WO2016157217A2

WO2016157217A2 - Technological device to assist user in workouts and healthy living

Info

Publication number: WO2016157217A2
Application number: PCT/IN2016/000085
Authority: WO
Inventors: Pratik SARAOGI
Original assignee: Saraogi Pratik
Priority date: 2015-04-01
Filing date: 2016-04-01
Publication date: 2016-10-06
Also published as: WO2016157217A3

Abstract

A system to track activity of a user comprising: a wearable device comprising: a motion sensor and, a force sensor. It further relates to a remote unit configured to determine start and end of an exercise activity, identify an exercise, evaluate and compare exercise; the remote unit comprising a signal attribute computation engine; a data accumulator; a software module configured to identify an exercise performed by user; and, a software module configured to determine deviation in the exercise performed by the user; and, a communication module configured to transmit data between the wearable device and the remote unit. It relates to a method to determine start and end of exercise activity, identify an exercise and predict its accuracy. It relates to a method to analyse deviation in an exercise form. It further relates to a method to derive continuous heart rate and Respiration Rate of a user.

Description

TECHNOLOGICAL DEVICE TO ASSIST USER IN WORKOUTS AND HEALTHY LIVING

Technical field

The present invention relates to a system to track an exercise regime of a user which includes a wearable device with a plurality of sensors, a memory storage element and a display. The system further includes a remote server configured for exercise identification, comparison of the data received from the wearable device with the data stored on the server, analyzing deviation in the exercise form or imbalance in the physical parameters of the user and for transmission of the deviation or imbalance data to the wearable device. The system further includes a communication module for data transmission between the wearable device and the remote server. The invention further relates to a method to assist the users in their workout regimes using the system of the invention. This device can be a stand-alone device or be part of other articles such as gloves, shoes, socks, headbands, wristband etc.

Description of the related art

People normally are not able to track their workouts at the gym. Manual tracking is possible but too cumbersome and time consuming to be used in today's fast paced and tech-oriented world. Also, manual tracking is limited to user's recollection of the exercises performed and the form used and may not always be accurate.

Improper/incorrect workout forms can lead to severe injuries and a large proportion of the population can't afford to hire or consult an expensive personal trainer. There are also scenarios where people are not aware of what exercise regime to follow, along with correct weights and their corresponding forms. Consequently, digital tracking of actual workouts currently includes only basic information such as a heart rate. In recent years various technologies have matured to allow user to monitor and track their daily activity/ exercises and performance. US9171201 discloses a personal computing device to create an exercise analysis application configured to receive data of bodily motion of a user in three dimensions and recording the data of the user performing a defined exercise, to generate a statistical model for the exercise. The system of US 9171201 enables generating a statistical model of exercise based on the combination of data from sensors, to identify the type of the exercise. However, the data received does not fulfill the purpose of reliable and accurate tracking and monitoring motion or activity and performance of user. The system of US 9171201 does not provide for the communication between multiple wearable devices for performance tracking. US 9174084 disclose a monitoring device having a sensor array configured to measure physical activity. The device allows for identifying and analyzing exercise or activity with only the motion sensor array. The US Ό84 monitoring device classifies exercise and non-exercise activity only based on the data from the motion sensors such as the accelerometer and the gyroscope and not utilizing any other sensor data for motion identification or grading. Moreover, the monitoring devices of US '201 and US '084 does not provide for multiple analyses such as balance of movements and identification of exercise activity by considering motion and pressure data at the same instance for each type of exercise or activity

Moreover while these prior art devices provide for real time data transmission, the transmission of the data is found to be relatively slow and a device which not only provides an accurate analysis of the exercise and physiological parameters but also provides for a quick real time data transmission is desirable.

The current disclosed invention provides solution to override drawbacks of prior attempts. It is observed that the wearable device is further enabled to classify exercise and non-exercise activity and grade each activity for correctness using a plurality of sensors in novel intelligent correlated manner.

SUMMARY

The present invention is a system comprising a wearable device witha motion sensor for determining motion of a user and a Force sensor for determining force applied, grip pattern, weights lifted and pressure magnitude. The system further comprises of a remote unit configured to determine start and end of an exercise activity, identify an exercise, evaluate and compare exercises on the remote unit and a communication module configured to transmit data between wearable device and remote unit. The system further includes the wearable devices including physiological sensors that can gather information like heart rate, respiration rate, body fat level, hydration, body fat level etc. This data or information stored at wearable device in memory storage element and transferred to the remote unit via external device. Moreover, the device is provided with buzzer or vibrator and notification alert system to help the user be alerted on crucial and important physiological conditions. These conditions are relating to excessive heart rate or under optimum heart rate, or low hydration levels, non-ideal breathing patterns or nonuniform weight distribution by grip patterns etc. The said remote unit may be remote server, computing-cloud, remote processing station, or any other computing unit accomplishing required computation. The remote unit of system configured to transmits data acquired from plurality of sensors for analysis and trigger comparative analyzed alerts and notification information in wearable device or external device. The remote unit configured to determine start and end of exercise using algorithm to track the activity or exercise of user. This algorithm includes signal attribute computation engine (SACE) to compute attribute of data. Also the exercise data accumulator (EDA) to perform marker based exercise separation and software module to determine deviation in exercise performed by user. The method includes steps of pattern identification adapted to identify an exercise activity performed by user by identifying start of a specific activity. The method includes SACE, EDA and WGBEI functionally coupled to identify activity and improve exercise prediction accuracy. The application contains the database of different types of motions and grip patterns that correlates it to a specific exercise. The algorithms executed on the remote unit and communication of data with wearable device via external device assist exerciser for tracking regime.

The invention includes method to identify an exercise activity performed by a user. The invention further includes method to improve prediction accuracy. The invention further includes method to analyze deviation in the exercise form performed by the user. The method further includes the robust heart rate and respiration rate measurement using multispectral peak analysis. The multispectral peak analysis and adaptive noise cancellation. The heart rate measurement includes. Also the respiration rate calculation method from spectral analysis of pulsatile signal.

An object of the present invention is to provide a device to identify, track and analyze the different types of exercises performed and to help guide users towards the correct form of range of motion and pressure application.

Another object of the present invention is to provide device which may gather information of the exerciser such as force applied/weights used, exercise time, number of repetitions/sets etc.

Yet another object of the present invention is to provide device which can gather and store biological and / or physiological information like heart rate, body fat level, hydration, respiration rate etc. Yet another object of the invention is to provide device which keeps track and analyze one's fitness activity using external computing device.

Yet another object of the invention is to provide a wearable device in plurality and provide for communication between such plurality of devices and then with remote server.

Yet another object of the invention is to provide system of alert and notification such as text notification, image of correct form/way of exercise, haptic vibration alert, audio alert, etc. in external device. Yet another object of the invention is to provide a device that intelligently tracks motion activity of user by remote server based comparative analysis and triggers interaction in a way that user learns and corrects the motion activity and grip pattern instantly.

Yet another object of the invention is to provide system of one or more wearable devices that tracks, analyzes the exercise activity using remote server where the algorithm executed for exercise and non-exercise classification and further form evaluation. The Summary is not intended to be used to limit the scope of subject matter, nor is it intended to be used to recognize features of claimed subject matter.

BRIEF DESCRIPTION OF DRAWINGS

For a more complete understanding of the invention, reference should now be made to the embodiments illustrated in greater detail in the accompanying drawings and described below by way of examples of the invention. Figure 1 shows system have wearable device 9with an instrument inserted on the inside surface of the gloves with comprising an on-board storage unit 4 on which data controller 4a and memory 4b is embedded.

Figure 2shows system have technical wearable device 9 with an instrument inserted palm side surface of the gloves with Weight/Force Sensor 5.

Figure 3shows system have wired/wireless communication module 3 which depicts the flow of the signal from the article/instrument of exercise to mobile phone/laptop/computer to remote server and getting instruction for exercises.

Figure 4 depicts system for the data transfer or communication module wherein said wearable device 9 is linked to remote unit7 over a network or external device 6.

Figure 5 shows system for readings on board motion sensor module 2, wired/wireless communication module 3, on board storage element 4 along with weight or force sensor 5 and the collected information or essential data 11 such physiological information like heart rate, body fat level, hydration, pressure applied etc.

Figure 6 and Figure 7 depicts the system for data/ information during weight lifting exercises, way of weight lifting is detected by the wearable device 9, collected and sent to remote server viz/via wired or wireless network. Figure 8 represents the system for maintenance of essential data 11 and physiological data 10 at data controller or memory device 4 and transfer of said data to the remote server.

Figure 9 depicts system for segregation of data depending upon the essential data or physiological data and analysis of the same by means of a flowchart.

Figure 10 depicts system with representative examples of the alerts and notifications in case of deviation in exercise from ideal or universally accepted forms and suggestion depending on the workout may be in the form of audio buzzers and/ or vibrations.

Figure 11 depicts the method of signal attribute computation engine (SACE) that extracts multiple attributes of signal by simultaneous and parallel processes 17.

Figure 12 depicts the method of accumulation of exercise data from workout session by using start 22, stop marker 23.

Figure 12A shows illustrative examples of start marker and stop marker.

Figure 13 depicts the method for workout group identification by weightage based exercise Predictor 47 using multiple attribute sets 43 and Identifier Networks 45.

Figure 14 and figure 15 depicts grip analysis by predefined pressure distribution 61 and improper pressure distribution 69 during lifting or carrying the weights 8 and quantifying the weights lifted with minimum disturbance to the user/ exerciser.

Figure 16 depicts the method for grip analysis mode and weights quantification mode

Figure 17 depicts the method for selection of LEDs or combinations thereof according to desired out-put and depending upon the type of skin.

Figure 18 depicts the method that detects whether pulsatile signal 87 is acquired or not and also manages the battery consumption 92. Figure 19 depicts method for selection of heart rate spectrum 99 by adaptive noise cancellation and mutli-spectral peak analysis 97. Figure 20 depicts the method for determining 105 heart rate peak and value by comparison of current and previous spectra 103, pulsatile signal 104 and relative motion data peak 106.a

Figure 21 A, 21 B, 21 C depicts graph of PPG signal data, Relative motion data and data from motion sensor module 2, time domain to frequency domain spectrum conversion, selected heart rate spectrum respectively.

Figure 22A, 22B depicts baseline PPG signal waveform and illustrative respiration rate spectrum having respiration peak. Figure 23 depicts method for estimating and displaying respiration rate by checking spectrum of respiration rate 113 and method for notifying 117 user to stay steady to estimate respiration rate.

It is to be understood that the drawing is not to scale and is schematic in nature. In certain instances, details which are not necessary for an understanding of the present invention or which renders other details difficult to perceive, may be omitted. It should be understood, of course, that the invention is not necessarily limited to the particular embodiments illustrated herein.

DETAIL DESCRIPTION OF PREFERRED EMBODIMENTS For a better understanding of the present invention, reference will be made to the following detailed description of the invention which is to be read in association with the accompanying drawings.

In describing the embodiment of the invention which is illustrated in the drawings, specific terminology is resorted to for the sake of clarity. However, it is not intended that the invention be limited to the specific terms so selected and it is to be understood that each specific term includes all technical equivalents that operate in a similar manner to accomplish a similar purpose. Unless otherwise defined, all technical terms having same meaning and commonly understood by a person with ordinary skill in the art.

An exercise may be used interchangeably with "activity", although "activity" has a broader meaning than exercise in that it includes any human activity that involves a physical aspect to it, including sitting down, standing up, sleeping, lying down, walking, jogging, running, etc.

Devices of system is unless otherwise defined, all devices having same use or function and commonly understood by a person with ordinary skill in the art.

As used in this specification and the appended claims, the singular forms "a," "an," and "the" include plural references unless clearly dictates otherwise. Any reference to "or" herein is intended to encompass "and/or" unless otherwise stated. The invention relates to a system to track activity of a user. The system comprises a wearable, device which include a motion sensor for determining motion and a force sensor for determining - force applied, grip pattern, weights lifted and pressure magnitude. The system further includes a remote unit configured to determine start and end of an exercise activity, identify an exercise, evaluate and compare exercise. The remote unit further comprise a signal, attribute computation engine to compute attributes of signal data received from the force sensor and the motion sensor; a data accumulator enabled to perform marker based classification with a start marker for determining start of an exercise activity and with an end marker to determine end of the exercise activity performed by user. The remote unit further includes a software module configured to identify an exercise performed by user and a software module configured to determine deviation in the exercise performed by the user and transmit an error alert message on the deviation to the wearable device.

The system further includes a communication module configured to transmit data between the wearable device and the remote unit.

The system of current invention communicates acquired data from wearable device to an external device. The external device referred here may be any computing device for example smartphone, laptop, desktop, etc. The data from such external device transferred to a unit which is capable of handling huge data in order to store and analyze based on algorithms. Such unit may remote server, cloud-computing, work station, etc. In the description of disclosed invention such unit is termed as "remote server" and/or "remote unit" must be referred with mentioned meaning.

One advantage of this device of system is that only motion sensor and weight/force sensor can collect all the information such as weights used or pressure exerted/applied, exercise time (using a time sensor, e.g., an inbuilt clock), number of repetitions/sets, physiological information or data like heart rate (via a heart rate sensor), body fat level (via a body fat sensor), hydration (via a hydration sensor), weights and provide a means for the exerciser to follow the correct type of exercise by comparing with the maintained database of different types of motions and grip patterns, the advice provided by others, such as trainers. The exercises may be showcased on mobile apps. Another important advantage of this device is that the exerciser can also receive instructions for a new exercise with continuous monitoring.

The detailed description provided below accompanying figures. The detailed description provided in connection with such example but not limited to any particular example. The scope is only limited by claims and numerous alternatives, modifications, and equivalent are encompassed. In order to provide a thorough understanding, numerous specific details with gloves are set forth in following description. The gloves hereafter described as example of a wearable device of system and can incorporated in any desired article such shoes, socks, wristband and headbands and physio-therapeutic instruments like belt or bed. Figure 1, 2 and 3 describe the system wherein technological wearable device 9 comprises a force/ pressure sensor 5, storage element 4, data controller 4a and memory 4b, a motion sensor module 2, and a data/wireless communication module 7.

Referring figure 4, the technological wearable device 9 of system of present invention identifies, tracks and analyzes different types of exercises and thereby guides the user towards the correct form of exercise and range of motion of particular exercise and pressure. This wearable device 9is to be worn while exercising with or without weights. This wearable device 9can be in the forms of gloves, wristband, shoes, socks, and headbands and physiotherapeutic instruments like belt or bed. The weight sensor, force sensor and related features such as grip pattern identification and weights quantification may be optional in some of the wearable devices. The number of wearable devices send data wired or wirelessly in a wearable device to store data from all sensors at one memory storage element4bof wearable device. The system comprises of complete physical or body posture and motion analysis of the user. This structure analysis is in at least three dimensions X-Y-Z (3D). The technological wearable device 9is linked to remote server 7 over a network or external device 6. In another embodiment, the system comprises of complete physical or body posture and motion analysis of the user. This structure analysis is in at least nine axis X-Y-Z (3D). This motion analysis is based on the comparison of the physiological data 10 of the user which guides or instructs the user for particular exercises. The outcome of this analysis shows most preferred and universally accepted motion and posture of the exercise.

In certain embodiments, the present invention is also useful to maintain physique of the exerciser as feed with guidance and timely alerts.

The invention further includes alert and notification system by using wearable device and external device.

In another embodiment, of the present invention the analysis can also inform the user about the speed and time of the exercise i.e., whether the user is in regular mode, normal mode, required mode, in slow speed or high speed by buzzer or vibrant or any other alert massage with instructions for proper mode at external device 6.

Software program can be prepared in any programming language required by person skilled in the art. The result or instruction or notification can be in any best form which can be understood by the user regardless of language.

In certain embodiments, for personal exercise a software module configured or the bodily motion of a user performing as defined exercise in one, two, or three dimensions, based on the physiological data analysis or physical data analysis, based on time and speed of the user's performance, based on the statistical data comparison for the exercise, based on the user's qualification for fitness, guidance of professional coach or personal trainer or expert. In certain embodiments, the exercises which are analyzed are of weight training exercise, a dumbbell exercise, a pull-down exercise, a dead lift exercise, a barbell exercise, a pull-up exercise, a sit-up exercise. Exercise based on body weight, height, gender, wingspan, repetition, or information and the like. In certain embodiments, the physiotherapy or pain management or pain rehabilitation exercises which may be analyzed are of Passive Range of Motion (PROM) exercises, Active Assistive Range of Motion (AAROM) exercises, Active Range of Motion (AROM) exercise, etc.

In some embodiments wearable device 9comprises of heuristic engine (figure 1 1) (Signal Attribute Computation Engine hereafter referred to as SACE) configured to compute attributes of signal data that may be acquired from motion sensor module2and/or force sensorit is further quantifies and grades the said attributes. In preprocessing said engine may gather the three dimensional motion data of acceleration and angular velocity from motion sensor module 2along with force sensor data to determine grip pattern and roll, pitch, acceleration magnitude, pressure magnitude, pressure distribution, weight lifted are collectively derived. This derived raw data (roll, pitch, acceleration magnitude, pressure magnitude, pressure distribution, weight lifted) and three dimensional data each of acceleration and angular velocity may provoke a frame from above said raw elements. The data of motion sensor module 2 is composite of number of such frames. The said SACE engine may be executed on remote server 7.

In some embodiment, a data frame is processed at a time by the attribute computation process. In each frame processing, signal attributes from said raw elements are calculated by operating not only by parallel but a fixed size window 12 on each said element. For further processing and for selection of next data, next frame may be transitionally shifted with 80 % overlapping.

For extraction of signal attributes, location of input sample data 15 varies depending on if quantification or grading of the attributes to be performed. The appropriate input data processed for classification of exercise and non-exercise activities. In certain embodiments, information related to motion of particular exercise, exercise time, number of repetitions/sets etc. and physiological information like heart rate, body fat level, hydration, pressure applied and the information of weight used or force applied for particular weight lifting exercise is collected from on board motion sensor module 2 and weight/Force sensor 5 respectively and stored into on board storage element 4b (data controller or memory) at a wearable device from plurality of wearable devices.

In certain embodiment wearable device 9 configured with SACE (figure 1 1) to take the decision according to the various conditions and processing environment for simultaneous and parallel extraction of multiple attributesl7.

The illustrative example of the framing of data 16 for taking decision of window size 12 selection, overlapping 131evel and calculating related parameters such aswindow index, number of frames in the data to name a few. For every transitional shift from current window to next window, overlap level/boundaries and number of frames may be selected. In certain embodiment wearable device 9 comprises of SACE (figure 11) configured to take the decision according to the processing environment, strategic decision for the selection of number of frames and such other conditions for simultaneous and parallel extraction of multiple attributesl7.Such extraction of signal attributes achieves faster signal computation as the process is divided into multiple level of computationl8. Such heuristic computation assures multiple and accurate signal attribute. The computation signal attributes comprises of number of domains of data such as time, frequency and wavelet domain to name a few. The numeric single value of each attribute is calculated from all the windows together by processing entire signal data. Each set of such values of the signal attribute is stored in memory 19. All the said sets are combined and stored may be at the external computing device 6 or at the remote server 7 for further processing20. These signals attributesl7 and each set of values of the signal attributed are further used for building statistical model. This heuristic method configured for training and testing of classification networks.

In some embodiment Exercise Data Accumulator (EDS, figure 12) processes exercise event data from complete activity or workout session data set and accumulate it in segments s as per the type and mode of performing activities. In another embodiment the said segment classification is either in exercise or non-exercise. The said non-exercise activity may be a movement that does not adding physical efforts or general activities carry out during said session, like drinking water, walking, resting, handling a cell phone etc to name few. During any said session user performs numerous activity or movements other than exercise activity that classified and separated for further analysis.

The invention further includes the method to determine the start and end of an exercise activity performed by the user.

In some another embodiments, exercise data accumulator is to wring the only exercise events from a complete data set. For example accumulating the barbell bicep curls exercise data set (As depicts in figure 6) from complete workout session. The exercise data analysis reveals the presence of differentiating number of frames at the any instance for example at start and end of the exercise set that indicates and helps to identify particular type of exercise. Such differentiating frame is herein after termed as marker (figure 12 A).The marker based classification and window size 12 selection is easier to implement and obtain reliable results. As shown figure 12 the said marker based exercise event separation is implemented in EDS. Depending upon the type of the exercise, passing appropriate argumentsl2, 13 to the SACE (figure 1 1), the attributes set is extracted and fed into the marker identifier 21.

In another embodiment SACE (figure 11) is configured for passing arguments such as higher window size and moderate overlapping. Higher window size provides better attribute extraction for long duration exercise repeating activity data. Using higher window size and moderate overlapping percentage gives a better result as compared to conventional windowing logic.

The invention further includes the method to produce compatible vector model of attributes based on the at least one computation method. In another embodiment the said extracted attributes sets are then arranged as a frame and the marker classification step21takes place. Start marker identifier 22and stop marker identifier26 are neural network and/or ensemble learning network based binary classifiers that processes data at the said remote server 7. These classifiers are trained by supervised learning processes. The supervised learning process provides pre-identified markers with probable results are provided to the neural network and/or ensemble learning network. In another embodiment as and when a start marker identifier22 identifies the start of the exercise event, an associated index of attribute data is stored23. A raw data index is then calculated from attribute data index24. In some another embodiment the raw data is segmented utilizing the said raw data index25. The data is accumulated as said exercise segment data until a stop marker identifier encounters a stop marker26 in same execution. Once the stop marker is encountered, the segmentation process is stopped and data is accumulated together and registers as Nth exercise data segment 27. In further process the segment number counter (N) is incremented by one and then the process is repeated from step 22till end of data32. If end of data 32step is tme then process is terminated 33. In another embodiment if start marker 22 is not found, next attribute set is selected 34 and the process is repeated from step 21. In another embodiment, if start marker 22is not found till end of attribute data29 the process is terminated 28.

In another embodiment, if stop marker 26is not found after start marker22 till end of attribute data30, the segment is accumulatedas Nth exercise segmentand segment number counter (N) is increased by one then process is terminated31. In current invention, the marker based accumulation process performsbetter as compared to processing entire signal.

A workout or exercise group consist of number of muscles targeted exercises such as bicep specific exercise events, chest specific activities, shoulder specific exercises. These workout or exercise regimes can also be classified in to domains or sub-domains of exercise such as crossfit, kettlebell, aerobics, cardio, etc. A "workout or exercise group" should be read in its broader sense, than limiting only with the mentioned specific movements or classes of workout. These workout or exercise groups^may be performed during free-style workout, gymnastics, physiotherapy etc. The invention further includes the method to a method to identify exercise activity performed by the user. In some embodiment, a wearable device comprises of the Workout Group Based Exercise Identifier (WGBEI) (figure 13)configured to processes exercise data segment created by EDA (Figure 12) and classify them in to particular exercise as per the pre-selected workout group(s). Once EDA (figure 12) accumulates data 37 from motion sensor module 2 then EDA (figure 12) starts segmenting data in to exercise segments 38. These exercise data segments , always have markers that include starting index and ending index. For attribute computation of these exercise events by SACE, WGBEI provides classified data from ED to SACE (37, 38). The SACE here is configured for Moderate window size 39and high window overlapping40. In some embodiment of wearable device 9, the software module of WGBEI has the Workout Group Based Identifier Network and Attributes Set Selector41 that takes external input for workout group selection 42 from user/exerciser without disturbing the flow and rhythm of workout session. This workout group selection from said workout event is user defined. The minimum intervention of the user is required before or after the workout session to select workout group(s) from the provided options in wearable device 9. Based on selected workout group, the identifier network and attribute data set is selected41. This group specific selection step 41 segregates identification process in parallel branches (attributes set 1, attributes set 2, attributes set 3 an dup to nth step 43) Meanwhile the network training management system 44 trains each exercise identifier network 45 for predicting accuracy of workout event which involves training a dataset of different types of workout events and motions that are stored at remote server 7.

The invention further includes the method to improve exercise prediction accuracy. In some embodiment of wearable device 9, the software module of WGBEI has the exercise identification process which is executed in parallel branches. Each branch process may have its own said set of attributes and said identification network41. The set of attribute and trained identification network, selected in particular branch in a manner that are most suitable for selected workout group. Each parallel branch process which is a part of the multiple (and/or "N" number of) branches of parallel processes, that are capable to produce its own prediction with percentage prediction accuracy46. In some another embodiment of wearable device, the software module of WGBEI has the weightage based predictor47 to collect predictions and percentage prediction accuracy46 from each said parallel branch. The accurate prediction is achieved by determining highest weighage prediction and analyzing its overall percentage accuracy. In order to illustrate the weightage based prediction step 47, the workout event which is provided here does not limit the weightage prediction step.Example: If five said parallel branches are used in process and out of which three branches predicted the workout event, as a dumbbell bicep curl each with 50%, 60%, 40% accuracy and remaining two branches predicted as bicep concentration curl with 80%) and 90% accuracy then weightage based exercise predictor will predict exercise as dumbbell bicep curl as it has higher weightage in this case. The percentage accuracy will be only analyses in cases of conflicts or tie-ups. Therefore, the implementation of weightage based classifier increases the exercise identification accuracy by greater margins.

The invention further included the method to analyze deviation in the exercise performed by the user.

In some another embodiment of wearable device 9, wherein workout group based Identifier Network is based on artificial neural network and/or ensemble learning network. The said group based Identifier network is trained with compatible vector model of attributes by using supervised learning method.

In some another embodiment of the wearable device9, the software module of WGBEI has the repetition counting process 48 that is based on the exercise detected by the previous stage. The number of repetition of particular exercise or activity may be quantified by executing autocorrelation of the acceleration magnitude and/or pressure magnitude and/or change in grip pattern.

In some another embodiment of the wearable device9, the software module of WGBEI has the form analyzer 49 that compares current detected and identified exercise with the maintained standard form database of different types of motions and/or grip pattern 77and/or pressure magnitude at remote server 7. Once the exercise is detected, the form analyzer 49 compares the exercise with the stored standard exercise form data set and grade the input exercise form 50. In some another embodiment, the grading is completely comparative and the form score generated 50 here is based on how much exercise form matches with standard data set that are stored at remote server 7.

In some another embodiment the process including step 41,43,45,46, 47, 48, 49 and 50 is repeated and group count is incremented 53 till end of the groups 51. After end of the groups the process is terminated 52. The pressure magnitude and grip pattern changes during the movement or course of the exercise activity. In another embodiment of wearable device 9, grip pattern during workout is analyzed by using force sensors data. The pressure magnitude changes with time during the movement of the exercise activity. However, the grip pattern change is repeatable, redundant and consistent for the same exercise. This helps for identification of exercises where range of motion is similar.

The figure 14 depicts various points on palm side where the pressure magnitude changes as per type of motion activity or exercise. The graphs 54,55,56,57,58,59,60 depict the pressure magnitude change over time. The time for one repetition of particular exercise is referred here. These graphs shows, the change in pressure magnitude for a particular repetition time of exercise for various points on palm side 70. This information is used as raw data in exercise identification algorithm that assures correct and robust exercise identification. The remote server database has grip pattern and pressure magnitude data ofdifferent exercises that may be used to carry out for comparative analysis with data from force sensor. This can be described with an example ofinclined, declined and flat bench press with barbell (Figure no.7 for inclined barbell bench-press)wherein range of motion is same but grip patterns are different. In figure 14 graph plotted 61 of data fromthree different force sensors for multiple repetitions of exercise. For different activities different grip patterns can be observed. This difference is very helpful in identification of an exercise activity and this force sensor data is used as one of the raw data in SACE for attribute computation. These three sensors assembly is arranged and aligned for grip analysis and using grip patterns obtained after analysis the exercise identification is done. The identification is carried out by said SACE.

The identification of free-weight exercise is difficult because of limited and similar range of motion. The force sensor data provides a novel solution for this problem by using different grip pattern data as raw data along with motion sensor data for free weight exercises identification. For instance, during normal push-ups and triangular grip push-ups, the range of motion is limited and similar but force sensor provides different grip patterns for both exercises. This difference between grip patterns is used as key matrix for identification of normal push-ups and triangular push-up.

In another embodiment force sensor assembly is arranged and aligned comprising of variety of shapes and sizes at pressure zones so that the comprehensive applied pressure data acquisition and analysis may be carried out. The analysis of pressure applied data to overcome conditions where the high weights lifted has higher magnitudes of transition phase of graph (figure, no 15and figure no. 14) to quantify the weights lifted/carried. The analysis is carried out with data from all sensors together to correlate with set values and find out differential readings between them. The force sensor data is further analyzed for the pressure magnitude to track grip pattern during movement and exercise activity. In figure 14 the graphs 62,63,64,65,66,67,68 depict the pressure magnitude change overtime. The time for one repetition of particular exercise is referred here in these graphs. These graphs shows that the there is absence of constant pressure magnitude in transition (changing) values of curve with particular repetition time of exercise for variety of point on palm side 71. This analysis interpreted as non- uniform weight distribution from graph 69. The all graphs of data from each force sensor and graph 69 shows the non-uniform grip pattern. This observation and interpretation used for grip analysis of exercise.

In another embodiment, weight quantification and grip analysis 72carried out during either grip or weight mode 74. For weight quantification during exercise data is acquired from number of force/pressure sensors 73. For accurate determination of weights carried or lifted during an exercise user must be in stationary or in no motion position. The weight determination from the force assembly data carried out by statistical method. In the next step 75 these statistical methods may be mean, mode, median, variance, etc. As referred in figure 16, the above steps of algorithms carried out for weight mode operation. In another embodiment, the grip pattern is used for exercise form analysis. For form analysis grip pattern of identified exercise76 by WGBEI (figure number 13) is compared with standard grip pattern 77stored in remote server database. The comparison is carried out by correlation of standard and obtained grip pattern, if correlation index is found to be below certain threshold it is considered as wrong form of exercise. And then the alert is triggered 78.As referred in figure 13, the above steps of algorithms of grip mode operations are used in exercise form analysis.

Here the physiological sensor of wearable device described for the method of collection of data and execution of algorithm set for analysis. In certain embodiment wearable device 9 configured using an optical technique for the measurement of changes in the blood volume within arteries which is wavelength dependent variation in light absorption coefficient. In this method sensor includes detector for capturing light reflected by from arteries or capillaries to measure and analyze heart rate data. Here sensor includes the emitter, as source of light that is ranging from 440nm to 940nm wavelength may be used. The multiple light emitters (such as Light Emitting Diodes [LED]) and their respective wavelength in various combinations may be used depending upon skin tone, and other bio-physiological factors (as obstacles) such as blood-flow in arteries or capillaries is continuous and produces very low pulsatile waveform to obtain desired data. In the present invention, for capturing pulsatile waveform light wavelength range used is 440nm to 940nm. The light detector is used and output of detector is amplified using trans-impedance amplifier. The amplified output is given to the processor for the further estimation of the heart rate using robust algorithm. The light intensity selection at every instant is done^'with the help of processor unit only after the analysis of previous output of light detector output.

In an embodiment, as shown in figure 5, the wearable device 9of present invention comprises of on board motion sensor module 2, wired/wireless communication module 3, on board storage element 4 along with weight or force sensor 5. The on board storage element4 include, for example, magnetic tape, any other magnetic storage element, any other optical storage element, punch cards, paper tape, any other physical storage element with patterns of holes, RAM, PROM, EPROM, FLASH-EPROM, any other memory chip or cartridge, or any other mediumfrom which a external computing device6 can read. The on board storage element4 particularly include, NOR flash memory that allows for both direct code execution and data storage in a single storage element.

In an embodiment, a system of present invention may include wearable device 9 and remote server 7, connected over a network or external device 6. On board motion sensor module 2 and weight/force sensor 5 collects information or essential data 11 such as correct form of range of motion, pressure exerted/applied, weights used, exercise time, number of repetitions/sets etc. and physiological information like heart rate, body fat level, hydration, pressure applied. Wired or wireless communication module 3 transfers collected information from sensor to the external device 6.

Also described herein, in certain embodiments, external device 6 can be in form of smart phone/laptop/computer (or any other device, including portable and non-portable devices) wherein collected data is analyzed which transfers collected information to remote server 7. The remote server 7 transfers back information to mobile app 6 which further guides the user for workout plans and regimes. As shown in figure 4 & 5, technical wearable device 9 of present invention identifies, track and analyze different types of exercises. More particularly, as shown in figure6, the weight 8 lifted by exerciser are on the palm side of exerciser as shown in figure. 7, where the weights 8 are going to rest on the wearable device9 in such a way to be able to detect the force and all related information of the exerciser in three dimensions and the recorded data is analyzed at external device 6 sent to theremote server 7 for comparison with maintained databases, then corrected data is displayed at external device 6 with buzzer.

Referring figure 6 & 7, during weight lifting exercises, way of weight lifting is detected by device 9. If there is any mistake in weight lifting by exerciser, then there is change in pattern recorded at wearable device 9 and this change is detected at the wearable device 9 which is analyzed at external device 6 and the data is transferred to remote server 7 and then corrected data is displayed at external device 6. As the wearable device 9 has an audio outlet which guides and talks about the correct form of motion and weight 8 lifted if and when any form is of particular exercise is disturbed.

Figure 8 represents the maintenance of essential data 11 and physiological data 10 at data controller or memory device 4. As the essential data 11 and physiological data 10 are collected and stored at data controller 4a, which is then with the help of wired or wireless module 3 transferred to external device 6 for analysis. After analysis, the data is transferred to remote server 7. In certain embodiment, referring to figure.9, critical information such as heart rate 11 and basic information such as no. of reps/set will be processed at the onboard processor of the wearable device 9, while non-critical 11 or physiological data 10 such as exercise identification, body fat level and hydration, respiration rate will be processed on the app 6. Other non-real time data or non-critical data is sent to the remote server 7 for analysis.

Referring to figure 10, the data which is collected at external device 6 is analyzed and when responded back from the remote server 7 to external device 6 then audio buzzer and vibrator for alerts in external device 6 such as smart phone can be of various types such as if heart rate is not optimum yet then there is notification such as run little more. Also in absence of server connection critical information like heart rate, number of repetition count of exercise interacted on wearable device or smart phone. The interaction for such critical information may be in the form of text, audio, graphics, etc. If there is a notification such as check balance, then exerciser have to follow unique weight distribution and grip on weights. If there is any alert such as weights are slipping out, then exerciser have to strengthen grip. Further, notification such as drink some water, extend your arm to 180° while expansion is provided as audio buzzer or vibrator at external device 6. These are representative alerts in wearable device 9 and text notification at external device 6.

In certain embodiments, the connection between a wearable device 9, smart phone/laptop/computer 6 and remote server 7 is either Bluetooth based or Wi-Fi based (or other forms of IEEE 802.11 standards based, as is well known in the relevant arts) or LTE (long-term evolution) networks which provides improved coverage and throughput even in dense urban areas. LTE also promises significant improvement in network performance and operational efficiencies.

In certain embodiments, the audio port is also connected to the storage space where music files can be played.

This whole device with all physical parts is also supported with software application to fulfill the object of the present invention. The software is to collect the data of the user, to send it to the computer/server and analyze it with a revert including a message about the exercise to the user. Thus this invention is a combination of software and device giving guidance to the user.

In certain embodiments, in the absence of guide, complete onboard memory notification and a display are configured for personal exercise analysis, suggestions and information related to improve the efficiency of the current workout. Thus the analyzed data guide the exerciser for proper physical exercises form, order, regime, weights, body vitals, range of motion, etc.

In current embodiment of wearable device software module configured to analyze previous data and update data regularly daily. The daily updated data comprises of the information of weights used, exercise time, number of repetitions/sets, correlation or disturbance between both hands etc., physiological information like heart rate, body fat level, respiration rate, hydration, pressure applied etc.

In another embodiment, the user can put daily stored and analyzed data in initial mode or learning mode in order to keep track on previously done exercise and planned exercises. This embodiment comprises hardware wearable device 9 and required software module to analyze previous data and daily update of exercises being performed. This data comprises of information of weights used, exercise time, number of repetitions/sets, correlation or deviation from ideal pattern of motion between both hands etc., and physiological information like heart rate, body fat level, hydration, pressure applied etc. In this way the data can viewed in user friendly and user defined manner. The wearable device 9 can also work without remote server connection, wherein if the said remote server 7 is not connected with the device, still the device shows the user's essential or physical activity and physiological data with instruction for the exercises of the day. Exercise analysis carried out at remote server site/level comprises of link between essential data 11 and physiological data 10 wherein the system correlate critical and real time information and reset values that can be stored in device and notify the user by buzzer or vibrate for the proper or essential information. In certain embodiments, the wearable device of the present invention can be in form of hand glove or sock. The above sensor device, external device6, remote server or other system is applicable to both. The same system can be explored to any wearable device for exercise like kneecap, waist belt etc. For the wearable device of system data comparison can be inserted accordingly. Like this, the present invention is useful to maintain body in painless state. This device can also be light filled for alert to detect any kind of mistake in current ongoing exercise.

The invention further includes method for robust heart rate and respiration rate measurement. The method (figure 17) describes combination of light emitter selection based on diversity of skin tone and other bio-physiological factors. Here light emitters 79 (such as LED) selection 81 is primarily based on skin tone adaptation. The algorithm is devised to cover all the subjects having different skin color tone and different wrist/body surface thickness. The algorithm is also devised to check if the desired out-put of LED signal detector is present or not 84. There are three green colored LEDs with varying intensities are used and depending upon the analysis of light detector output, particular LED and/or combination thereof is selected for a particular user/exerciser 81. The number of such combinations of LED may be implemented to cover various intensities 82 for measurement of heart rate in diverse group of users. By checking the output of the light detector either a single LED or combinations of LEDs are selected 81. This step is carried out at every measurement set and hence device gets adapted with the situations for individual user/exerciser 85. If the desired output 84 is not acquired, then the undesired (out-put) data is discarded and step 82, 81, 80, 84 is repeated till the desired out-put is acquired. The heart rate zone for exercises is predefined in data controller and memory storage element. Therefore user may be assisted for achieving standard heart rate for exercise he is performing. There are three conditions which are determined depending on the nature of the output such as;

1. whether the device is worn by the user or not,

2. whether the device is kept on a rigid or an opaque base but not on the skin,

3. whether the device is kept in open space facing directly towards light.

In this algorithm, to determine whether signal is from the individual's body surface or not (i.e. whether the signal is containing HR data or not) pulsatile signal from light detector must be present. The consumption of battery can be managed by keeping device in open space facing directly towards light that help to save unnecessary LED lighting. There are three conditions such as whether the device is worn by the user or not, whether the device is kept on a rigid or an opaque base but not on the skin, whether the device is kept in open space facing directly towards light which are determined by the nature output so that the selection and combination of the LEDs can be carried out. In another embodiment the, detector output 86 checks for pulsatile signal 87, if the pulsatile signal (Also termed as PPG signal or photoplethyosmography signal) as shown in figure 21 A is present then the heart rate determination 88 is continued. The Figure 21 B shows the spectrum conversion of time domain to frequency domain. If the pulsatile signal is not present then the failure is counted 89 and checks the nature of the failure 90. Depending upon the nature of the failure the LED selection 81 is carried out.

In another embodiment if the wearable device 9 is not on the body but may be placed on an opaque surface or in open space, the number of failure(s) 91 is counted which results in to turning off of the LEDs (after certain number of failures) in order to manage battery consumption 92. If the device is on the body then LED selection step 81 is carried out again.

The output data from a light detector is in the form of PPG signal 93 and another output data 94 is in a form of skin motion noise data from the same detector are acquired. Along with heart rate data from detector, motion sensor data (which is mounted on the device) is acquired. Both the data 95 of motion sensor module 2 and relative motion data from detector are considered for adaptive noise cancellation (ANC) 96 from raw PPG signal. Multi-spectral peak analysis 97 is carried out on ANC data. In this the ANC data is compared and analyzed with raw relative motion data 94 from the detector and motion sensor data 95 from motion sensor module 2. The PPG signal _quality analysis is carried out that includes analysis of the nature of the raw PPG signal and PPG spectrum. Multi-spectral analysis 97 is basically determination of heart rate indicating peak 99 (As shown in figure 21 C) by analyzing all the peaks available in the spectrum of all the three basic data i.e. PPG signal, relative motion data from detector and motion sensor data 98. Decision for heart rate spectrum selection is taken if the signal quality is good 98. To determine whether the signal quality is good, signal analysis is carried out using DC amplitude of raw PPG, peak to peak signal strength and the number of PPG signal in spectrum. The DC level of the raw PPG signal should be within the upper and lower bounds. Also the peak to peak value of the raw PPG signal should be above the threshold limit, the number of peaks of the PPG signal should not be more than predefined threshold.If every situation is true or satisfied then the heart rate value 108 is displayed 107. If false then new signal data set is taken and heart rate value 108 for display is estimated from past data spectral analysis. If the signal quality is not good the data is discarded 100 and the next step for acquiring the new data 101 from the detector out-put 86 is carried out.

As shown in figure 20, after heart rate spectrum analysis when heart rate spectrum is selected 102 then in order to display heart rate values 108, PPG signal spectral strength at previous instant and PPG signal at current instant are compared. The basic value of heart rate 108 is dependent on the spectral analysis but when there are multiple peaks in the output of spectral analysis, the selection optimistic HR value 108 plays an important role. The value or peak of heart rate within the spectrum is determined 105 based on comparison of the previous HR value spectrum and current HR value spectrum 103, comparison between previous pulsatile signal and current pulsatile signal 104 and relative motion data peak 106. The determined heart rate value 108 is displayed in wearable device 9 or external computing device 6 orreceived in audio format.

The reflectance method is more sensitive to the oscillation in the venous pressure during the respiratory cycle. The amplitude of a PPG waveform (figure 22 A) at a particular wavelength varies synchronously with respiratory cycle. This phenomenon can be termed as amplitude modulation of cardiac synchronous PPG waveform or respiratory induced intensity variations in the baseline of PPG signal. Baseline fluctuations in the PPG signal with a frequency of 0.2-.05 Hz are synchronous with RR. For the measurement of the respiration rate, the respiratory induced variations in the PPG signal waveform is measured. The reliable respiration rate estimation from PPG analysis would improve user acceptability by reducing the burden of wearing large monitoring systems. As shown in figure 24, for estimation of respiration rate, it is important to analyze the complete data when there is no motion data 112. It is because motion spectrum peak (figure 21) and respiration spectrum peak falls in the same region or region nearby, which leads to corruption of respiration rate data. As motion data is having larger amplitude than the respiration rate data, the respiration rate data gets totally suppressed.

For determining respiration rate, the no-motion data must be true. At first the motion sensor data from motion sensor module 2 is checked and then it is confirmed that the no-moti n data is true. Then system checks whether the spectrum in respiration rate ranges from 0.2- 0.5 Hz 113.Then the system estimates peak 114 and displays the respiration rate 115. As shown in figure 2 IB, spectral analysis is carried out in the range lower than the heart rate or PPG spectral range. The spectrum is analyzed and respiration rate is determined and displayed. If no-motion data is false then no respiration rate is displayed 116 and it also notifies 117 the user to stay steady.

In some embodiment of wearable device 9 amore sensitive reflectance method with oscillation in venous pressure during respiratory cycle in comparison to less sensitive transmittance method with venous compression pressure that arising due to confounding effect of probe is used during transmittance.

Therefore in another embodiment, when respiration estimation mode is ON user/exerciser may get continuous feedback and alert notification at the wearable device 9, unless the data required to measure the respiration rate is acquired. The respiration cycle and amplitude of a PPG waveform exhibits proportionality in the nature in way of variations in the waveform. The amplitude modulation of the variations in the baseline PPG waveform 109 (figure 22 A) gives respiration rate information.

In some embodiment wearable device 9configured for estimation of respiration rate when there is no voluntary motions in order avoid baseline fluctuation of waveform 109 in the acquired data. As motion spectrum peak and respiration spectrum peak falls in the same region or region nearby (figure 21 B), the identification of respiration rate peaks is difficult and /as it leads to corruption of respiration data. The respiration rate determination is obligated for the least voluntary motions during gathering heart rate sensor data through PPG signal 93 so as to avoid superimposing of respiration rate peak (figure 21 A, 2 IB) with motion noise data. As in the most of cases amplitude of motion data causes suppression of respiration peaks. It is mandatory to check and conform that the negligible motion data is conflicting during respiration rate estimation mode. Spectral range of respiration rate value 111 is lower than the spectral range of heart rate 108 (figure 21C) in the complete spectrum obtained from PPG signal. The determined respiration rate is saved at remote server 7 and used for fatigue level estimation.

In one embodiment, the working of whole system using said wearable device 9 for analysis and tracking of exercise is described as following: The wearable device 9 identify, track and analyze the different types of exercises which is performed by exerciser and collect information regarding use of weights, exercise time, number of repetitions/sets. The wearable device 9 receives motion and weight/force information from the onboard sensor 2 and weight/Force sensor 5 and through the wired or wireless communication 3 collected information is transferred to a network connector 6 such as smart phone/laptop/computer. Now, smart phone/laptop/computer analyses the information (e.g., against a pre-determined and stored information such as exercises, range of motion, correct form of exercises, rehabilitator measures, etc.) and then sends information to remote server 7. The remote server 7 maintains a database of different types of motions and grip pattern/ pressure magnitudefor a particular exercise which is universally accepted or advised by the top trainers and correlates the received information with the maintained databases. Based on this correlation if there is any imbalance or deviation in the exercise performed by exerciser than universally accepted information related to weights used, exercise time, number of repetitions/sets etc., is detected by smartphone/laptop/computer/server 6,7 and which further guides the user for correct workout plans and regimes as audio functionality.

In certain embodiments, the corrected data is also provided with buzzer or vibrator or other notification alert system to help the user be alerted on crucial and important physiological situations relating to excessive heart rate, under optimum heart rate, low hydration levels, non- ideal breathing patterns or non-uniform weight distribution 'etc.

In certain embodiments, exerciser also receives instruction for the next exercise and an audio outlet or onboard display is provided which guides and talks about the correct form of motion and weights to be lifted. During activity with wearable device 9 dynamic force or pressure and three dimensional motion data is acquired. On comparative analysis of this data with database from remote server, user notified for proper grip on weights 8. In certain embodiments, the wearable device 9 of present invention continuously provide feedback on one's activity and helps exerciser to set the targets and drives accordingly in a right direction even from the remote server 7 also.

In certain embodiment for analyzing hydration level and body fat level,' body impedance analysis technique is used, where wrist located logical sensor is configured for estimation. The physiological sensors comprises of impedance measuring sensors/electrodes such as of hydration sensor and body fat sensor which may be in the form of metal or conductive fabric as example.

The said hydration sensor and body fat sensors configured to measure resistance/impedance using electronic circuits. The body fat level and body hydration level measurements from wrist are proportional to the body impedance, hence it can be estimated. Here, wrist located impedance measurement electrodes/sensors are configured to display the optional hydration level and body fat percentage in external device6 may be in the comprehensive report.

In certain embodiments, on board motion sensor 2 can be in form of accelerometers and gyroscopes, which track fitness on a comprehensive level.In certain embodiments, on board motion sensor 2 can be in form of heart rate tracker that changes color on a display or it can be the breathing pattern analyzer and the hydration tracker that alerts users if their hydration levels are too low and also estimate body composition by using a method called bioelectrical impedance analysis.

The impedance sensors comprises of at least four electrodes. Out of said four electrodes, two are sensing electrodes that may be placed on wrist and in-between the two current carrying electrodes. Other arrangement can be placed in such a way that current carrying electrodes can come in contact with another hand for torso fat and hydration level measurement. The comprehensive report at external device 6, wearable device display the body fat and hydration level can be displayed.

In certain embodiment, referring to figure 9 the physiological data collected can be processed and stored onboard in wearable device 9 in situation such as no remote server connections. In such cases the report for body fat level and hydration level information displayed on wearable device. During unavailability or connectivity obstacles of remote server connection fat and hydration level treated as essential datall.

In certain embodiments, wearable device 9 keeps on tracking and analyzing one's fitness activity and plan or suggests the exercise regime according to the analyzed data even from the remote server 7 which is connected to smart phone 6 and the device 9. Further, the wearable device 9 can also be provided with buzzer or vibrator and notification alert system to help the user be alerted on crucial and important physiological conditions relating to excessive heart rate 11 or under optimum heart rate 11, or low hydration levels, non-ideal breathing patterns or nonuniform weight distribution, grip on weights etc. Referring to figure 9, data collected from wearable device 9 can be of two types. First type is of essential data 11 including critical heart rate 11 and number of reps/set. Now, onboard sensor 2 senses data and processes at data controller element 4, further respective alerts triggered in external device 6. This data processed from sensor of device 9 and output data transferred to external device 6 for alerts such as vibration in external device 6 in case of elevated or high heart rate, respiration rate and audio buzzers and LED-displays in case of completion of reps/ set .The external device analyze data and triggers real time alerts and notification such as audio, buzzers or vibration at the wearable device 9depending on level of critical and severe conditions. If there is any critical deviation in exercise performed by exerciser than professional user also has the option to override such commands and turn this functionality off in above mentioned case. Second type of data is physiological data 10 such as body fat level, hydration, pressure applied which is stored in specific format at external device 6and then this data 10 is sent to remote server 7 where the data compared with maintained databases. Remote server 7 sent back the corrected data to external device 6 and the corrected data is represented as audio, buzzers or vibration at wearable device 9.

In certain embodiments, the collected data is just not processed at the remote server 7 only, but also processed at the mobile application 6 and at the wearable device 9, depending on the type of data.

In certain embodiments of wearable device 9,the remote server 7 configured to estimate fatigue level by analyzing all collected data from all the said sensors, comprising of analyzed values such as weights lifted, duration of exercise session, etc. and related physiological information such as heart rate, respiration rate, hydration level, etc.

Description of a Specific Working Example: A wearable device such as hand gloves is discussed in the current example tohave complete and clear understanding of the best way to operate disclosed method and system. Hand gloves illustrated in current example is not limiting the scope of invention but provides the best way of working of system and method. Before starting exercise, exerciser wears and syncs hand gloves with smartphone wirelessly. The smart phone connects to remote server. The user starts his session by warming up on the treadmill. He runs on the treadmill and gets audio and/or text notifications on the screen of the wearable the moment his heart rate enters the exercise zone. He also keeps checking what is heart rate is and what the distance covered by him already is. Post this, the user starts his gym exercise session. He decides to workout on biceps and chest today and selects that as the muscle groups on the wearable device and/or mobile App. He starts exercise session with first exercise as dumb bellbicep curl which is needed to be performed with both hands moving together during flexion and extension of arm. Exerciser starts reps of said dumbbell bicep curl meanwhile obtains one or more of the audio and/or text notifications like "Heart rate is not in zone", "May slow down movement during extension", etc. oft smart phone and on hand gloves display unit.

The system also displays image of the grip pattern on the smart phone by the ongoing exercise and/or after completion of exercise. Exerciser reacts to the notification by improving grip on dumbbells or by correcting his/her form and proceeds for exercise. Exerciser completes the first set of dumbbell bicep curl. The number of sets exerciser performs, displays on hand gloves display unit and on smart phone also. The system provides brief report on smartphone and informs exerciser for weights used, heart rate zone, and respiration rate during exercise and fatigue level. In between two exercises system alerts exerciser for dehydration level as "Drink some water", if necessary. Exerciser starts second set of dumbbell bicep curl and gets an audio and/or text notification for grip pattern on smart phone. The user learns for non-uniform weight distribution from image of grip pattern, said alert and then improves the grip. The smartphone plays audio as "Check Balance". Exerciser also gets a vibration alert on hand gloves. Exerciser checks and starts sets with proper weight distribution and completes all sets of dumbbell bicep curl. On completion of exercise the system provides report that informs exerciser exercise performed, number of repetition of exercise, number of sets of exercise, quality of form of exercise, intensity of exercise, duration of exercise, duration of rest, fatigue level of exerciser to name few. The alerts and notification may be displayed in smart phone software application as graphical representation of data such as pie-chart, histogram, line chart, etc., based on which userlearns and track exercise regimen. The analyzed values such as heart rate/heart rate zone, respiration rate, workout performance intensity, time and number of repetitions of exercise etc., may be displayed as grade, class, etc., in smart phone software application. The system and method counts only the number of reps of the dumbbell bicep curl that the exerciser performs successfully with full length of motion and displays on hand gloves onboard display unit. If user performs any random activities like carrying weights, drinking water, handling smart phone, etc. between exercise sessions then, the system eliminates all those random activities or non-exercise activities and delivers results specific to exercise activities only.

Before exerciser starts next exercise, barbell bench press, he/she gets an alert for heart rate and respiration rate by vibrations on hand gloves from last exercise brief report. . Exerciser rests and relaxes or continues to warm up until he/she gets alert of achieving optimum level of heart rate and respiration rate by vibrations on hand gloves. Exerciser wears wireless headsets (connected to smartphone) that alerts about achieving optimum heart rate and respiration rate by playing sound, such notification can be by stopped user. Exerciser can stop all the notifications triggered at the headset. Once the heart rate and respiration rate is optimum to carry out the next exercise, exerciser gets alerts such as "Get ready for next exercise" on smart phone. Exerciser wrongly performs barbell bench press deviating from interval duration oflowering the barbell that usually takes about twice of raising the barbell takes. The deviation of interval duration above threshold triggers alert for exerciser in brief report after completion of exercise. Exerciser improves and completes barbell bench press exercise accordingly.

On completion of the day's exercise session the system generates comprehensive report and displays on the smartphone through software application. The comprehensive report includes heart rate throughout session, type of exercise performed, number of sets and reps completed and weights used for each exercise, exercise performance of day in percentage and body fat level. Exerciser checks his/her progress in weekly, monthly reports and tracks his/her exercise regime.

Another example of disclosed wearable device as a wrist bandwhich the exerciser wears on wrist during exercise incorporated with onboard sensor modules. The method and system generates all possible reports and notification as similar as generates in hand gloves based on onboard sensor module does.

While the present invention has been described with respect to a number of preferred embodiments, those skilled in the art will appreciate a number of variations and modifications of those embodiments.

Claims

CLAIMS We claim:

1. A system to track activity of a user wherein the system comprises:

a. a wearable device comprising:

i. a motion sensor for determining motion of a user; and,

ii. a force sensor for determining force applied, grip pattern, weights lifted and pressure magnitude.

b. a remote unit configured to determine start and end of an exercise activity, identify an exercise, evaluate and compare exercise; the remote unit comprising:

i. a signal attribute computation engine to compute attributes of signal data received from the force sensor and the motion sensor;

ii. a data accumulator enabled to perform marker based classification with a start marker for determining start of an exercise activity and with an end marker to determine end of the exercise activity performed by user;

iii. a software module configured to identify an exercise performed by user; and,

iv. a software module configured to determine deviation in the exercise performed by the user and transmit an error alert message on the deviation to the wearable device; and,

c. a communication module configured to transmit data between the wearable device and the remote unit.

The system of claim 1 wherein the wearable device further comprises a physiological sensor module comprising at least one of a heart rate sensor to measure heart rate; a body fat sensor for determining fat composition of an individual; and, a hydration sensor for determining levels of hydration of the user.

The system of claim 1 or 2 wherein the wearable device further comprises a memory storage element for storing data received from the force sensor, the motion sensor or the physiological sensor module

The system of Claim 1, wherein the remote unit further comprises a software module to compare the exercise performed by the user with a standard exercise data stored on the remote unit.

The system of claim 1 , wherein the wearable device further comprises

a. a display for providing output of exercise related information or error alert messages to the user; and,

b. at least one of a buzzer, a notification system and an audio outlet to alert user on deviation of exercise form, grip patterns and imbalances or current status of biological and physiological conditions.

A method to determine a start and an end of an exercise activity performed by a user, the method comprising:

a. identifying start of an exercise activity with a start marker identifier ona data accumulator configured on the remote unit;

b. computing raw data index on the data accumulator from a raw data associated with the exercise activity;

c. segmenting the raw data using the raw data index computed in step b;

d. creating an exercise data segment based on calculated attributes from the raw data until a stop marker identifier on the data accumulator indicates an end of the exercise activity;

e. separating the exercise data segment; registering it with a unique number and storing it on a remote unit; and,

f. feeding data to Workout Group Based Exercise Identifier configured on remote unit.

A method to identify an exercise activity performed by a user, the method comprising:

a. receiving segmented data index from an exercise data accumulator ona remote unit;

b. calculating at least one attribute set from the segmented data index; c. feeding the at least one attribute set to a plurality of exercise identifier networks enabled to predict the exercise activity with percentage prediction accuracy; and, d. identifying the exercise activity based on weightage analysis of prediction accuracies from each exercise identification network by a weightage based exercise predictor.

8. The method of Claim 7, further comprising configuring the remote unit for quantification of attributes of the data stored on the memory storage element in multiple domains with windows of overlapping margins to produce compatible vector model of attributes, based on at least one computational method.

9. A method to improve an exercise prediction accuracy, the method comprising:

a. creating a plurality of workout groups based on at least one of targeted muscles or, exercise domain and exercise subdomain on a remote unit;

b. receiving information from a user of a workout group to be performed during a workout session;

c. selecting a workout group from the plurality of the workout groups on the remote unit based on the information received from the user in step b; and

d. identifying exercise activity performed by the user from the selected workout group limiting identification of exercise activity within the selected workout group to improve the exercise prediction accuracy.

10. A method to analyse deviation in an exercise form, the method comprising:

a. creating a database of at least one of correct exercise form and grip pattern on a remote unit; ;

b. obtaining force and motion sensor data from a user during an exercise activity; c. computing a plurality of attributes from raw data using a signal attribute computation engine;

d. segmenting the raw data using an exercise data accumulator;

e. identifying an exercise by using a workout group based exercise identifier; f. comparing identified exercise data attributes with the correct exercise form and grip pattern stored on the remote unit for that particular exercise; and, g. alerting users on deviation in the exercise form and the grip pattern via at least one notification sent to at least one of a wearable device and an external device via a communication module.

11. A method to derive robust continuous heart rate of a user, the method comprising:

a. adjusting intensity level of a plurality of LEDs in a Heart Rate Sensorbased onbio-physiological parameters of at least one user;

b. acquiring atleast one photoplethyosmography signal from a plurality of optical detectors in a Heart Rate Sensor placed on skin of user;

c. acquiring relative motion data from at least one optical sensor placed on skin of the user and motion data from a motion sensor;

d. refining photoplethyosmography signal obtained in step b using adaptive noise cancellation using motion sensor data obtained in step c as reference signal;

e. performing Multi Spectral Peak Analysis onthe refined signal data obtained in step d using a signal data obtained in step c to select heart rate spectrum peak based on a predetermined frequency range;

f. analysing signal quality by comparing current photoplethyosmography Signal with previous photoplethyosmography signal over nature of data, output signal strength and spectral analysis to eliminate incorrect samples; and,

g. calculating heart Rate from the selected Heart Rate Spectrum Peak in step eand performing frequency to time domain conversion.

12. A method to derive Respiration Rate of a user, the method comprising

a. adjusting intensity level of a plurality of LEDs in a Heart Rate Sensor based on users bio-physiological parameters;

b. acquiring at least one photoplethyosmography signal from a plurality of optical detectors in a Heart Rate Sensor placed on skin of user;

c. acquiring relative motion data from at least one optical sensor placed on skin of user and motion data from motion sensor to ensure user is stationary; d. performing Spectral Analysis over photoplethyosmography signal in a predetermined frequency range; and,

e. calculating Respiration Rate from selected Spectrum Peak in step d and performing frequency to time domain conversion.