PERSONAL HEALTH TRACKER AND WELLNESS ADVISOR
CROSS-REFERENCES TO RELATED APPLICATION [0001] This application claims priority to and is a complete specification of Application No. 201941004950, titled “PERSONAL HEALTH TRACKER AND WELLNESS ADVISOR”, filed on May 07, 2019.
FIELD OF THE INVENTION [0002] The disclosure generally relates to wearable technology in healthcare and, in particular, to personalized health tracking and wellness recommending devices.
DESCRIPTION OF THE RELATED ART [0003] Developments in portable electronic devices have greatly increased in the past decade. Today, smart phones are used for numerous applications including voice and video calls, photography, playing media, navigation, socializing, etc. The growth of multi-purpose smart devices like mobile phones in the market has inspired new innovations to satisfy customer needs. New types of smart devices, such as wearable devices, have been developed that provide physiological information of the user. With more people wanting to upgrade their fitness, health, and performance, and prevent lifestyle diseases like diabetes, obesity, hypertension, and heart diseases, the demand for wearable fitness tracking devices is increasing.
[0004] Generally, fitness tracking devices include number of different sensors to measure various health parameters, such as heart rate, step count, calories burnt, sleep analysis scores, etc. Some advanced precision fitness wearable devices also provide information that are very useful to users especially trainers and athletes. However, the advanced wearable devices may be expensive and may not address fitness and health problems holistically.
[0005] Specifically, the body vitals that are considered to estimate or measure certain physiological parameters often do not provide reliable and accurate results. Existing wearable devices also do not take into account important body parameters, nutrition data, mental health and stress levels, workout regimes, etc. Moreover, in terms of reference standards, the reference ranges with respect to the vitals and measurements provided by

wearable devices are typically population specific, and not subject specific. So recommendations on fitness, health and nutrition provided to wearers may not be specific to a user but based on analysis of a standard population. Therefore, a holistic approach for automatically analyzing the body parameters and providing accurate interpretations and predictions to the users in real time is necessary.
[0006] Various publications have attempted to address some of the challenges. US9427053B2 discloses a wearable device with a plurality of magnets that are magnetized through their widths or thickness. US20090105560A1 relates to a system and method which monitors physiological parameters in people, and scheduling activities based on the monitored parameters. US20140073486A1 discloses physiological measurement systems, devices and methods for continuous health and fitness monitoring. US9949691B2 relates to systems, methods and devices utilizing flexible and stretchable electronics for sensing and analysis. However, these publications do not discuss wearable health tracking devices that address the aforementioned problems.
SUMMARY OF THE INVENTION
[0007] According to one embodiment of the present subject matter, a wearable
electronic device for improving health and fitness is disclosed. The device includes a memory unit, one or more sensors configured to measure a plurality of body parameters of a user, and a processor coupled to the memory unit. The processor is configured to determine one or more physiological metrics based on the plurality of body parameters. The processor collects a plurality of user data from one or more data sources and compares the one or more physiological metrics with one or more user-specific thresholds. Next, the processor is configured to predict one or more recommendations based on the compared physiological metrics and collected user data. The device also includes a display unit configured to display the predicted one or more recommendations to the user.
[0008] In various embodiments, the device includes a speaker configured to recite
the predicted recommendations to the user. In some embodiments, the one or more data sources comprise: a database, social media database, a smartphone, and a medical record. In some embodiments, the one or more sensors include heart rate sensor, blood pressure sensor,

step count sensor, calorie sensor, sleep analysis sensor, temperature sensor, motion sensor, and pressure sensor. In various embodiments, the device includes a communication unit configured to communicate with one or more smart devices, databases, and servers. In some embodiments, the memory unit includes a learning module configured to train one or more learning models and recommendation models.
[0009] According to another embodiment of the present subject matter, a method of
improving health and fitness using a wearable device. The method includes measuring a plurality of body parameters of a user using one or more sensors in the wearable device. One or more physiological metrics are determined based on the plurality of body parameters. The method involves collecting a plurality of user data from one or more data sources and comparing the one or more physiological metrics with one or more user-specific thresholds. One or more recommendations are predicted based on the compared physiological metrics and collected user data and displayed to the user.
[0010] In various embodiments, determining one or more user-specific thresholds
includes: receiving historical body parameters of a user from one or more data sources.
determining one or more historical physiological metrics based on the historical body
parameters; collecting a plurality of user data and historical user states from one or more
data sources; determining a relationship between the user states and the user data, historical
body parameters and historical physiological metrics; estimating one or more user-specific
thresholds based on the determined relationship. In some embodiments, the one or more
physiological parameters include heart rate variability, energy expenditure, central aortic
pressure and cardiac output, and VO2max. In some embodiments, the method includes
reciting the predicted recommendations to the user.
[0011] This and other aspects are described herein.

BRIEF DESCRIPTION OF THE DRAWINGS [0013] The invention has other advantages and features, which will be more readily apparent from the following detailed description of the invention and the appended claims, when taken in conjunction with the accompanying drawings, in which:
[0014] FIG. 1 illustrates an environment of personal health tracker and wellness advisor device, according to an embodiment of the present subject matter.
[0015] FIG. 2 illustrates a perspective view of a wearable electronic device for improving health and fitness, according to one embodiment of the present subject matter. [0016] FIG. 3 illustrates a block diagram of the wearable fitness and health tracking device, according to an embodiment of the present subject matter.
[0017] FIG. 4 illustrates a block diagram for aggregating data from different data sources, according to an embodiment of the present subject matter.
[0018] FIG. 5 illustrates the wearable device with components on a distributed network, according to an embodiment of the present subject matter.
[0019] FIG. 6 illustrates a flow diagram for a method of improving health and fitness tracking device, according to an embodiment of the present subject matter. [0020] FIG. 7 illustrates a flow diagram for determining user-specific thresholds, according to an embodiment of the present subject matter.
[0021] FIG. 8 illustrates user states for a day, according to one embodiment of the present subject matter.

DETAILED DESCRIPTION
[0022] While the invention has been disclosed with reference to certain
embodiments, it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted without departing from the scope of the invention. In addition, many modifications may be made to adapt to a particular situation or material to the teachings of the invention without departing from its scope.
[0023] Throughout the specification and claims, the following terms take the
meanings explicitly associated herein unless the context clearly dictates otherwise. The meaning of “a”, “an”, and “the” include plural references. The meaning of “in” includes “in” and “on.” Referring to the drawings, like numbers indicate like parts throughout the views. Additionally, a reference to the singular includes a reference to the plural unless otherwise stated or inconsistent with the disclosure herein.
[0024] In the following detailed description, references are made to the
accompanying drawings that form a part hereof, and that show, by way of illustration, specific embodiments or examples. Although various aspects of the disclosure will be described using with regard to illustrative examples and embodiments, those disclosed embodiments and examples should not be construed as limiting.
[0025] The present subject matter in its various embodiments relates to methods, devices and systems for tracking, improving, and managing health and fitness. The device and method analyze body parameters and provide accurate subject-specific predictions in real-time.
[0026] A system network for tracking and managing health and fitness of users is illustrated in FIG. 1, according to an embodiment of the present subject matter. The system network 100 provides a cloud based machine learning platform for improving health and fitness of users of the platform. The system network 100 may primarily include one or more wearable devices 102-N, one or more trainer devices 104-N (hereinafter referred to as “first user device”), one or more manager devices 106-N (hereinafter referred to as “second user

device”), one or more data sources 108-N, and a server device 110, communicatively coupled over a network 112.
[0027] Each user uses the wearable device 102 and owns the first user device 104 to track and manage their own health and fitness data. The devices 102, 104 may communicate over the network 112 with other devices, for example, the user may automatically share health and fitness data with a second user device 106, owned and operated by a trainer, a coach, and/or doctor. The wearable device 102 and first user device 104 may also automatically synchronize with the one or more data sources 108-N to store, collect, retrieve health and fitness data. The devices 102, 104, and 106 may also communicate with the server 110 configured to host a health and fitness management platform wherein users, such as trainers, athletes, students, employees, coaches, teachers, gurus, mentors, and the like, may collaborate and update latest health and fitness statuses in real-time.
[0028] A wearable electronic device 102 for improving health and fitness is illustrated in FIG. 2, according to one embodiment of the present subject matter. The interactive wearable personal health tracking device 102 may continuously track a user’s body parameters, such as vital signals, to estimate physiological parameters and provide alerts and personalized recommendations on fitness and health. The wearable device 102 has a unique design in the form of a wrist band, which may include an inner ring 202, which may be in contact with the user’s skin all the time, and an outer ring 204 having a user interface 206. The user interface 206 may include one or more user selection options 208 in the form of buttons, text, or icons. The user interface 206 may also include a graphical user interface element 210 indicative of body parameters, physiological information, recommendations, etc. The device 102 may include various sensors to capture data seamlessly without any drop in signal and the signal is transferred for processing and communicated back to the user through the device’s outer ring 204. In some embodiments, the device may include wrist band strap made up of silicone material. The silicon straps may be non-toxic, no smells, not harmful to skin. The strap may be durable, with good tensile, wear-resistant, easy to clean, and provides comfort.
[0029] A block diagram of the device 102 is illustrated in FIG. 3, according to one embodiment of the present subject matter. The device 102 may include a memory unit 302,

one or more sensors 304 configured to measure a plurality of body parameters of a user, and a processor 306 coupled to the memory unit 302. The processor 306 may be configured to determine one or more physiological metrics based on the plurality of body parameters. The processor 306 may collect a plurality of user data from one or more data sources and compare the one or more physiological metrics with one or more user-specific thresholds. The processor 306 may predict one or more recommendations based on the compared physiological metrics and collected user data. The device 102 may also include a display unit 308 configured to display the predicted one or more recommendations to the user. In some embodiments, the display unit 308 may include one or more of vibration motor, control buttons, a touch-sensitive display like mono color OLED display (96 x 24 pixels).
[0030] In various embodiments, the device 102 may include a speaker 310 may be configured to recite the predicted recommendations to the user. In some embodiments, the one or more data sources comprise: a database, social media database, a smartphone, and a medical record. In various embodiments, the device 102 may include a communication unit 312 may be configured to communicate with one or more smart devices, databases, and servers. In some embodiments, the one or more sensors 304 include heart rate sensor 314, pressure sensor 316, proximity sensor 318, axis accelerometers sensor 320, blood pressure sensor 322, step count sensor 324, and sleep analysis sensor 326.
[0031] In one embodiment, the sensors 304 may continuously capture blood pulse waves from the skin surface of the user, using pulse transducers including, but not limited to, photo plethysmogram (PPG), force sensors and impedance plethysmogram (IPG) implemented in the tracking device.
[0032] In some embodiments, the memory unit may 302 include a plurality of modules, such as goal setting module 328, data collection module 330, a data cleaning module 332, a learning module 334, a physiological metric computation module 336, a threshold computation module 338, and a recommendation module 340. In some embodiments, the modules may be implemented as software code to be executed by the processor 306 using any suitable computer language. These software codes may be stored as a series of instructions or commands in the memory unit 302. In various embodiments, the modules

may be implemented as one or more software modules, hardware modules, firmware modules, or some combination of these.
[0033] The goal setting module 328 may be configured to receive one or more targets the user wants to achieve. The goal setting module may be configured to display a plurality of types of goals related to general fitness, weight loss, performance, injury recovery, and the like, on the display unit 308 for user selection. The user may select one or more of the goals and the goal setting module 328 may subsequently track the progress made by the user to achieve the target and provide regular updates and alerts to the user for achieving the target.
[0034] The data collection module 330 may be configured to receive and retrieve various types of data associated with the user from one or more data sources 108-N. A simplified diagram of the various data sources is illustrated in FIG. 4, according to one embodiment of the present subject matter. Data may be broadly classified as general user data 402 and body parametric data 404. User data 402 may include, but not limited to, user’s personal health conditions data, medical record data, genetic data, user’s images, etc., and body parametric data 404 may include, but not limited to, such as vital signs, sleep cycle data, etc. The user data 402 and the body parametric data 404 may be associated with timestamps, which may be used to classify some of the data as historical data 406.
[0035] The data may be obtained from various sources including medical records 108-1, which may include patient data, surgery history, drug allergies, etc. The data may also be received from a user device 108-2, for example the user may manually send data from a mobile smartphone. The body parametric data 404 may be received from the sensor data 108-3. Some of the examples of the body parametric data may include heart rate, heart rate variability, body temperature, breathing rate, etc. Other sources of user data may include images, content, media from the web, including social media networking websites 108-N.
[0036] After the data is classified as historical data and current data, the data is prepared to obtain structured data. In one embodiment, the data may be segregated into training dataset, validation dataset, and the test dataset. The segregation may be performed randomly or according to a timestamp associated with the data. For instance, the older data may be

apportioned to the training dataset and the relatively newer data may be apportioned to the validation and test datasets. The data preparation 408 involves various steps of data preparation – data cleansing, data standardization - on the collected data. The prepared data may be stored in the master database 410, which be utilized by other modules of the device 102.
[0037] Referring back to FIG. 3, the physiological metric computation module 334 may be configured to determine physiological metrics, such as energy expenditure, central aortic pressure, cardiac output, blood pressure, VO2-max, and the like. based on the plurality of body parametric data. In some embodiments, the module 334 processes the recorded signals and provides beat-to-beat pulse interval, equivalent to inter-best intervals (RRI) of ECG in a continuous manner to estimate heart rate variability (HRV). Further, blood oxygen level and VO2max may be used along with the HRV to estimate the real-time energy expenditure (EE) of the individual or calories burnt.
[0038] Further, the physiological metric computation module 334 may be configured to determine VO2max using HRV, HR and/or other recorded vital signals. The module 334 may be configured to determine the blood pressure from the radial artery using the recorded heart rate and the received PPG signal. The module may continuously track the blood pressure from 24 hours ambulatory BP values and considers the the difference in the blood pressure from the radial and central arteries along with HR, HRV, and respiratory quotient. The module 334 may determine central aortic pressure using VO2max and oxygen saturation level in the arterial system. The determined central aortic pressure combined with heart rate could be used to obtain cardiac output in real-time.
[0039] The memory unit 302 may include a learning module 336 configured to train one or more learning models and recommendation models. Machine learning techniques may be implemented in the plurality of modules. In some embodiments, the plurality of modules may be machine learning algorithms, such as linear regression, logistic regression, decision trees, support vector machine, naïve Bayes, k-nearest neighbors, random forest, etc. The learning module 336 may use the training dataset to develop a prediction model, which may be fine-tuned using the validation dataset. The testing dataset may be used as input to the fine-tuned prediction model to make new predictions. In various embodiments, the

prediction model may be configured to predict recommendations for users, user’s performance, injury recovery estimate, etc.
[0040] The threshold computation module 338 may be configured to determine user-specific thresholds instead of population specific thresholds. The threshold computation module 338 may be configured to obtain user data, user states, and historical data from the one or more data sources. User states may include the user activity, mood, emotions, and the like. The module 338 may be configured to determine a relationship between the user states and the user data, historical body parameters and historical physiological metrics.
[0041] The module 338 may estimate one or more user-specific thresholds based on the determined relationship. The user-specific thresholds may vary for each user based on the user data, user states, and historical data of the user. Since each user’s physiology is different, the device determines different thresholds for providing recommendations. For instance, the nutrition and training data for a female athlete during normal time and during the period cycle of the month may be different. Data associated with the user’s states or situation may be used by the device and accordingly provide information on the training, diet and recovery for the female athlete.
[0042] The recommendation module 340 may be configured to provide one or more recommendations to the user. The recommendations may include objective identification of athletes’ highest and lowest abilities, objective identification of readiness and fatigue, and the like. Monitoring recommendations may be given percentage of training increase or decrease associated with generic recommendations for better recovery (sleep, recovery technique, Nutrition). The use of various body parameters, such as energy expenditure, heart rate, heart rate variability, activity information, and user state enables accurate predictions and recommendations. The continuous tracking of the vital data also allows to predict events and to predict game day performance of the athletes.
[0043] In various embodiments, the device 102 enables female users to track their menstrual cycle as part of personal informatics. The method predicts the various phases of the menstrual cycle of the female user using HRV, which is also an indicator of the level of stress. The information on the menstrual cycle will help the user to track her health through

the menstrual phase and the method further provides recommendations on nutrition and activities to relieve stress. The recommendation algorithm also considers the individual’s recovery time.
[0044] In some embodiments, the device may also include a language module (not shown in figure) configured to display information to the user in a plurality of regional languages, such as Tamil, Hindi, Punjabi, in addition to the official languages. This may allow engaging more users using protocols, supported research papers, and videos that act as training tools athletes, coaches and researchers to get trained which will be backed by science.
[0045] A distributed configuration of the fitness and health tracking is illustrated in FIG. 5, according to another embodiment of the present subject matter. In the distributed configuration, the plurality of modules may be implemented in one or more devices in a network. As shown, the learning module, recommendation module, and the processing module may be implemented outside of the wearable tracking device.
[0046] A flow diagram for a method of improving health and fitness using a wearable device is disclosed in FIG. 6, according to one embodiment. The method includes measuring a plurality of body parameters of a user using one or more sensors in the wearable device at block 602. One or more physiological metrics are determined based on the plurality of body parameters at block 604. The method involves collecting a plurality of user data from one or more data sources at block 606, and comparing the one or more physiological metrics with one or more user-specific thresholds at block 608. One or more recommendations are predicted based on the compared physiological metrics and collected user data at block 610 and displayed to the user at block 612. In some embodiments, the one or more physiological parameters include heart rate variability, energy expenditure, central aortic pressure and cardiac output, and VO2max. In some embodiments, the method includes reciting the predicted recommendations to the user.
[0047] In various embodiments, determining one or more user-specific thresholds includes receiving historical body parameters of a user from one or more data sources at block 702. The method includes determining one or more historical physiological metrics

based on the historical body parameters at block 704. A plurality of user data and historical user states are collected from one or more data sources at block 706. The method involves determining a relationship between the user states and the user data, historical body parameters and historical physiological metrics at block 710. Finally, the method includes estimating one or more user-specific thresholds based on the determined relationship at block 712.
[0048] In various embodiments, the body parameters and the physiological data may include heart rate, heart rate variability, activity or user states, step counts, speed, VO2max, energy expenditure, and food intake information. The heart rate may be measured from the heart rate sensor and monitored continuously 24/7. The user may gauge the effectiveness of their workout using visualizations in real time on the wearable and historical data. The device may also provide stats like average heart rate, maximum heart rate tagged with the activity performed, etc. When the user pushes their heart rate to their maximum heart rate a notification alert is sent to the user.
[0049] HRV may be tracked to calculate stress using stats like base heart rate variability, stress undergone, and recovery that that body has undergone. The method correlates stress with activities for correlated visualization of energy burnt and then recommends the number of hours of sleep required to recover to the fullest. HRV may be monitored 24/7 to provide these physiological statistics.
[0050] Activity or user states may be tracked by using accelerometer which is highly reliable and optimized to work on a wide range of users. The method may track activities such as walking, running, biking, deep sleep, light sleep, workout on/off, rest. By mapping this data with time, HR and HRV deep insights may be to the users. The different states of the user determined based on the body parameters and physiological metrics on a given day is illustrated in FIG. 8.
[0051] Further, the step counts may be calculated to understand the energy expended by an individual. Speed may be derived by the method which processes the accelerometer data. The method may use speed to predict the VO2 max parameter and we also tag this data with running activity to get more insights on HR and HRV rise. Additionally, food intake may

provide data about calorie intake. With this information, the method may provide statistics on total calorie intake and energy expended for the day.
[0052] Examples
[0053] Monitoring of HR of athletes continuously during physical exercise is crucial since heart rate variability highly reflects the intensity of the exercise and gives insights on the customization of training protocols to improve the fitness efficiently. The different HR zones depending on the intensity of the exercise are: Low exercise intensity: less than 50% of your maximum HR, Moderate exercise intensity: 50% to about 70% of your maximum HR, and Vigorous exercise intensity: 70% to about 85% of your maximum HR.
[0054] In order to estimate the heart rate of a person during these exercising zones, a study was conducted using the wearable device wrist-based Optical Heart Rate (OHR) tracker and Polar H10 chest strap was used as the reference device. Also, a separate study was conducted to validate the activity classification by the device during different exercise zones because monitoring of physical activity is essential in the case of athletes, since the level of physical activity and sedentary behavior off-training of young athletes may reveal the quality of recovery from training and highlight health-related issues. It is also useful for personal training professionals and coaches who work with athletes to monitor their training programs.
[0055] The objective of the study was to validate HR by the device against HR from Polar chest strap, collecting data from wrist based OHR tracker during different exercise phases to validate the HR zones against the reference Polar chest strap, and validate activity classification by the device during multiple activities.
[0056] Example 1A: Heart Rate Validation
[0057] The study participants included volunteer satisfying the following criteria:
[0058] Age of Men: 15 years - 40 years
[0059] Age of Women: 15 years - 50 years
[0060] Meet the ACSM criteria for low risk
[0061] BP ≤140/90 mmHg

[0062] 13 subjects were selected for the study with a mean age of 22.76 ± 2.24 with mean height of 172.46 ± 8.67 cm and mean weight of 81 ± 12.25 kg. As the HR measured by wearables is affected by skin tones, the scale of skin tone is also considered. The firmness of the grip around the wrist is dependent on the muscle mass around the wrist and the skin hair under the wearable.
[0063] HR data was collected from 13 subjects for a duration of 10 minutes using the following protocol:
[0064] Phase 1- Resting phase 0 - 2min: The subject was asked to sit in a comfortable manner and relax for 2 minutes. The baseline HR data of the subject is collected during this phase.
[0065] Phase 2 - Walking phase 2 - 4min: The subject was asked to walk at a speed of 3km/hr on a treadmill for 2 minutes. During this phase, the HR of the subject increased greater than the base HR. The subject reached moderate/vigorous exercise intensity zone during this phase.
[0066] Phase 3 - Resting phase 4 - 6min: After walking phase, the subject was asked to relax by sitting in a comfortable manner for a period of 2 minutes. Here, HR decreased and tried to come back to the base HR value.
[0067] Phase 4 - Running phase 6 - 8min: The subject was asked to jog/run at a speed of 6km/hr on a treadmill for 2 minutes. Again, an increasing trend in HR was noted during this phase. The subject even reached his/her maximum HR during this phase.
[0068] Phase 5 - Resting phase 8 - 10min: The subject was again asked to relax for a period of 2 minutes to bring back the HR to the normal state.
[0069] Beat-to-beat HR was computed from the raw data obtained using the device at a sampling frequency (fs) of 25Hz. Pre-processing techniques like signal smoothing and wavelet denoising were implemented to remove the noise and motion artifacts from the signal and to improve the signal quality for HR computation. In the next step, valid peaks and their corresponding locations were extracted from the signal and beat-to-beat HR was computed using the formula:

[0070] HR = (sampling rate / Difference in number of samples between consecutive peaks) *60
[0071] Also, HR values averaged over 3-seconds and 5-seconds were also computed to compare the accuracy of the HR against the Polar reference device.
[0072] Beat-to-beat HR, 3-seconds averaged HR and 5-seconds averaged HR computed using the above formula were plotted against the HR obtained from the reference Polar chest strap. The corresponding HR comparison plots for some of the subjects were plotted.
[0073] Example 1B: Heart Rate Variability validation
[0074] Heart rate variability (HRV) is the physiological phenomenon of variation in the time interval between heartbeats. HRV is a measure that indicates the variation in your heartbeats within a specific timeframe. The unit of measurement is milliseconds (ms). HRV was measured by the variation in the beat-to-beat interval, also known as the inter-beat interval. As a side note, HRV can be computed from R-R interval from ECG signals, and in the case of the PPG signal, it can be measured from the inter-beat interval (IBI) between two pulse waves.
[0075] The HR varies relative to the body's physical needs, such as the need to absorb oxygen and excrete carbon dioxide, physical exercise, sleep, anxiety, stress, illness, ingesting and drugs. Furthermore, at a given heart rate, women typically have a higher heart rate variability than men. There are no generic guidelines for optimal HRV values – which is understandable considering there are several ways to both track and calculate it. HRV is becoming one of the most used training and recovery monitoring tools in sport sciences. The possibility of applying HRV on such variety is based on the fact that cardiovascular autonomic regulation is an important determinant of training adaptations, before also being responsive to training effects. In general, athletes exhibit a different HRV profile to sedentary control subjects, with an overall increase in HRV and parasympathetic cardiac modulation.
[0076] Each contraction of the heart results in a blood volume pulse that propagates through the bigger arteries towards the small capillaries. The blood volume change is tracked optically in case of PPG signals, which is based on the absorption of certain

wavelengths of light when reflected towards blood veins. When HRV is tracked through PPG, the inter-beat-intervals (IBI) are measured, i.e., the start of a new interval is the steepest increase in the blood volume signal prior to its actual peak.
[0077] HR was computed from the IBI in ms obtained by the OHR tracker using the formula, HR = (60/IBI)*1000 in ms
[0078] Example 1C: Activity classification validation
[0079] Accelerometers and gyroscopes are being used widely for activity classification, step count and also to calculate the energy expenditure by integrating vertical acceleration over time. Activity classification in coordination with the heart rate data was used to determine the HR zone of the individual and used to increase or decrease the intensity of the exercise of the athlete. The device classifies basic activities like walking, running, cycling, etc. and few fitness trackers can classify gym activities like rowing, elliptical, weight-lifting, dumbbells, etc. The device also classifies various activities like resting, walking, running, biking, random activity, and rhythmic activity. It also classifies various stages of sleep like light sleep and deep sleep. The following table shows the various activities and its respective class number as shown by the device.
Table 1: Various activities and respective class numbers
Class given by the device Activity
0 Rest
1 Other
2 Walking
3 Running
4 Biking
5 Other rhythmic
6 Sleep: other
7 Sleep: light
8 Sleep: deep
[0080] Staircase climbing, and combined activity were not defined as separate classes by the device, they were assumed as class 2 and class 1 respectively. The missing data points or samples were given class number 10 for ease of visualization.

[0081] The device was configured to collect data during various physical activities including walking, running, biking, etc. and validate the same. The volunteers for the experiment were selected based on the same criteria used during heart rate validation.
Table 2: Protocol for activity data collection from 20 subjects for a duration of 50
minutes
Phase Duration (in Activity Activity description
minutes)
Resting The subject is asked to sit in a relaxed and
1 2 comfortable manner for 2 minutes to record the
resting phase data
Walking The subject is asked to walk on a treadmill at speed
2 5
of 3km/hr for 5 minutes
Resting The subject is asked to sit in a relaxed and
3 2
comfortable manner for 2 minutes
Running The subject is asked to run/jog on a treadmill at a
4 5
speed of 6km/hr for 5 minutes
Resting The subject is asked to sit in a relaxed and
5 2 comfortable manner for 2 minutes or until the
subject is fully relaxed
Random Subject is asked to do some random activities like
6 5
activity writing, typing, talking, eating, etc. for 5 minutes
Resting Subject is asked to sit in a relaxed and comfortable
7 2
manner for 2 minutes
Rhythmic Subject is asked to do some rhythmic activities like
8 5 activity stretching exercises, bending exercises, etc. for 5
minutes
Resting Subject is asked to sit in a relaxed and comfortable
9 2
manner for 2 minutes
Biking Subject is asked to ride a bicycle at a constant
10 5
phase for 5 minutes
Resting Subject is asked to sit in a relaxed and comfortable
11 2
manner for 2 minutes
Staircase The subject is asked to climb 3 floors upstairs and
12 3
climbing downstairs
Resting The subject is asked to sit in a relaxed and
13 2
comfortable manner for 2 minute
Combined The subject is asked to do all the above activities
14 5 activity (walking + staircase climbing + jogging) in a
random manner for 5 minutes

[0082] After the data collection and activity classification prediction, the accuracy, precision, recall values were generated. The table below shows the overall accuracy, prediction, and recall percentage values for all the 20 subjects.
Table 3: Overall Accuracy, Precision and Recall percentage values for all 20 subjects
Class Accuracy (%) Precision (%) Recall (%)
0 88.17 87.47 69.78
1 73.04 40.39 24.12
2 76.62 41.87 74.48
3 94.33 84.52 56.1
4 88.95 48.22 80.90
[0083] Since HR is computed from PPG signals acquired optically, there were variations observed in HR for people with different skin tones. Activity classification works good for walking, running and biking class. It was also observed that there is higher data loss during resting phase.
[0084] The device may be designed bearing in mind the wearer’s comfort in wearing it at all times, without causing irritation to the skin and discomfort to the user. Unlike wearables in the space of personal informatics that estimate energy expenditure (EE) of the wearer, often by considering only the heart rate (HR) which has certain limitations and cannot be relied upon entirely to provide accurate results, the personalized health tracking system considers other important physiological parameters including heart rate variability and peak oxygen uptake. The invention is built robustly to ensure that the recommendations on fitness, nutrition and to reduce stress levels will be seamlessly passed on to users and fitness aspirants through the wearable tracking device and through an on-mobile application of the tracker to motivate users to achieve their fitness objectives.
[0085] The disclosed device may have wide application in the field in the field of sports, for example, in analyzing the real time physiological parameters of athletes. Particularly, it may allow coaches and physiotherapists in getting a complete picture of athletes’ performances and conditions, and estimate unnoted fatigue or potential injuries developing in body parts. Therefore, the device may not only be used in measuring real time physiological parameters of athletes accurately but also act as early warning system in alerting stakeholders about the potential injuries.

[0086] Although the detailed description contains many specifics, these should not be
construed as limiting the scope of the invention but merely as illustrating different examples
and aspects of the invention. It should be appreciated that the scope of the invention includes
other embodiments not discussed herein. Various other modifications, changes and
variations which will be apparent to those skilled in the art may be made in the arrangement,
operation and details of the system and method of the present invention disclosed herein
without departing from the spirit and scope of the invention as described here.
[0087] While the invention has been disclosed with reference to certain embodiments,
it will be understood by those skilled in the art that various changes may be made and equivalents may be substituted without departing from the scope of the invention. In addition, many modifications may be made to adapt to a particular situation or material the teachings of the invention without departing from its scope.