signals emitted, for example, can be used by the inference engine 320 to help identify an activity and / or transitions between activities. In some cases, all activity engines can process feature data substantially simultaneously, while in other cases only a selected number of activity engines or activity engines can be selected to process feature data. According to some embodiments, the presence of a large error signal for a currently identified activity, for example, can cause the inference engine 320 to re-identify an activity for newly received data.

In some embodiments, the characteristic parameters of an activity that include any data calculated by a 340-m activity engine can be provided and managed by the 360 ​​data services for subsequent presentation to the user, on-board storage, on-board analysis, and / or storage on a remote storage device. Additional data (for example, date, time and duration of the activity) can be provided to the 360 ​​data service. The data can be formatted or packaged by the data service in any suitable format, and can be formatted to include header information. Data service 360, for example, may comprise on-board data storage, eg, memory 120, and a transceiver 140 for transmitting the data to a remote device, such as a computer. In some implementations, data service 360 ​​may comprise a temporary buffer that is measured to hold activity data for a time interval between about 5 minutes and about 30 minutes or longer periods. In some embodiments, the data service may additionally or alternatively include an application operating on a remote computer or a remote server configured to receive data from the health monitor and further record and / or process the received data. In some implementations, data 362 from data service 360 ​​may be retrieved by processor 110 for further processing, for example, to determine health benefit levels of the recorded data.

In some embodiments, a health monitor 100 may be additionally or alternatively configured to store the user's goals in memory (e.g., number of steps per day, distance traveled, a duration of exercise during a specific activity, a number of credits of health over a period of time, etc.). Processor 110 may be configured to determine progress toward the user's goal (s) based on broadcast from activity engines, according to some embodiments. The health monitor, for example, can be further configured to provide an audible, visible, or tactile indication to the user to indicate progress towards goals and / or indicate when a goal is reached. For example, a health monitor can project LEDs, beep, or vibrate when a user has reached an assigned health credit goal during a one-day interval.

Since a health monitor may include an accelerometer and components for processing motion data, in some embodiments, a health monitor 100 may be further configured to recognize specific motion gestures that a user may execute (e.g., health monitor, hand movement in a circle, rotate the health monitor). A health monitor, for example, can include one or more activity engines adapted to recognize such gestures. Recognized gestures, for example, can be used as an interface method to perform specific functions on the health monitor (for example, turn on, turn off, clear data, set a goal, display progress towards one or more goals, etc.) .

IV. Identification of types of activity

Identification of types of activities, according to some modalities, will be described in more detail. In some embodiments, the data generated by an accelerometer 130 and / or one or more physiological sensors can be processed to produce characteristic features Λ that are provided to the inference engine 320 for identification of the type of activity. The inference engine 320, for example, can be configured to receive processed data, and in some cases raw data, and further process the received data to identify and / or classify a detected activity. In some embodiments, an activity that is identified can be one of a plurality of different types of activity that are performed by a human and that the health monitor is programmed to recognize.

According to some embodiments, the inference engine 320 can classify activities into different categories. Categories can include, for example, non-human activities, recognizable human activities, and non-recognizable human activities. Classification can occur as the inference engine tries to identify a type of activity. According to some modalities, the data for each classification can be handled differently to determine the health credits for a subject, as further described below.

In some embodiments, the inference engine 320 may process the received activity data to identify an activity type in any suitable way. In some implementations, the inference engine 320 can identify an activity based on the acceleration derivative data D_n only (as described in US Patent Application No. 13/840098 which is titled Versatile Sensors with Data Fusion Functionality, mentioned above), or a combination of raw acceleration derivative data and physiological data and / or a set limited of characteristic features. In some embodiments, the inference engine 320 can identify an activity using one or more raw accelerometer traces and / or characteristic features generated from one or more raw traces. In some cases, the raw data or acceleration derivative data can be used in combination, or in combination with physiological data to identify or classify an activity. Additionally, in some embodiments, the inference engine may additionally or alternatively qualify the activity data (for example, identify a quality level of recognized activity data or identify an aspect of the recognized activity such as a location of the health monitor in the subject) using any combination of the acceleration derivative data D_n, raw accelerometer data, physiological data and related characteristic features.

According to some implementations, the inference engine 320 may employ one or more identification algorithms to identify or distinguish activities. For example, in some embodiments, the inference engine 320 may employ a pattern recognition algorithm to recognize an activity based on a waveform defined by the acceleration derivative data D_n and / or one or more raw accelerometer traces. In some embodiments, the inference engine 320 may additionally or alternatively employ fuzzy logic to identify an activity based on a number of values ​​in sets of characteristic features F_n = {f, ... f_n}. The inference engine 320 thus, in some embodiments, may employ a combination of pattern recognition and fuzzy logic to identify an activity. In some modalities, for fuzzy logic recognition, the fynCÍ0ne5 of ΠΊ§ [ηΐ2Γ65ίά ñptf® ^ 3 different activities can be defined and downloaded to the health monitor. It should be appreciated that other recognition algorithms or combinations of such algorithms may be used additionally or alternatively.

As just one example of activity recognition, data received by the inference engine 320 that has a predetermined number of values ​​that fall within a membership function scale can be identified by the inference engine as a function-related activity membership For some implementations, fuzzy logic may be suitable for recognizing a large variety of different activities without placing a heavy data processing load on processor 110. For example, fuzzy logic, in some embodiments, may only require determining whether a plurality of characteristic features of the trait data F_n falls within certain ranges of values. In other embodiments, fuzzy logic can additionally or alternatively evaluate a cost factor for each candidate activity and identify an activity based on the cost factor evaluation.

For the purposes of understanding only, and not limiting the invention, an example of activity identification is described with reference to Figures 4A to 4B. Figure 4A represents three traces (x, y, z. Top three) of raw acceleration data from a first type of activity (walking in this example). The lower trace D in Figure 4A represents the acceleration derivative data calculated from the upper traces. Figure 4B represents the corresponding traces of data obtained from a second type of activity (riding a bicycle in this example). For the data shown in Figures 4A to 4B a health monitor 100 was supported below the shin of the subject.

As can be seen from the traces of Figures 4A to 4B, there are a number of differences in the traces. The differences include maximum and minimum acceleration values, periodicity of the traces, number and shapes of the peaks in the traces, width of the peaks, and distances between the peaks among other things. Differences, for example, can be captured in sets of characteristic features F_n for each trace.

Although Figures 4A through 4B represent only accelerometer derived motion data, it will be appreciated that additional data from one or more physiological sensors can be processed in a similar way to generate feature sets for physiological data. A health monitor can process physiological data in combination with activity data to enhance information about a subject or related activity by the subject. For example, physiological data can provide an indication of an intensity level of an activity.

Continuing with the previous example, in some modalities, a set of characteristic features F_n for each trace within a measurement interval T_m can be built with the following inputs:

F_n = {maximum trace value (max); trace min value (min); number of peaks with a width smaller than the samples mi (N_p) ', number of valleys with a width smaller than the samples m₂ {N ^ ·, average rate of change of the trace at the maximum mean of the peaks (Ri_/2Y_r average distance between peaks (ΔΡ)}.

It should be appreciated that a wide variety of characteristic features can be generated and used to identify the different activities. As understood from this example, the differences in the traces in Figures 4A to 4B, for example, can be captured as numerical differences in the characteristic feature sets. In some embodiments, the inference engine 320 can then distinguish between activities based on such numerical differences, for example, by comparing the numerical values ​​of various characteristics against the corresponding values ​​for recognizable types of activity. A judicious choice of values ​​to include in feature sets, in some embodiments, can reduce the computational load on the inference engine 320 and allow quick identification of different types of activities.

According to some modalities, a membership role for each of the activities, for example, can be defined as follows:

<td></td><td>^F(max_ava - Smax) &lt;max &lt;(max_avg + 3max) f</td>

<td></td><td>(min_av5 - 3min) &lt;min &lt;(max_flW5 + dmax);</td>

<td></td><td>^ p, avg ^— dN_p) - Np - ^ p, avg + ^^ ρ)&gt;</td>

<td>^ activity ⁴</td><td>(N_v#_vg - dN_v) &lt;N_v &lt;(N_Viavg + dN_v}·</td>

<td></td><td>ÍRi -dRi} &lt;Ri &lt;ÍRi + a /? IY</td>

<td></td><td>X 2 / z \ zJ</td>

<td></td><td>. (AP_avg - dbP) &lt;ΔΡ &lt;(RAP_avg + 5ΔΡ),</td>

(1) where the subscript av ^ 'designates an average or expected value, and the quantities identify a predefined scale within which a measured value can be considered to qualify as belonging to the membership function. In some embodiments, when characteristic features are received by an inference engine 320, for which all values ​​qualify as belonging to the membership function, then the detected activity can be identified by the inference engine. The inference engine can then, for example, route the data from the feature generator, using a multiplexer 330 to an appropriate activity engine for further analysis.

In some implementations, by testing a large number 6053 ^ 05 ρθΓ0 ádlVÉl faith WiSCions in each URO of characteristic feature values ​​can be observed and statistics regarding variations can be determined. Statistical results, for example, can be used to help build membership functions for fuzzy logic activity identification. For example, it can be seen that a maximum value of acceleration derivative traces D for running has a 2-sigma variation of 5 measurement units. A membership function for running, may for example include the specification {(120 -5) &lt;maximum value of D &lt;(120 + 5)}, where 120 units of measurement are determined to be an average of the maximum value for the trace D for the running activity.

In some cases there may be a partial overlap of the functions of the membership. For example, one or more scales for characteristic features in one membership function may overlap or be coincident with corresponding scales in a second membership function, which may or may not belong to the same type of activity. Although there may be a partial overlap of membership functions, in some implementations, an activity type can be identified based on the evaluation of a plurality of characteristic features and their locations within the membership functions for the activity types. . For example, each activity can receive a record (for example, a value between 0 and 1) for each characteristic that falls within a role of membership for that activity. In some embodiments, after matching the records for each activity, the one that receives the highest record can be selected as the identified activity.

In some modalities, the records based on the described membership functions can be all or nothing, for example, any characteristic f_n it is measured and determined to be within the limits of their corresponding membership role and contribute to a record, or it may be off limits and contribute nothing. Other embodiments may additionally or alternatively employ membership functions such as those depicted in Figures 5A to 5C. The graphs represent the membership functions Mj that have been constructed for n traits of characteristics (denoted by the subscript /) and for two activities (denoted by the subscript The membership functions M_h¡ for example, they can be constructed from statistical analyzes of many measurements, as described above. Although membership functions are shown as trapezoidal, they can take any shape, for example round, semi-circular, semi-elliptical, Gaussian, parabolic tops, representative distributions of measured statistical values, multimodal distributions, etc.

In some implementations, when a characteristic feature is determined, for example, f_t 510-1, a corresponding value for each role of the activity membership for that trait can be determined. For that case shown in Figure 5A, for example, the second membership function of activity M_2r contributes to a value 512 where the first function of the activity membership contributes no value. The values ​​can be normalized as represented in the graphs. For a second measured characteristic 510-2 shown in Figure 5B, both membership functions for example, may contribute different values ​​522, 524. In this case the membership functions overlap. For another measured characteristic 510-n shown in Figure 5C, both membership functions, for example, can contribute identical values.

In some modalities, a cost factor G can be calculated for candidate activity (denoted by the subscript / ') based on the detected characteristic values ​​510-1, 510-2, ... 510-n and the functions of Míj default membership according to the following relationship:

ς; = í ^ · ^ · ^). ς;₌₁^ (2) where represents a weight factor for characteristic f^h of the activity f^h. The weight factor, for example, can be selected to emphasize some characteristics and to downplay other characteristics for the purpose of identifying an activity. In some modalities, an activity with the highest cost factor C, can be selected as the identified activity type. For example, referring to Figures 5A to 5C and considering only the membership functions shown and the measured trait characteristics 510-1, 510-2, ... 510-n, represented as open circles, the second activity M₂ it can be selected as the identified activity, according to some modalities.

For EC. 2, if the membership functions are each normalized, as represented in Figures 5A to 5B, the value of the cost function will range from 0 to 1. By multiplying the cost function by 100, a degree of match as a percentage confidence level.

PC¡ = 100 x C_t (3)

For example, a feature set that has each feature value that falls under the peaks of the membership functions for a particular activity will produce 100% confidence in matching and identifying the activity type.

In some cases, the membership functions may completely overlap, so that it is not possible for the inference engine 320 to identify an activity between the two membership functions. This situation can occur, for example when a new membership role is added to the inference engine 320 for recognition of a new activity not previously recognizable by the inference engine. For example, if the health monitor 100 is configured to identify cycling and is subsequently updated to identify elliptical training activity, a new membership function to identify elliptical training can be completely overlaid with the membership function preset for biking as two activities are similar.

In some embodiments, the membership functions and traits used by the inference engine 320 can be expanded, so that additional characteristics and associated membership functions can be added to the system for activities recognizable by the health monitor. Additional characteristic features and revised membership functions, for example, can be added to distinguish two activities that have previously substantially overlapped membership functions. The addition and updating of membership functions and characteristics can be accomplished, for example, by means of of communication between a health monitor with an external device, for example, a personal computer or computer connected to the Internet.

Through the use of the membership and / or pattern recognition features, a health monitor can identify the types of activity. In some implementations, a health monitor can further classify detected activities into multiple categories. Categories may include, for example: (i) recognizable activities, (i) unrecognizable activities that are likely to be performed by the subject, and (iii) non-human or non-animated activities (for example, activities performed by a machine or animal). The recognizable activities may correspond to the types of activity (for example, walking, walking, biking, swimming, etc.) that the health monitor has been programmed to recognize (for example, the monitor includes at least the functions of membership for these types of activity). In various modalities, the activities that fall within this first classification can be analyzed to identify a specific type of activity, and they can be candidate activities that can be further evaluated by the health monitor to determine a level of health benefit or credit. health for the subject based on the identified activity. For recognizable activities, an activity type can be determined by finding a high confidence match of characteristic traits for membership functions for a particular activity. In some embodiments, the confidence level can be expressed as a percentage of a maximum value that is obtained for a best match.

Activities that fall with the second ranking (unrecognizable activities that are likely performed by the subject) may or may not be further evaluated by the health monitor to determine a level of health benefit for the activity. In some modalities, a health monitor can be configured to admit or reject an unrecognizable activity for health benefit analysis based on the similarity of the activities to a recognizable activity. For example, a health monitor can be configured to distinguish between unrecognizable activities that are similar to the types of recognizable activity (for which the health monitor is programmed to recognize) and unrecognizable activities that are not similar to the types of recognizable activity. In some modes, unrecognizable activities that are similar to recognizable activity types can be weighted and evaluated according to the metrics for most similar recognizable activity. Unrecognizable activities that are determined not to be similar to recognizable activity types may or may not be given credit toward health benefits. Health benefit analysis methods are explained in more detail below for various categories of activities.

As a practical example to evaluate the similarity of activities, some exercise regimens that are set to music include a variety of repetitive body movements in the form of dance movements where new movements are created on a regular basis. Although a health monitor can be programmed to recognize some motion sequences, it may not recognize newly entered motions. In some cases, the movement and / or physiological data generated by a new movement can produce characteristic features f_n that are similar to other activity recognized by the health monitor. The other activity can be a recognized dance move, a recognized exercise (eg, jumping rope), or a move performed on a specific exercise machine.

A determination of similarity or non-similarity of a detected activity compared to a recognizable activity type can be based on a value of the cost function or a value of a confidence in the match. For example, a threshold value can be selected for any of these quantities, and match evaluations that produce results greater than the threshold value and less than the match value can be determined to be similar activities. Match evaluations that produce results below the threshold value can be determined to be dissimilar activities. As a further example that is not intended to limit match and threshold values, a threshold value of about 50% can be selected to determine similarity, and a match value of about 80% can be selected to determine positive match. Values ​​that are 80 or greater can be considered a positive match, according to this modality. Values ​​that are at least 50 but less than 80 can be considered similar, and values ​​that fall below 50 can be considered dissimilar for recognizable types of activity.

As noted above, in some implementations, feature sets for unrecognizable activities can be stored by a health monitor for subsequent identification by a user. For example, after downloading data from the health monitor to a computing device, the user can be made to identify an activity type related to a set of traits for which the health monitor is capable of recognizing activity. Identification can be based on the time in which the activity is performed. In some modalities, once identified by a user, the data stored for the activity can be analyzed for health benefits. The health monitor can later be updated to recognize the newly identified type of activity. In some implementations, there may be insufficient data stored for health benefit analysis, and the health monitor updates to recognize activity during future use.

In some examples, for example in an aerobics class or boot camp where there may be several new movements, it may be inconvenient for a health monitor user to remember and identify each new movement. In such cases, a user may prefer to have the health monitor automatically with activities related to similar activities as described above, instead of storing sets of traits and having the user identify each activity. According to some modalities, A health monitor may be user configurable to store feature sets for new non-recognizable activities for subsequent user identification and analysis or to automatically link non-recognizable activities to similar activities as described above. A software configuration can be provided to the user to configure the health monitor's handling of unrecognizable activities.

In some modalities, a health monitor can be configured to identify specific non-human activities, for example, representative precise repetitive circular motion of strapping a health monitor to a rotating object, representative repetitive back and forth motion strapping a health monitor to a reciprocal target, etc. Non-human and non-animated activities may correspond to activities that exhibit traits that cannot be performed by a human. For example, accelerometer data exhibits a very high degree of repetitive motion accuracy over extended periods of time or excessive speeds, accelerations, or movements not sustainable by a human. In some implementations, a health monitor can detect such traits, identify or classify the activity as non-human, and then finish further processing of the activity data. Activities that fall within the third classification (non-human activities) can be excluded from further analysis to determine a level of health benefit to the subject.

V · Analysis of health benefits

In various embodiments, a health monitor can be configured to determine one or more levels of health benefit for at least some activities performed by the subject and detected by the health monitor. The inventors have recognized Q | gunO5 monitor ^ AiÍJi $ 6 pyrite δ ^ ρΐϋί to produce health credit data conforming to recognizable health standards, and that such adaptation can reduce inaccuracies and wide variability related to count of steps through from different activity monitors. Health credits can, in some modalities, be determined based on the type of activity, and at least the duration of the activity, level of intensity of the activity and standards established by a health entity. According to some modalities, health credits can be determined from calorie burning rates, and calorie burning rates can be more accurately calculated using a combination of heart data and motion data rather than based on cardiac data alone or movement data alone. The determination of health credits can provide a more reliable figure of merits to evaluate the health benefits derived from a subject's exercise program than conventional metrics such as number of steps performed, number of kilometers traveled, number of minutes of exercise. exercise.

In some modalities, the calculated health credits of an exercise can be formulated as a system of points, which can be used across all types of activities. For example, a subject may be given some portion or multiples of a health credit point, also referred to herein as a health credit, for each minimum creditable time unit (MCTU). ) during which an activity is performed continuously at a particular intensity level. The number of points can depend on the intensity level, and the intensity level and MCTU can be set by a recognized healthcare organization. As an example, a subject may be given one health credit for each MCTU during which an activity is performed continuously at or above a first intensity level, and two points for each MCTU during which it is performed. activity continuously at or above a second intensity level for MCTU. In some implementations, a subject may be given a fractional point for each MCTU during which an activity is performed at a level below the first intensity level, for example, as an incentive to do at least some activity.

In some implementations, calculated health credits can be converted to total step counts using a standardized conversion formula, for example to provide a comparison to legacy step count devices. Converting health credits, which can be calculated more accurately than conventional step counts, to step counts can provide a more reliable comparison of total step counts between different health monitoring devices.

Demarcation levels for intensity and MCTU may correspond to guidelines established by a recognized health organization. For example, the MCTU can be a one minute interval and the first intensity level can be a calorie burning rate that falls within a scale between about 3.5 kilocalories (kcals / min) and about 7 kcals / min (a scale identified as moderate activity by the Center for Disease Control (CDC) and the World Health Organization (WHO). The second level of intensity can be a calorie burning rate that is at or exceeds about 7 kcal / min (a scale identified as vigorous activity by the CDC and WHO). Other scales and values ​​can be used in other modes.

Health credits can be accumulated over a day or week, stored, and subsequently provided to a subject or doctor as a record of subjects' exercise achievement over a period of time. For example, a subject may be active in one day and earn 40 health credits during 40 minutes of moderate activity effort and 10 health credits during five minutes of vigorous activity effort, according to some modalities. In some implementations, the same number of health credits can be earned during the same time intervals for moderate and vigorous activity, although the credits can be separated into moderate and vigorous records. A health credit record can provide a more convenient and accurate indicator of a subject's exercise and fitness level than a record of steps taken as calculated by different health monitors using different algorithms, for example.

In some modalities, a health monitor can be configured to determine and record Enhanced Health Credits (EHCs) for a subject. For example, the CDC and WHO have found that additional health benefits are obtained when an activity is performed continuously at a moderate or vigorous activity level for time intervals greater than about 10 minutes. In some implementations, a health monitor can determine when a subject performs one or more activities at moderate or vigorous activity levels for periods exceeding approximately 10 minutes and earns EHCs for the subject. Additional aspects and features will be described in relation to the following methods and systems for determining health credits.

Referring now to FIG. 6, an illustrative example of a method 600 for determining health credits according to some embodiments is depicted as acts arranged in a flow chart. The modalities of method 600, for example, can be implemented in a usable health monitor, and can be used to determine health credits for a subject from one or more activities performed by the subject and detected by the health monitor. The flow chart provides a snapshot of some acts for just one example of a method 600, according to some implementations, and there may be more or fewer acts than those illustrated in the figure in various modes.

According to some embodiments, the activity data can be received from at least one accelerometer of a health monitor, and the health monitor can be configured to identify 610 an activity type from the data. In some embodiments, the activity data may comprise only motion data. In some implementations, the activity data may further include physiological data received from one or more physiological sensors (eg, a cardiac waveform sensor). Once the system has identified 610 the activity, the system can determine 630 if the identified activity has been continued for a minimum creditable unit of time. If it is determined that the activity has been continued for a creditable minimum time, the system may determine 660 a health credit value for the activity and add a corresponding health credit value for an activity buffer. If the system determines 630 that an activity has not continued for a minimum creditable unit of time, the system may add 640 a null value to the activity buffer.

In various modes, the system can determine 670 whether the activity buffer is full. If it is determined that the activity buffer is not full, the system may return to a state of receiving activity data and identify 610 an activity related to the received activity data. If the system determines 670 that the activity buffer is full, the system can copy 680 the data in the activity buffer to another location and analyze the data. According to some embodiments, the system can store 690 a summary of the activity that occurred during the filling of the activity buffer.

Figure 7A depicts a sub-process for identifying an activity 610, according to some embodiments. The sub-process for identifying an activity 610 may comprise receiving 612 activity data. For example, the received activity data may comprise motion data received from an accelerometer of the health monitor. In some embodiments, the received data may further comprise physiological data that can be received from one or more physiological sensors attached to a subject. The received activity data can include data that has been processed by a preprocessor 305 and feature generator 310, and can also include raw data, according to some embodiments. In some embodiments, the sub-process for identifying an activity 610 may further include processing 614 the activity data. Processing 614 of the activity data can be carried out as described above in connection with FIG. 3.

During the processing of the activity data, an activity type can be identified, for example, as described above in relation to Figures 5A to 5C. Under some modalities, a detected activity can broadly fall into one of three classifications. For example, an activity can be a recognizable activity that results in data produced by an accelerometer and / or physiological sensor that has at least one set of related characteristics for which the health monitor is programmed to recognize. Examples of recognizable activities may be running, walking, biking, or swimming, although recognizable activities may not be limited to just these activities. Other recognizable activities may include different types of team sports activities. In some embodiments, the health monitor can detect activity that is unrecognizable but has characteristics that indicate that the activity was likely performed by a human. In some implementations, a health monitor can detect activity that has non-human characteristics, for example, precise repetitive motion or excessive forces that are not likely to be sustained by a human.

In some embodiments, a health monitor can determine 616 if a detected activity is not from a human. If it is determined that the detected activity is not human, the health monitor may return to a state of receiving 612 activity data and processing 614 the activity data. According to some implementations, non-human activities can be ignored for further data processing. In some implementations, certain types of non-human activities can be identified by the health monitor (for example, movement in a washing machine, circular or reciprocal movement indicating that it is a machine). In any case, detected non-human activities can be recognized by the health monitor, so that no health credit is earned by a subject.

If it is determined that the detected activity can be related to human activity, the health monitor may attempt recognition 618 of the activity. In some embodiments, determining whether the activity is recognizable comprises an attempt by an inference engine 320 to determine a type of activity. If the activity is recognizable, the system can identify 620 an activity type related to the received activity data. In some embodiments, the system may further determine 625 an intensity value related to the type of activity identified. The system may then proceed to step 630 of determining if the activity continues for a creditable minimum amount of time.

According to some modalities, if a detected activity is not recognizable but has characteristics that indicate that it can be performed by a human, the system can determine 619 whether the activity is similar to a recognizable activity (that is, an activity for which the health monitor is programmed to acknowledge). If the activity is determined 619 to be similar to a recognizable activity, the health monitor may identify 620 an activity type to relate to the detected activity, and determine 626 an intensity value for the detected activity. As just one example, the received activity data can be determined to be more similar to a running activity, but does not include the characteristics necessary for a defined run identification.

With reference again to FIG. 7A, according to some embodiments, the similarity of an activity to a recognizable activity may be based at least in part on the features and / or characteristics of motion data received from the accelerometer. In some cases, physiological data can be used in combination with motion data to determine 619 similarity of an unrecognizable activity to a recognizable activity. As an example, motion data, and in some cases physiological data, can be processed to generate feature sets as described above in connection with Figures 5A to 5C. According to some modalities, an activity that is determined to be similar to the recognizable activity may have a set of characteristics for which at least some of the characteristic data falls within one or more membership functions for one or more recognizable activities. . Data falling within one or more membership functions may not provide a high enough confidence level to make a definite match to any other recognizable activity. However, an unrecognizable activity can be determined to be similar to a recognizable activity based on the closeness of the measured characteristics to a defined match for one of the recognizable activities. For example, an activity such as jumping may generate a set of activities with values ​​that are closer to, or partially within, the membership functions for running than the other membership functions, such as walking or cycling. The jumping activity can then be evaluated as similar to a running activity.

An example for determining similarity is graphically represented in Figures 5A to 5C. In this example, the characteristics 530-1, 530-2, ... 530-n represent a set of characteristics measured by the health monitor for an activity that is not recognizable. The activity can be determined to be unrecognizable, because the measured characteristics do not provide a percentage confidence level that can give a definite match for any of the activities represented by the membership functions M¡ or M₂. In this example, characteristic 530-1 does not fall within a membership role M₁₄ or M₂₁. The measured characteristic 530-2 falls within a membership function M₁₄ for the first activity, although the confidence level of feature 530-2 is low within this membership role. The measured characteristic 530-n falls well within the membership function M_{1&gt; n}, but the combined confidence of all characteristics for this activity may remain too low to provide a definite match for any activity or According to some modalities, a health monitor can determine that the measured characteristics 530-1,

530-2, ... 530-n, as represented in the example, are more similar to an activity represented by the membership functions Μ_14ι M ^, and Mi, n- As you can see from the graphs, these measured characteristics are closer to the membership functions for the first activity than they are to the membership functions for the second activity.

In other embodiments, a health monitor can calculate a distance between one or more measured features, eg 540-1, and the closest membership functions. The health monitor can determine which membership functions for an activity are closest to the characteristics measured for the unrecognizable activity. If the measured characteristics are found to be closer to the membership functions for one activity than any other activity, then the health monitor may determine that the non-recognizable activity is similar to a recognizable activity for which the membership functions are more close to the measured characteristics for the unrecognizable activity.

According to some modalities, an activity that is determined not to be similar to any other activity may have a set of activities for which the data does not fall within any membership function of any recognizable activity, or falls below a value default threshold. An example of dissimilar activity is also depicted in Figures 5A to 5C. In these representation characteristics 540-1, 540-2, ... 540-n represent measured characteristics for an activity that can be determined 619 by not being similar to any other recognizable activity. In this case, the measured characteristics are all outside the membership functions for the recognizable activities. For example, none of the measured characteristics 540-1, 540-2, ... 540-n fall within a membership role for a recognizable activity. According to some modalities, a health monitor can determine such activity as being unrecognizable and not similar to a recognizable activity.

According to some embodiments, when an unrecognizable activity is found to be not similar to a recognizable activity, then a health monitor may assign 622 an unknown activity type to the activity. In some embodiments, a reserved character designation or bit sequence can be used for unknown types of activity. For example, a bit sequence of zeros (0000) can be assigned for an unknown activity type, although any other designation can be used. The reserved designation can then be used by the health monitor to identify periods of time in which unknown activities are performed, for example, to later cause the user to identify the activity for that period of time.

Referring again to FIG. 7A, after recognizing an activity or determining whether an activity is similar to a recognizable activity or not similar to a recognizable activity, the health monitor can determine an intensity value for the activity. The determination 625, 626 of an intensity value can be performed in any suitable way as described above for example. In some modes, motion data can be used to determine an intensity value for an activity. For example, motion data can be processed to determine a speed or rate at which an exercise is performed, and an intensity value can be based on or be proportional to the predetermined rate or rate. In some embodiments, a characteristic motion waveform of an activity can be a Fourier transform to determine the spectral energy for the waveform. In some implementations, the raw acceleration data can be squared and integrated, for example, to provide a value for spectral energy. An intensity value for the activity can be based on (eg proportional to) the measured spectral energy. Additionally or alternatively, in some embodiments, the determination of an intensity value may be based at least in part on physiological data. For example, heart rate data (eg, RR interval) and / or respiration rate data (eg, encoded in R waves over multiple beats) can be used to determine an intensity value for an activity. As just an example, based on a repetition rate and / or physiological data (e.g., heart rate, breathing rate) a health monitor can relate or assign gait in a step that has an equivalent heart rate and speed of breathing for activity.

According to some embodiments, for unrecognizable activities, determining 626 an intensity value for the unrecognizable activity may further comprise, for example, multiplying a determined intensity value for a similar activity based on the physiological data by a factor scale. This can be implemented, for example, because the activity was not positively identified and was instead judged to be similar to a recognizable activity and an intensity value is assigned based on the similarity and / or physiological data. Determining an intensity value may initially be similar to or equal to a process used for the corresponding recognizable activity, but can then be leveled by a factor of less than 1, according to some modalities. In some embodiments, the scale factor may be a constant value, or it may be inversely proportional to a level of confidence in a match between the unrecognizable activity and the related recognizable activity. In some embodiments, the intensity value may be a low value compared to an intensity value for a similar recognizable activity. For example, the reduced value may comprise a scale factor with a value of between about 50% and about 80% in some embodiments. In other modes, other values ​​can be used for the scale factor. The intensity rating scale for an unrecognizable but similar activity can toggle whether or not the activity is eligible for health credits.

If it is determined 619 that the received activity is not similar to a recognizable activity, the system may assign 622 a value of the unknown activity type for the detected activity. The system can also assign an intensity value 627 to the unknown activity type, in some modalities. According to some implementations, the intensity value can be determined from one or a combination of: repetition rate, heart rate, and breathing rate. In some implementations, the allocation may additionally or alternatively be based on the acceleration data, eg, repetition rate, detected rate, and / or energy spectrum of the acceleration data. After completing data processing for an activity that is not recognizable, the system may proceed to the step of determining 630 whether similar activity or types of unknown activity continue for a creditable minimum unit of time.

Referring now to Figure 7B, an exemplary embodiment of a health credit determination sub-process 660 is depicted. The exemplary method can be used to determine health credits using only motion data, for example, in the modalities where a health monitor may include motion detectors but lack physiological sensors. As depicted, the sub-process 660 may comprise an act to receive 662 activity type data. The activity type data may, for example, comprise an identifier for an identified activity type and include an intensity level that was determined by the health monitor for the identified activity type. Sub-process 660 may further include searching 664 for metabolic equivalents (METs) for the type of activity based at least in part on the determined intensity level where the activity was performed. According to some embodiments, METs can be determined from a look-up table as depicted in FIG. 8A.

METs can provide a measure of calories burned per kilogram per unit time for type of activity and intensity of activity. In various embodiments, METs can be obtained from a recognized compilation of metabolic equivalents for various types of activity and intensities. An exemplary compilation of METs can be found in the 2011 Compendium of Physical Activities: A Second Update of Codes and MET Values, by Barbara E. Ainsworth et al., Published by the American College of Sports Medicine, the full contents of which are incorporated herein by reference.

According to some embodiments, the lookup data table 802 may comprise at least activity type data 810, intensity data 812, and metabolic equivalent data 814. For the example shown, the lookup table includes data for varying intensity levels of running and aerobics The look-up data table may also comprise data for rest. In the example shown, the data is represented as binary data, although in other modes any method can be used to represent the data.

According to some modalities, once the activity type and intensity have been determined, they can be used to look up the METs in the 802 metabolic equivalent data table. For example, if the activity type is determined to be running, can be represented by the binary sequence (0110). If the intensity of the activity is found to be representative of the subject's gait between 6 to 8 mph, the intensity level can, for example, be represented by the binary sequence (01000) representing 8 mph. The health monitor then, using the activity type (0110) and intensity level (01000), uses these values ​​to look up the metabolic equivalent data table 802 to find the related metabolic equivalents (01011) for the activity. In this example, the number of METs for the identified activity can be 11 (METs). In some embodiments, the METs can be rounded to the nearest integer in the data table to reduce data storage space. In other modes, fractional values ​​can be stored in the data table, although this may require more memory in the health monitor.

Once the metabolic equivalents are determined, the health monitor can calculate a 665 calorie burning rate for the subject. Calculating the rate of calorie burning may, for example, include multiplying the METs by the weight of the subject. In some embodiments, calculating the rate of calorie burn may also include converting the units of time to minutes. Although Figure 8A shows an illustrative example of a look-up data table 802 that can be used to determine METs and calorie burn for an activity according to some modalities, other methods can be used to determine calorie burn and / or metabolic equivalents in other modalities.

According to some modalities, calorie burn can be calculated using a combination of heart data and movement data. In some implementations, the calorie burn rates can be calculated for a subject who is walking or running using heart and movement data, and the obtained burn rates can be used as calibration values ​​to calculate the calorie burn rates. for other activities.

For example, calorie burn, in some modes, can be more accurately calculated based on a maximum VO2 level.₂ of the subject and heart rate, rather than calculating calorie burn from the heart rate data and a subject's weight, age and duration of exercise or from the movement data and MET conversion. For example, when a maximum VO2 level₂ of the subject is known, the calorie burn can be calculated with improved accuracy from the following expression.

Cal = 14.34 (^ + C₂HR + C₃VO₂ + C₄W + C₅A] T (3) where HR represents the measured heart rate (in units of beats / min) of the subject during the activity performed, VO₂ represents a maximum value of VO₂ (in units of mL kg¹ min¹) for the subject, W represents the subject's weight in kilograms, A represents the subject's age in years, and T represents the duration of exercise in hours. The constants C_n in EC. 3 have been determined empirically for subject populations and differ from male and female subjects. According to some modalities, the constants C_n can have values ​​shown in Table 1. In some modes, other values ​​can be used for the constants C_n that are more specific to the subject, for example, are determined from a population study for which the subject has more in common with the population studied.

TABLE 1

<td>Constant</td><td>Male</td><td>Female</td>

<td>Cx</td><td> -95.7735</td><td> -59.3954</td>

<td>c₂</td><td> 0.634</td><td> 0.45</td>

<td>C₃</td><td> 0.404</td><td> 0.380</td>

<td>C₄</td><td> 0.394</td><td> 0.103</td>

<td>C₅</td><td> 0.271</td><td> 0.274</td>

In some cases, a user of a health monitor can enter their maximum VO2 value₂ if you know the value. However, in most cases, a user does not know their maximum VO value.₂. In some cases, a value of VO₂ max for a subject can be determined from the following equation, for example, although other expressions can be used to calculate VO₂ max in other modes.

VO₂ me

-------- + b₄

HR * T_c¹ (4)

The EC. 4 has been shown to provide an accurate estimate of VO2₂ max for a population of subjects. (See, Weyand, PG et al, J. Appl. Physiol. 91: 451-458, 2001.) In this expression, T_c represents the foot contact time during gait and the heart rate (HR) is determined over a period of time during which the foot contact time is approximately constant. A health monitor that includes an accelerometer located below the knee may be able to determine the foot contact time in some modalities. The quantities mi and b, can be gender specific and have values ​​shown in Table 2. In some modalities, other values ​​of m, and b_t can be used that are more specific to the subject, for example, determined from a population study for which the subject has more in common with the population studied.

TABLE 2

<td></td><td>me</td><td>bi</td>

<td>male</td><td> 34.4</td><td> 11.1</td>

<td>Female</td><td> 30.9</td><td> 10.3</td>

In some implementations, foot contact time can be determined from the speed of a subject, for example, when an accelerometer is located at a different location than below the knee. It has been found that the foot contact time T_c and speed 5 are linearly related according to the following expression.

S = m₂T_c - ¿&gt;₂ (⁵)

The m values₂ and b₂ depend on whether the subject is running or walking and may have the values ​​shown in Table 3. In some modalities, other values ​​of m₂and b₂can be used that are more specific to the subject, for example, determined from a population study for which the subject has more in common with the population studied.

TABLE 3

<td></td><td>m₂</td><td>b₂</td>

<td>walk</td><td> 3.15</td><td> 945</td>

<td>To run</td><td> 2.58</td><td> 258</td>

In some embodiments, a value for F ^ can be determined according to EC. 5, for example, from accelerometer data by a fixed health monitor to a moving subject. The health monitor can be configured to identify a subject's running activity as described above and the subject's speed. The health monitor can use the value of T_c calculated according to EC. 5 in EC. 4, for example, and a detected heart rate HR recorded during the subject's activity range for which T_c was calculated to calculate a value of VO₂ max according to EC. 4 for the subject. The calculated value of VO₂ max then can be used in EC. 3 to calculate the calorie burn for the subject. Based on calorie burn in VO₂ max calculated in this way can improve the accuracy of health credits calculated compared to metrics based on only heart rate or accelerometer data.

According to some modalities, a value of VO₂ max can be calculated automatically by the health monitor for a subject at regular intervals (e.g. daily, weekly, biweekly, etc.), so that the accuracy of the calculated calorie burn which is based on VO2₂ max (eg calculated according to EC. 3) is regularly updated. Calorie burn, for example, can be calculated using updated VO values₂ max and EC. 3 for various types of activities, whether or not the activities are recognizable, according to some modalities.

Referring again to FIG. 7B, after the calorie burning rates have been calculated, the sub-process 660 may further comprise determining 667 the health credit for the activity. In various modalities, the health credit can be determined based on at least the rate of calorie burning that was maintained during a measured time interval for the activity. The health credit can also be based on the guidelines established by a health entity. For example, the Center for Disease Control and the World Health Organization have determined that an exercise level that produces a calorie-burning rate between about 3.5 kcal / minute and about 7 kcal / minute is at a first level of exercise with health benefits. Accordingly, a health monitor, for example, can be configured to measure and process activity data in time intervals of approximately one minute, and gives a health credit for each minute of exercise in which the burn rate of calories falls within this calorie burning rate scale. The CDC and WHO have found that exercise that produces calorie burning rates greater than about 7 kcal / minute provides a second level of exercise that is beneficial to health. Accordingly, a health monitor, for example, can be set for two health credits for every minute of exercise during which the calorie burning rate is equal to or greater than about 7.0 kcal / minute. In some modalities, activities that produce a calorie burning rate below 3.5 kcal / minute can be earned without health credits. In other modalities, activities that produce a calorie burning rate below 3.5 kcal / minute can earn a partial health credit, for example, half a credit. Other amounts of credit and calorie burning speed scales can be used in other modalities.

As can be seen from the EC. 3, in some modalities it may not be necessary to identify an activity type to determine calorie burn and health credits. For example, an updated value of VO₂ max for a subject and a heart rate measured during an activity can be used to determine the rates of calorie burn for the subject for that activity. In some embodiments, a health monitor, for a subject's convenience, may identify or attempt to identify the activity performed. Figure 7C is an illustrative example of an alternative method of determining health credits for a subject.

According to some embodiments, a 660-a method for determining health credits may comprise receiving 662-a activity data for running and analyzing the activity data to determine 662-b a current value of VO2.₂ max for one subject. The value of VO₂max can be determined using heart rate and speed values ​​in the ECS. 4 and 5, for example. A method 660-a may further comprise subsequently receiving 664-a activity data for a recognizable or unrecognizable activity performed by the subject. The activity data can include RR intervals and / or cardiac waveform data from which a heart rate of the subject during performance of the activity can be determined. According to some modalities, calorie burning rates can be calculated 665-a for activity, for example, using EC. 3. In some implementations, a health monitor can also evaluate an energy spectrum of accelerometer data to verify that heart rate is related to physical activity performed by the subject. Health credits can be determined 667 from calculated calorie burn rates and 669 added to an activity buffer, according to some modalities.

Although VO₂ max can be calculated from EC. 4 based on running, in some modalities, it can be calculated from a different equation or set of equations based on another activity in other modalities.

Equations other than EC. 3 can be used, in some modes, to calculate calorie burn and health credits. For example, the following equation can be used in some implementations:

Cal = 14.34 [Ci + C_ZHR + C₃W + C ^ A] T (6)

For this expression, the values ​​of the constants C_n can be as shown in Table 4. According to some modalities, the CE. 6 can be calibrated for a subject using EC. 3 to adjust some of the constants, for example, C₂ and C₃, so that the calorie burn is approximately the same for the two equations. In some mod ^ li ^^ j, other values ​​for the constants C „can be used that are more specific to the subject, for example, they are determined from a population study for which the subject has more in common with the population. studied.

TABLE 4

<td>Constant</td><td>male</td><td>Female</td>

<td>Ci</td><td> -55.0969</td><td> -20.4022</td>

<td>c₂</td><td> 0.6309</td><td> 0.4472</td>

<td>c₃</td><td> 0.1988</td><td> -0.1263</td>

<td>C₄</td><td> 0.2017</td><td> 0.074</td>

As you can see, a health monitor can earn, not earn, health credits on an ongoing basis as they exercise. For example, activity data can be pooled to the processor which analyzes the data on a continuous basis to calculate health credits. In various modalities, the 660 determination of health credits can be done on a minute-by-minute basis, although any suitable time interval can be used.

As activity data is produced and analyzed and as health credits accumulate, the health monitor can generate a health credit data stream 805, as shown in FIG. 8B. The health monitor may temporarily store 669 the health credit data for an activity buffer. An example of an activity buffer 910 is depicted in FIG. 9A. The health monitor, in some modalities, may also add 669 an activity type indicator related to any health credits earned to the health credit data stream 805.

The health credit data stream 805 may, as a single example, comprise a series of bit sequences that are generated by processor 110 and stored in temporary memory. The bit streams can include at least two bit streams having M bits and N bits for each data input 820 in stream 850. In some embodiments, each data entry 820 may contain information about the health credit earned for an activity performed during the particular time interval, and may not include an activity identifier. In the example shown, two bits are used to represent the health credit value and four bits are used to represent the type of activity. Therefore there can be four values ​​related to health credit (for example, 0, 1, 2, 3), and eight values ​​that can be used to identify different types of activity for the modality shown For the example shown In Fig. 8B, at least two types of activity are identified during the generation of the health credit data stream 805. The first type of activity is represented by bit sequence 0110, which corresponds to a type of running activity according to the example shown in FIG. 8A. A second type of activity, which has the bit sequence 0001, corresponds to an idle state of the subject During buffer filling for two time intervals, the subject earns a health credit for the running activity and for one time interval subject earns two health credits for running activity. No credits are earned during the rest interval. Other modalities may use more or fewer bits to represent the health credit and type of activity and may include additional information or less information in each data entry 820. In some modalities, the type of activity can be identified only at the beginning or conclusion of an extended time interval during which the activity is performed.

After adding 669 the health credits and the activity type to the activity buffer, the done health monitor determines 670 whether the activity buffer is full. The activity buffer can be any suitable size and contain a plurality of data entries 820. The activity buffer, for example, can be measured to contain between 5 and 100 health credit data entries 820. According to some embodiments, the activity buffer contains at least 10 health credit data entries and in some implementations it may contain at least 20 data entries. Figure 9A depicts an activity buffer 805 containing 20 data entries wherein each data entry comprises a health credit value and an activity identifier, according to some embodiments. When it is determined 670 that an activity buffer is full, the activity buffer can be copied to additional storage, so that the copied contents of the activity buffer can be analyzed. In some embodiments, after the activity buffer is copied, the activity buffer can be cleared or overwritten with subsequent health credit data.

In various embodiments, the size of the activity buffer 910 is small, so it does not consume the important memory available in the health monitor, and so the processor can easily process and compile the data between each fill in the buffer memory. In some implementations, the data processing comprises reducing the data into a summary, so that the amount of memory required in the health monitor to store the health credit data and an activity log is less than otherwise. form may be required to store data for each minute of health monitor operation.

Once the data in the activity buffer 910 has been analyzed, the health monitor may store 690 a summary of the data in the buffer. According to some embodiments, summarizing the activity buffer data reduces the amount of data represented in an activity buffer full. The summary of the intermediate activity data may form an onboard stored data stream 860 as depicted in FIG. 8C, according to some embodiments. The onboard stored data stream may comprise timing data 850, health credit data 852, enhanced health credit data 854, and activity type data 856, according to some embodiments. The stored onboard data stream 860 may comprise more, less, or different types of information data in other modes. In some implementations, health credit data 852, enhanced credits 854, and activity type data 856 may correspond to analyzes of at least one full activity buffer 910. As the activity buffer 910 is repeatedly filled and analyzed, additional health credit, enhanced credit, and activity type data can be added to the stored on-board data stream 860.

In some implementations, 850 timing data can be added only to the start and / or end of the 860 stored data stream. In some cases, 850 timing data can be added to the start or end of each summary of activity data corresponding to the analysis of a full activity buffer 910. The timing data 850 can be used, for example, to keep track of when and how long an activity can be performed. In some embodiments, where the activity buffer 910 spans a fixed length of time, each activity summary entry in the onboard data stream 860 represents a known length of time. In such embodiments, timing data 850 may not be required between each entry of the activity summary. Instead, the duration of an activity can be determined by the number of sequential entries in the data stream stored onboard 860. In some implementations, timing data may be fed into the onboard stored data stream after breaks in activity, for example, when the health monitor is inactive for a period of time or interrupted for a period of time.

Referring again to FIG. 9A, one embodiment of a full activity buffer 910 is depicted. The exemplary activity buffer shows periods of 20 time intervals and a data entry for each time interval includes two pieces of information ( health credit values ​​and type of activity). For the example shown, the first data entry 912 indicates that a first type of activity that is identified by the bit sequence (0110) (for example, run according to the example in Figure 8A) received a value of health credit of 1 (represented by the bit sequence (01)). The full buffer 910 shows that the first type of activity was sustained for five time intervals. In the third time interval, represented as the third entry 914, the intensity level at which the subject performed the activity increased, such that the subject received two health credits for the activity during this time interval. Two health credits were also earned during the fourth time slot. In the sixth time interval, a second type of activity, defined by the bit sequence (0011) was performed by the subject. This activity (eg, walking) did not receive health credit. For example, the walk may have been very slow in pace or it may have lasted a full time interval. In the following time interval, the subject was in a resting state, represented by the bit sequence (0001). In the next three time intervals, the subject switched between a state of running at a low speed, which received no credit values, and the state of walking that did not receive health credits. During the eleventh time interval, represented by the eleventh data entry 918, the subject performed an activity that was not recognized by the health monitor. In this example, an unrecognized activity can be given a bit sequence of all zeros. During this time interval, the subject may be presented with some activity, such as stretching, but the health monitor cannot determine or identify the type of activity and / or the subject's heart rate may be insufficient to earn health credits.

During the fourteenth time interval, represented by data entry 920, the subject exhibited appreciable activity that was not recognized by the health monitor. For example, movement data and / or physiological data may have indicated that there was no significant movement by the subject and an increased heart rate indicate a high calorie burning rate. However, the health monitor cannot identify the type of activity. For example, during this period the subject may have performed calisthenics or some type of warm-up exercise. Due to detected movement and physiological data, the health monitor can be configured to award the subject a partial health credit or more. In some embodiments, a partial health credit may be indicated by bit sequence (11), or another special bit sequence. In some cases, a partial health credit can be one-half of a health credit. In the fifteenth time interval, the subject performed a different type of activity identified by the bit sequence (1001), for example, riding a bicycle. The subject then cycled at a moderate activity level, and subsequently increased to a vigorous activity level for the next two time intervals. At the eighteenth time interval, represented by data entry 924, the subject temporarily stopped receiving health credits. During this time interval, for example, the subject could have stopped for a traffic light. The subject then resumed cycling for the remaining two time intervals as recorded in the activity buffer 910.

According to some embodiments, a full activity buffer 910 can be analyzed to determine a total number of health credits received since the activity buffer was filled, a number of enhanced health credits (EHCs) received, and / or to identify the predominant activity that occurred during buffer filling. In some cases more than one activity type may be logged with the activity buffer full 910 analysis. An analysis of an activity buffer full 910 can result in an activity summary data entry 930, as shown in Figure 9B, which can be stored in long-term memory in the health monitor.

The activity summary data entry 930, for example, may be an entry that is added to the on-board stored data stream 860 that is depicted in FIG. 8C. As described above, the data entry may include 852-n health credit data, 854-n enhanced health credit data, and 856-n activity type data. Health credits can be determined by summarizing the total number of health credits received while filling the buffer for activity 910. In the examples shown, the total number of health credits received during 20 time intervals gives 15 health credits, represented by the bit sequence (001111). In the example shown, no enhanced credits are earned.

In some embodiments, the health monitor can be configured to determine a predominant activity that occurred during the activity buffer fill. In some cases, the predominant activity may be an activity that was performed for the greatest number of time intervals within the activity buffer 910. In some implementations, the predominant activity can be determined as the activity that receives the most credits during the buffer fill of activity 910. In the example shown, the predominant activity is selected to be the activity for the which the highest number of health credits were received, which is cycling in this example.

By compiling the activity full buffer 910 into an activity summary data entry 930, the amount of data stored on board the health monitor can be appreciably reduced. In the example shown, one hundred and twenty bits that are required for the activity full buffer 910 can be reduced to fourteen bits for the data input of the activity summary 930, which represents approximately a data reduction of 10: one. Consequently, the activity data can be monitored and analyzed at high resolution, for example every minute and the meaningful information about the activity can be stored with reduced data sympnincative, The algorithms of ®ιηρΜ ¡fe data Anafe can be used additionally or alternatively to further compress the activity summary data entry in some modes. Data compression may be desirable so that longer time intervals can be analyzed and recorded between downloads of the data from a health monitor which may require some manual interaction on the part of the user.

In some implementations, analyzing an activity full buffer 910 may further comprise tabulating transport time data 952 and transport credit data 954 as depicted in FIG. 9B. The leftover data 950 can be used, for example, in order to determine the improved health credits 854-n, as will be explained in more detail below. As described above, 854-n Enhanced Health Credits can be earned when performing an activity at or above a creditable health intensity level continuously for a time spanning multiple time intervals recorded in the memory buffer activity 910. For example, enhanced health credits can be earned when an activity that receives health credits is performed continuously for at least 10 minutes in some modes, or at least 20 minutes in some modes. In other embodiments, a time interval for receiving EHCs can be longer than or shorter than 20 minutes.

Determination of improved health credits 850-n can be carried out when analyzing the activity full buffer 910. According to some embodiments, the health monitor can review each successive data entry in a buffer of the activity fills 910 to determine a number of successive data entries and time intervals that have received at least partial health credits or more health credits. If the number of successive data entries that receive partial health credits or more health credits is greater than a threshold number, then the enhanced health credit can be assigned to the subject. As an example, each data entry in an activity full buffer 910 may be representative of an activity performed over a period of one minute. If the activity buffer full shows 20 data entries each receiving at least one partial health credit or at least one credit, then an enhanced health credit can be assigned to the subject. For example and as shown in Figure 9A, some data entries did not receive current health credits and therefore an improved health credit is not assigned in this example.

Enhanced Health Credits can provide an additional piece of information about a subject's exercise regimen. For example, enhanced health credits can be used to easily determine if the subject's exercise is LOW to LOW or exháUSt¡VQ&gt; Enhanced health credits can also provide a qualitative assessment of the subject's overall physical condition. For example, a subject who is able to sustain a health credited activity for an extended period of time may have a higher level of fitness than a subject who cannot and does not receive enhanced health credits.

Awarding of Enhanced Health Credits can be done in any suitable way. In some modes, Enhanced Health Credits are awarded for each extended time interval in which an activity is performed. For example, an EHC can be earned for each twenty minute extended interval that creditable exercise is performed continuously. In other modalities, 854-n enhanced health credits can be awarded for each health credit received during the extended period of exercise. For example, if during an extended period of twenty minutes of exercise a subject receives twenty-four health credits, then the subject can receive 24 enhanced health credits. In some modalities, separate point systems can be used for Health Credits and Enhanced Health Credits. In still other modalities, the 854-n enhanced health credits may be representative of a percentage of the 852-n total health credits that received enhanced health credit. For example, during an exercise interval of twenty minutes a subject receives twenty-eight health credits of which twenty-one are improved health credits, then the value of the improved credit 854-n may represent a value of approximately 75%. In some modes, a percentage value can be rounded to almost 10.

As can be appreciated, the extended activity intervals can span more than one activity buffer full 910. In the example shown in Figure 9A the first activity, running, ended during the fill of the activity buffer and a second activity, riding a bike, started during the second half of the activity buffer fill for activity 910. For such cases, the data to go ahead 950 can be stored and used by a health monitor to go ahead, to the next filling of the activity buffer 910, the information about an activity that was carried out during an extended period of time during a pre-fill of the activity buffer you can count towards the enhanced health credits.

For the example of FIG. 9A, an extended period of time (from the sixth data entry 916 to the thirteenth data entry) passed during which no health credits were received. Consequently, the 910 activity full buffer analyzes may not produce improved health credits. However, in the fifteenth time interval the subject began the activity identified as (1001), riding a bicycle, which persisted until the end of the activity full buffer 910, except for a brief interruption represented by data entry 924. According to some modalities, the health monitor can be configured to ignore the brief interruption, so that cycling activities can be followed as a continuous sequence of time intervals during which creditable health activity was performed continuously. . For this example, the transport time data 952 may reflect five time intervals that can count towards an extended time interval for the next activity buffer fill and analysis. For example, the five time slots can be added to an extended time slot, starting from the start of the next fill of the activity buffer, which may be eligible for enhanced health credits. Transportation credit data 954 can also be recorded to reflect the amount of health credits received during the transportation time. In this example, eight health credits were awarded during transportation time.

The data to go on 950 may be buffered temporarily or it may be stored in long-term memory according to some embodiments. In some implementations, go-ahead data 950 can be erased or overwritten for each fill of the activity data buffer. The go-ahead information 950 may comprise fewer or more bits than those shown in FIG. 9B, according to some embodiments.

According to some modalities, the health monitor may be tolerant of brief interruptions in exercise when determining whether or not the improved health credit should be awarded to a subject. As described above in relation to FIG. 9A, data input 924, representing an interruption in cycling activity (eg, stopped by a traffic light) can be ignored by the health monitor when it is determined that the sequence of time intervals extending from the fifteenth to the twentieth time interval were effectively performed continuously as the creditable health activity. In some embodiments, a health monitor can review the data entries in an activity full buffer to determine the number of sequential data entries that do not receive health credits. If the number of successive data entries that do not receive health credits are less than a threshold value, then these data entries can be ignored when determining whether the activity surrounding these data entries was actually performed continuously as the activity creditable health. According to various modalities, the threshold number can be any value between zero and six, although higher numbers can be used depending on the time span for each 912 data entry. As just an example, and with reference again to the Figure 9A, the threshold number can be three. Therefore, the data entries for the six to thirteen time intervals represent a definitive cessation of creditable health activity. Any transportation time that occurs prior to this termination may be overridden or not added to an extended interval that occurs after the termination. However, data entry 924 represents an interrupt (less than three intervals) that can be ignored by the health monitor for purposes of determining the improved health credit. Any transportation time and / or extended time that occurs prior to the outage can be added to an extended activity interval that occurs after the outage for the purpose of determining improved health credits.

In some embodiments, a time slot that receives a partial health credit can be treated as a time slot that does not receive health credit for purposes of determining the improved health credit. For example, an interval of time that you receive a partial health credit may count toward a cessation or interruption of activity. In other embodiments, a time slot that receives a partial health credit can be treated as a time slot that receives one or more health credits for the purpose of determining the enhanced health credit. For example, a time slot that you receive a partial health credit can count toward an extended time slot during which the health credit activity is performed.

It will be appreciated that health credits can be converted to step counts, in some modalities, for comparison to legacy systems. For example, health credits that are determined from calorie burn rates can be expressed in terms of METs. From the METs value, a walking speed can be determined for a subject and from the subject's walking and entry speed (as can be determined with the motion sensor) the equivalent steps taken during an exercise interval to obtain the same result. METs can be determined. In some modalities, a health monitor can be configured to accumulate equivalent steps for all activities performed by the subject receiving the health credits.

SAW. Additional aspects of activity data processing

Determining parameters, such as distance and speed of a walking or running person, that accurately represent a detected activity can be difficult without proper calibration techniques. Thus, in some embodiments, to ensure that a health monitor 100 provides data that accurately reflects various parameters related to a detected activity, activity-dependent calibration factors may be employed, for example, Such factors can be used by a 340-m activity engine when calculating activity-related data from data received from a moving direction and preprocessing circuitry. Activity-dependent calibration factors for each activity, for example, can be maintained and updated in memory 120.

In some implementations, the calibration factors can be used in one or more equations used by a 340-m activity engine to calculate a measure of activity intensity. An example of such an implementation is described, for example, in U.S. Patent No. 4,578,769 (incorporated by reference above) which describes deducing a runner's speed based on foot contact time that is detected by a sensor that is placed on footwear. In some embodiments, there may be a number of different calibration factors required for a health monitor configured to recognize a number of different types of activity. Such calibration factors, for example, can be determined prior to type, eg, through laboratory testing and experimentation, and then loaded into memory 120 of a health monitor 100 before use.

In some embodiments, calibration can be performed in conjunction with a user review of the data and user input. For example, a user can jog for 3.2 kilometers, and record a time, 14 minutes, 0 seconds (14:00), that it takes him to jog 3.2 kilometers. A health monitor, for example, uses a pre-defined calibration technique to identify activity such as running, and the 340-m activity engine can calculate a running pace of 6:50 minutes / kilometers. In such an implementation, the user can then, through a computer-based interface with the health monitor, run a calibration routine where the user can first select the identified activity and calculated rate, and then enter a known rate. for the activity. The system can then adjust or replace an internal calibration factor used by the health monitor 100 with a new calibration value for that activity. In this way, calibrations for various activities can be made specific to individual users of the health monitor, which can improve the accuracy of the device for each user.

In some embodiments, a health monitor 100 may additionally or alternatively be configured for automatic calibration or self-calibration for one or more activities. Such calibration routines can be run for one or more recognizable activities. As just an example, the auto-calibration for running will be described. When a health monitor includes an accelerometer that is placed on the foot or ankle, the accelerometer will temporarily enter weight along the direction of running (which is taken as x-directed in this example) like the soles of the foot on the I usually. When the foot is fixed, the rigid velocity xd i of the accelerometer is zero, and this can serve as a reference point for calibration. When the foot is then set, the x-directed velocity again returns to zero. By integrating the x-directed acceleration data twice, a distance between the two successive soles can be determined. The distance can be corrected using y- and z-directed acceleration values, since the orientation of the accelerometer changes as the foot moves forward. Once the distance is determined, a time between the soles of the feet can be determined from an internal clock of the processor 110, for example. Time and distance can then be used as the speed of the runner or walker. Velocity can be determined from two successive soles or more to obtain an average value, and the process of determining velocity can be repeated at separate time intervals. In some modes, the calculated speed can be used to update or correct the internal calibration values ​​used by a health monitor 100. For example, the calculated speed can be used to correct a calibration value used to calculate running speed based on foot contact time.

The calibrations can be used additionally or alternatively by the health monitor in a different way, and such calibrations can be referred to as location-dependent calibrations. For example and with respect to running or walking without being limited to just these activities, when a health monitor 100 is placed on the ankle or foot, for example, a more accurate measurement of activity can be made than if the monitor were placed on belt or placed in a trouser pocket. This can be seen, for example, in the accelerometer raw data traces of Figures 10A to 10C. When the monitor is worn on the ankle (Figure 10A), the z and x waveforms are more pronounced when the monitor is worn on the belt (Figure 10B). The foot strike time and / or foot contact time can be more precisely determined using data from a health monitor worn on the ankle or foot.

In various modalities, a health monitor can be additionally or alternatively configured to recognize a type of activity independent of the location where the motion sensor is used and is further configured to identify where the motion sensor is used for activity. . Just as the data traces in Figure 10A can be identified as walking by the inference engine 320 as described above, the traces in Figure 10B can be identified as walking where the motion sensor is worn as a belt and the Traces of Figure 10C can be identified as walking where the motion sensor is located in a bag. For example, the traces in Figure 10B can generate characteristic features f_n that belong more closely to one or more membership functions that can identify activity such as walking, motion sensor on the belt

In some embodiments, when an activity is identified where a health monitor includes an accelerometer mounted in a non-optimal location, a different calibration or scale value or values ​​can be used by the 340-m activity engine to calculate one or more parameters related to the activity, according to some modalities. For example, different calibration values ​​can be related to each identifiable activity and location of the motion sensor In other modalities, when an activity is identified where the motion sensor is mounted in a non-optimal location, the information that was gathered from the use above of the health monitor, when the motion sensor was mounted in a more optimal location, It can be used additionally or alternatively to lower or estimate activity parameters with the motion sensor in the non-optimal location. For example, the walking data collected when the motion sensor is worn on an ankle can be used to determine step lengths that correspond to different walking step frequencies or cadences. Thus, when the motion sensor is used in a non-optimal location, (for example a belt or bag) a detected cyclic frequency or cadence, for example, can be used in conjunction with previously obtained data to infer or calculate a length of steps for the activity. The calculated step length can be user specific. In some implementations, different calibration techniques and / or calibration values ​​can be related to each identifiable activity and location of the motion sensor.

In some embodiments, the characteristics of an activity can be additionally or alternatively inferred by a health monitor from the above high-quality data and an intensity for the activity can be calculated accordingly. Again with reference to the example of Figures 10A to 10C, the health monitor 100, for example, may store in memory 120 or provide storage in an external memory device one or more samples of high quality data (Figure 10A). when such data is collected and the motion sensor is used in an optimal or near optimal location to characterize the activity. In some modes, when the activity is repeated and the inference engine identifies the activity but with the motion sensor used in a non-optimal location (Figure 10B or 10C), the health monitor can retrieve more quality data from memory. high with a cadence that matches the currently detected activity. Higher quality data, for example, can be repeatedly provided to the activity engine 340-m for subsequent processing. As the currently detected cadence changes, different samples for example can be retrieved from storage. In some embodiments, samples retrieved from storage may depend on additional values ​​of currently detected signals other than cadence, eg, peak values, peak widths, minimum values.

In some embodiments, a health monitor may additionally or alternatively provide a confidence measure in conjunction with data output by an activity engine 340-m. For example, the monitor can indicate a level of confidence in the recognition of the activity (for example,&gt; 90% confidence,&gt; 75% confidence,&gt; 95% confidence), and can also indicate a quality level of the data (eg better, fair, poor). Confidence can be determined, for example, by how central each measured trait falls within a membership function or is based on a calculated value of the cost factor, (for example, value calculated according to EC. 2 ) for an activity, or how well a measured pattern matches a reference standard (for example, using least means squares difference algorithm). The quality of the data can be determined, for example, based on an identified location of the motion sensor of the health monitor as it is used by the user during the identified activity.

As described arrive in some modalities, the calibration values, characteristic features, membership functions and / or computation algorithms used by a health monitor 100 can be added and / or revised when the device interfaces with a computer by means of of a 140 transceiver. Thus it should be appreciated that in such modes the health monitor can be customized to become more accurate for a particular activity, for a particular location of sensors being used, and / or for a particular user. For example, when the device detects activity data that cannot be recognized by the inference engine 320, the device can record the acceleration data, and optionally physiological data, along with any characteristic features generated from the data, time, and duration of activity not recognized. In some embodiments, when subsequently in communication with an external device that has a user interface, such as a computer, smartphone, PDA, or similar device, the system, for example, may present a query to the user to identify the activity, intensity of effort and / or location of the motion sensor. In such modalities, information, for example, can be returned to the device, and a new membership role, characteristics and / or identification algorithm can be defined for the activity. The function of the membership and / or identification algorithm, in some modalities, can be produced external to the health monitor and downloaded. In some modalities, one or more 340-m activity engines can be added additionally or alternatively in order to customize a health monitor 100.

In some embodiments, a health monitor 100 can be useful for broad community challenges, allowing users to be able to easily compare themselves. A health monitor, for example, can be used to more accurately give an advantage to users of different performance capabilities. Some modalities of the Health Monitor 100, for example, can allow healthcare providers, insurance companies, and employers to more accurately assess fitness levels, exercise regimens, and health benefits of exercise for individuals and therefore provide appropriate incentives.

According to some embodiments, a health monitor can be configured to calculate a fitness metric that is indicative of a subject's overall fitness and / or health level. For example, a health monitor can collect activity data, which can include physiological data (for example, blood oxygenation data, respiratory data, heart data (and health credit data, and calculate a fitness metric of results using a standardized algorithm. In some modalities, the fitness metric may be representative of a VO2 max value.₂ for the subject or it can be a standardized or recognized physical condition or health index (HI). According to some modalities, a fitness metric F_m It can be calculated by a health monitor using the following formula.

^{Fm = β} R * HR ⁽⁷⁾

In the EC. 7, β represents a proportional constant, P can represent a rhythm, or equivalent rate calculated from a different activity by means of metabolic equivalents, R can represent a respiratory rate, and HR can represent a heart rate. The fitness metric in CD. 7 has been formulated in such a way that the higher number indicates a better level of fitness. For example, for a given rate P, an individual with a lower respiratory rate R, and a lower heart rate HR will be a healthier or fitter individual, as they are able to achieve the same rate more efficiently (by rating the cardiovascular system by a smaller amount).

Other fitness metrics can be additionally or alternatively calculated. As just one example, an aerobic fitness metric F_to can be calculated from the following equation for a subject.

F_to = 1 / (T_C x HR} ₍₈₎

In this expression, T_c represents the foot contact time, which can be determined from running speed according to EC. 5, and HR represents heart rate.

Some modalities may use other fitness metrics, or a metric where a lower number, rather than a higher number, indicates a better level of fitness. Other factors (eg, body mass index, blood pressure, blood oxygenation, resting metabolic rate, rate of calorie burning, metabolic rate) can be used in addition to or alternatively to those shown in the EC. 7 to determine a fitness metric. In some modalities, a fitness metric may include an assessment of the subjects' health credits and / or improved health credits accumulated over a period of time (e.g., one week, month, several months, one year, etc.) A fitness metric can comprise a weighted combination of various fitness indicators. An index of physical condition such as the one shown in K. 7 ü K. 8 can provide a summary of the subject's level of condition and / or health.

Other fitness metrics can also be determined with a health monitor that is configured to provide heart and movement data. As just one example, a health monitor can process activity data to determine when a subject is sitting or in a prone resting state, and process heart data during a resting state to determine a subject's resting heart rate ( RHR).

In some embodiments, motion data can be processed by the health monitor processor to determine when a subject is in an active state and performing activity at near maximum capacity. The performance of a near maximal capacity of the activity can be determined, for example, from the heart rate, speed of the subject, and / or repetition rate corresponding to the activity. In some implementations, an activity near maximum capacity level may be indicated by a detected heart rate that is greater than 75% of a maximum detected heart rate for the subject or greater than 85% of a maximum detected heart rate for the subject. in some modalities, or greater than 90% of a maximum detected heart rate for the subject in other modalities. A health monitor can then be configured to determine from motion data when the subject stops activity and assumes a resting state or recovery state. The health monitor can then process cardiac data after cessation of activity to determine a heart rate recovery time for the subject. A shorter recovery time may indicate a higher level of fitness. According to some modalities, the determination of heart rate recovery times can be performed automatically by the health monitor following any performance of a near maximum capacity of the activity by the subject.

According to some modalities, heart rate variability (HRV) can be determined by a health monitor during periods of activities performed at different intensities to assess the stress levels induced in a subject by the activity. Less reduction in HRV for higher levels of effort in an activity may be indicative of a fitter subject. In some implementations, a health monitor can be configured to record and process cardiac waveforms for extended time intervals (for example, intervals greater than 30 seconds) to determine the low-frequency and high-frequency spectral properties of the waveforms. . Low-frequency (LF) components, for example, can comprise frequencies between about 0 Hz and about 0.15 Hz. High frequency (HF) components, for example, may comprise frequencies between about 0.15 Hz and about 0.4 Hz. According to some embodiments, a spectral energy ratio of LF / HF can be calculated during periods of activities performed at different intensities to evaluate level them? stress induced in a subject by the activity.

In some implementations, a health monitor can be configured to evaluate the spectral energy of HRV and / or LF / HF to evaluate a condition of a subject. As just one example, a health monitor can evaluate the LF / HF spectral energy in the morning when an athlete wakes up but is still in a prone position. A LF / HF spectral energy value taken at this time may indicate that the athlete has fully recovered or not, from work the previous day. In some modalities, a ratio between approximately 0 and 0.15 may indicate that an athlete has not fully recovered from work from a previous day. The value can also be used to indicate, for example, a level of activity that can be performed by the athlete during the day with a lower risk of injury to the athlete.

As another example, a value for a spectral energy ratio of LF / HF may indicate that an athlete has or has not warmed up enough before harder work. For example, a LF / HF spectral energy ratio calculated by the health monitor to be greater than 0.4 may indicate that an athlete has not warmed up enough, and may further indicate that an HR * T product_C may not be accurate in evaluating the various fitness parameters. In some implementations, a LF / HF spectral energy value can be used to assess when subjects having a particular medical condition (e.g., CHF, COPD, advanced stages of cancer or diabetes, etc.) are in a state of tension. For example, a ratio between approximately 0 and 0.15 may indicate that a patient is in a state of stress.

One or more fitness metrics or health index value can provide ways in which to monitor the health conditions of subjects. For example, certain ailments can adversely affect a fitness index for an individual. Conditions that can adversely affect a fitness index include, but are not limited to, COPD, CHF, arthritis, dementia, diabetes, depression, PAD, hypertension, and obesity. In type two diabetes (T2D) there are studies that indicate that an increase in a max level of VO₂ of the subject is directly correlated with reduced dependence on medication. For COPD patients, a fitness index (for example, VO₂ max) can provide a diagnostic marker showing the stage and progress of the disease.

Some studies have shown that physical activity can help correct a disease state, reduce medication, or slow the onset or progress of a disease state. The CDC / HHS recommends moderate to vigorous activity approximately 150 minutes a week for the normal population. Although some activity monitors can monitor basic parameters of an activity, many conventional monitors cannot provide a metric or objective view of how well a patient or athlete is progressing on their fitness journey. A health monitor configured to calculate health credits and fitness metrics as described above can provide a convenient device for measuring and tracking a subject's fitness level on a daily basis.

There is currently little or no guidance for sick individuals as to what level of exercise is most effective for their condition. Although the look-up tables have been generated for various ages, for example, listing maximum recommended heart rates by age and gender, these tables do not take into account disease states or drug regimens for individuals. Thus, with respect to exercise-based treatment of a medical condition, it is not known whether the level of exercise is too little to be effective or too much to put additional health risk. A networkable health monitor can provide data for large user populations, of which exercise guidelines can be established based on age, gender, condition, and stage of condition. The accumulated statistical data can be used to update the health monitors with recommended exercise guidelines, in terms of health credits and improved health credits in some modalities, for various health conditions of a subject.

When treating an ailment, there may be times where the drug prescribed to treat a disease has a harmful effect on the subject, for example by lowering the body's ability to metabolize oxygen. Monitoring a fitness index that correlates with blood oxygenation for a subject suffering from one or more of these conditions can provide a convenient way to evaluate the effectiveness of various types of treatment, including pharmaceutical and exercise therapy and track the subject's recovery. For example, a health monitor that summarizes exercise in terms of health credits and / or fitness index can allow a doctor or individual to observe, almost immediately, any impact that a prescribed treatment may have on the human motor ( heart system pul mones-circu lato rio).

Although the above examples for determining health benefits are structured primarily in terms of two levels of activity intensity (moderate and vigorous) to determine health credits and enhanced health credits, other modalities may use additional activity intensity levels or Minors and Criteria for Determining Health Credits, Enhanced Credits, and Fitness Metrics. For example, a health entity can also refine intensity levels and criteria standards to evaluate the health benefits of activities, so that the scales and number of intensity levels can be altered in accordance with a health entity. In various modalities, a health monitor can be immediately reprogrammed to accommodate such changes. Additionally, enhanced health credit gradations can be implemented in some modalities. For example, different values ​​or point systems for Enhanced Health Credits can be earned based on one or both of the activity intensity and long duration of the activity. As an example, an additional point system can be used for creditable health activities performed continuously for 40 minutes, for example, so that a subject or doctor can better assess an endurance fitness metric for the subject.

All literature and similar material cited in this application, including, but not limited to, patents, patent applications, articles, books, treatises, and web pages notwithstanding the format of such literature and similar materials, are expressly incorporated by reference into its entirety. In the event that one or more of the incorporated literature and similar materials differ from or contradict this application, including but not limited to defined terms, use of terms, described techniques or the like, this application controls.

The section titles used herein are for organizational purposes only and should not be construed in any way limiting the subject matter described.

Although various inventive embodiments have been described and illustrated herein, a variety of other means and / or structures will readily be conceived by those skilled in the art to perform the function and / or obtain the results and / or one or more of the advantages described herein. , and each such variation and / or modification is considered within the scope of the embodiments of the invention described herein. More generally, those skilled in the art will readily appreciate that all parameters, dimensions, materials, and configurations described herein are considered exemplary, and that actual parameters, dimensions, materials, and / or configurations will depend on the specific application (s) for which they are being used. use the teachings of the invention. Those skilled in the art will recognize, or be able to determine using no more than routine experimentation, many equivalents to the specific inventive modalities, described herein. Therefore, it should be understood that the foregoing embodiments are presented as an example only and that within the scope of the appended claims and equivalents thereof, the embodiments of the invention may be practiced in a manner other than as described and claimed specifically. The embodiments of the invention of the present disclosure are directed to each individual feature, system, article, material, and / or method that is disclosed herein. Also any combination of two or more of these characteristics, systems, articles, materials and / or methods, if said characteristics, systems, articles, materials and / or methods are not mutually inconsistent, is included in the scope of the invention herein. description.

The above-described embodiments of the invention can be implemented in numerous ways. For example, some modalities can be implemented using hardware, software, or a combination thereof. When any aspect of a modality is implemented at least in part in software, the software code can run on any suitable processor or collection of processors, whether provided on a single computer or distributed across multiple computers.

In this regard, various aspects of the invention, for example feature generator 310, preprocessor 305, inference engine 320, activity engines 340-m, and data service 360, and versatile sensor network functionality can be represented by less in part as a computer-readable storage medium (or multiple computer-readable storage media) (for example, computer memory, one or more floppy disks, compact discs, optical discs, magnetic tapes, flash memories, circuit configurations in Field Programmable Gate Networks, or other semiconductor devices, or other tangible computer storage medium or non-transient medium) that are encoded with one or more programs, which, when run on one or more computers or other processors, perform methods that implement various modalities of the technology discussed above. The computer-readable medium (s) may be transportable, such as the program or programs stored therein may be loaded into one or more computers or other processors to implement various aspects of the present technology as discussed above.

Various aspects of the health monitor described above can be implemented in hardware, software, firmware, or a combination thereof. For example, any of the operational aspects of a health monitor that involve data processing, data handling, and / or communications can be implemented as stored machine-readable instructions that are executable by a microprocessor and are characterized in at least a tangible, computer-readable storage device. The instructions can be executed or put into operation in a digital processor of a health monitor. In some implementations, the instructions can be put into operation in a central cube or server that operates in conjunction with the operation of a health monitor.

The terms program or software are used herein in a generic sense to refer to any type of computer code or set of machine-executable instructions that can be used to program a computer or other processor to implement various aspects of this technology. as discussed above. The term processor can be used to refer to a microprocessor, microcontroller, or any programmable logic device including, but not limited to, field programmable gate networks. Additionally, it should be appreciated that according to one aspect of this modality, one or more computer programs that when executed perform methods of the present technology do not need to reside in a single computer, processor, or microCrOCQntroller but can be distributed in a modular form between a number of different computers, processors or microcontrollers to implement various aspects of the present technology.

Computer-executable instructions can be in many forms, such as program modules, executed by one or more computers or other devices. Generally, program modules include routines, programs, objectives, components, data structures, etc. that perform particular tasks or implement particular abstract data types. Typically, the functionality of the program modules can be combined or distributor as required in various modes.

Also, the technology described herein can be represented as a method, of which at least one example has been provided. The acts performed as part of the method can be ordered in any suitable way. Accordingly, modalities can be constructed where the acts are performed in a different order than illustrated, which may include performing some acts simultaneously, although they are shown as sequential acts in illustrative modalities.

All definitions, as defined and used herein, should be understood to cover dictionary definitions, definitions in documents incorporated by reference, and / or ordinary meanings of the defined terms.

The indefinite articles a and an, as used herein in the specification and in the claims, unless clearly stated otherwise, should be understood as at least one.

The phrase and / or, as used herein in the specification and claims, should be understood as implying either or both of the elements thus put together, that is, elements that are present together in some cases and present in a disjunctive manner in other cases. The multiple elements listed with and / or should be considered in the same way, that is, one or more of the elements thus combined. Elements other than the elements specifically identified by the clause and / or, whether related or unrelated to those specifically identified elements, may optionally be present. Thus, as a non-limiting example, a reference to A and / or B, when used in conjunction with unlimited language such as comprising may refer, in one embodiment, to only A (optionally including elements other than B); in another embodiment, to only B (optionally including elements other than A); in yet another embodiment, both A and B (optionally including other elements), etc.

As used herein in the specification and in the claims, or shall be understood as having the same meaning as and / or as defined above. For example, when elements are separated in a list, o and / or should be interpreted as being inclusive, that is, the inclusion of at least one, but also including more than one, of a number or list of elements, and optionally additional unlisted items. Only terms that clearly indicate otherwise, such as only one of or exactly one of, or, when used in the claims, consists of, will refer to the inclusion of exactly one element of a number or list of elements. In general, the term or as used herein should be construed solely as indicating exclusive alternatives (i.e. one or the other but not both) when preceded by exclusive terms, such as any of, one of, only one of, or exactly one of. What essentially consists of, when used in claims, shall have its ordinary meaning as used in the field of patent law.

As used herein in the specification and in the claims, the phrase at least one, with reference to a list of one or more elements, shall be understood as indicating at least one element selected from one or more elements in the list. of items, but not necessarily including at least one of each and every item specifically listed within the item list and without excluding any combination of items in the item list. This definition also allows elements other than those specifically identified to be optionally present within the list of elements to which the phrase refers at least one, whether related or unrelated to those specifically identified elements. Therefore, as a non-limiting example, at least one of A and B (or, equivalently, at least one of A or B or, equivalently, at least one of A and / or B) may referring, in one embodiment, to at least one, optionally including more than one, A, with no B present (and optionally including elements other than B); in another embodiment, to at least one, optionally including more than one, B, without A present (and optionally including elements other than A); in yet another embodiment, to at least one, optionally including more than one, A, and at least one, optionally including more than one, B, (and optionally including other elements); etc.

The claims should not be construed as limited to the described order or elements unless stated to that effect. It should be understood that one of skill in the art can make various changes in form and detail without departing from the spirit and scope of the appended claims. All modalities that are within the spirit and scope of the following claims and their equivalents are claimed.