Vous êtes sur la page 1sur 6

2015 IEEE International Symposium on Nanoelectronic and Information Systems

A Smart Gym Framework: Theoretical Approach


Abhay Jain
Department of Computer Science and Engineering
Indian Institute of Technology (BHU)
Varanasi, India
abhay.jain.cse11@iitbhu.ac.in

Abstract In this paper, we propose a framework for a smart ubiquitous devices. This technology must ensure a reduced
gym. The framework consists of a simple model and prototypes power consumption which will make it energy efficient so
which can be implemented in a gym. The framework is as to support low carbon economy.
basically an implementation of the smart city based on IOT.
The foremost objective of the proposed framework is to
d) Privacy concerns: IOT being a pervasive
develop an intrusive technology that will not only provide
valuable and innovative services to the user by integrating the technology usually produces a large amount of private and
user and the smart devices but also provide an efficient sensitive information about the user. In order to thwart any
utilization of the resources present in the gym. possible attack on the user privacy by intruders, we need to
maintain a strict check and protection system. This will
Keywords- IOT; Sports training; Tele training, Smart Gym; maintain the integrity of the information as well as resolve
the privacy concerns of the user.
I. INTRODUCTION
This work provides a conceptual framework of a smart gym.
Of late design and development of smart cities has attained The major problem faced by a user or a customer going to a
great significance especially in the context of urban gym is regarding maintaining a gym schedule. This schedule
development in developing as well as least developed is based on basic user objectives and progress made in the
countries. Smart cities comprise of ubiquitous devices direction of those objectives. The user inputs and the
which provide a new generation user-centric services
progress depend from user to user. The personalization of
affecting the way, the user communicates with the
surrounding environment. Following are the major issues workout schedule for a user becomes a costlier affair as it
and challenges which one confronts in the course of involves expert manpower which eventually increases the
development of smart cities: cost of the schedule. Another problem faced by the user is to
maintain the past history of the progress made by the user
which affects the present and future schedule. Further,
a) Design issues: At the inception of the development during the execution of the workout schedule, proper
process, the major challenge faced by the developer is utilization of the gym resources are one of the challenges
concerning the design of a prototype that fits and satisfies faced by the gym owner/instructor. The proposed
all possible requirements of every smart city. But framework attempts to address these issues.
practically, a single model cannot be emulated in all locales
due to varying social, cultural and economic conditions in
different smart cities. Therefore, we need to provide rather II. RELATED WORK
localized design patterns, according to the different With steady progress in the field of IOT, new applications
prevailing conditions as well the availability of resources to which are meant to improve the quality of life of an
different smart cities. individual are emerging. The proposed framework is based
on previous studies done in the field of strength training and
b) Commencement cost: Smart objects involve computer science. For instance, work has been done to
sensors/devices which communicate with users and services enhance the tracking of the user and keeping the trainer up
and perform various computations. These communications to date about the accuracy of movements performed by the
and computations with a high level of accuracy and user, using motion sensors [1]. A similar proposal is given
intricacy increase the cost of the system manifold. So, one in [2] with the main difference that the conception takes into
of our main objectives is to implement the system as account gathered data from sensor-equipped machines to
frugally and economically as possible. propose a fuzzy logic based evaluation of strength training
exercises. In [3] the authors proposes an application with a
c) Running cost: In general, we require small music recommendation and exercise scheduling system on a
footprint technology which can be embedded in various light and small smartphone which is used to prompt

978-1-4673-9692-9/15 $31.00 2015 IEEE 191


DOI 10.1109/iNIS.2015.32
participation in the interval training without any large or In order to achieve the aforementioned objectives we have
expensive equipment. Moreover in [4] and [5] the authors designed various modules in our system (Fig 1).Following
developed a low-cost framework for golf training using are the main components of our system:
inertial sensors and body sensor networks. In [6] the authors
propose a motion analysis framework which uses wearable Fitness Device: The fitness device acts as a communication
sensors to track the orientation of athletes body interface between the user and the system. It recommends
movements. Moreover, a normative data is generated which the user about the next set of exercise he needs to perform
is compared to the subjects motion data for any deviation in from different exercises which form a part of his dynamic
order to identify potential injury related factors. workout schedule. Along with this, it also keeps track of any
abnormalities in the vital stats of the user like blood
pressure, the rate of pulsation etc.

Gym Equipment: Gym equipment is the representation of all


the smart machines present in the gym. These smart
machines record the number of sets and repetitions
performed by the user, the force applied and the time-
elapsed. It acts as an interface between our system and the
workout machines present in the gym. Whenever other
components require any information regarding the
occupancy of the machine or the feedback of the machine
with respect to any user, they can interact with the smart
machines using this interface. Other system components can
also prescribe appropriate load to workout machines using
the Gym equipment module.
History Management: After a user performs exercise on a
particular machine all the information regarding the sets and
repetitions of the exercise is stored in the database. This
procedure resembles a typical bookkeeping procedure.
Whenever a new schedule is to be formulated it takes
cognizance of the past records of the user.
Identity Management: The Identity management module
manages the ID of different users by providing new users
Fig1: Class Diagram with new IDs and by deleting the IDs of old users.
Whenever a user enters a smart gym environment the
identity management identifies the user and retrieves any
III. SYSTEM DESCRIPTION appropriate information about the user from the database.

The system has been designed to achieve the following three Level Management: The Level management module
objectives: manages all the levels which are present in our system.
These levels are dynamic and have completely flexible
Personalized workout regimen: To provide user with boundaries. The flexibility of these levels is managed by the
personalized workout suggestions which are generated by level management component. In addition to this, the level
the system based on the vital statistics, machine feedback management module is also concerned with promotion or
and the workout history of the user so that the user obtains demotion of users to different levels based on their
optimum gains according to his/her fitness goals. performance on the set targets.

Personalized bookkeeping and comprehensive tracking: As Workout Management: Workout management is the part of the
system where the dynamic workout for each and every user is
suggested in the previous objective to raise the level of
generated using the information retrieved from other system
effectiveness accomplished by the exercises it is necessary
components. The information comprises of the history related
to track the user activities in a comprehensive manner as information along with the machine scheduling and feedback
well as maintain individual history records of each and related information. The level management module puts the
every user. user into a particular level and provides this level regarding
information to the workout management component. The
Improve the user involvement: We need to adjust the workout management utilizes all this information and creates a
workout regimen according to the available resources and dynamic workout for the user.
thereby increasing the involvement of the user with the
surrounding environment.

192
Occupant recognition: Occupant recognition component of the machine. This prediction along with precedence
identifies whether a machine is in idle state or not. If an constraints of the exercises is used to formulate a dynamic
occupant is present, the occupant is recognized using the workout for the user. For the purpose of implementation, we
unique user ID and the time since when the machine is divide our proposed framework into 3 different parts (Fig2).
engaged by a particular user is also identified.

Feedback generator: Feedback generator generates the


feedback based on the number of sets and reps the user
ought to perform and the number of sets and reps that the
user actually performed.

Scheduler: Scheduler component schedules the workout


such that in a multiuser and multi-machine environment, the
most time efficient workout is recommended to all the users.
In addition to this, it takes care of particular exercise
preceding another exercise as a necessary constraint.

Machine Management: Machine management module


interacts with the gym equipment directly and retrieves
information which it further supplies to occupant
recognition, feedback generator, and scheduler module.
These modules format the information according to their
requirements and provide the machine management module
with the formatted information. All this information is
supplied to the main module whenever asked for.

Main module: Main module directly interacts with the user


via the fitness device. It provides the user with
recommendations along with the dynamic workout which
the user needs to perform. In addition to this, it also
manages different information from other components.

IV. IMPLEMENTATION
Fig 2: Connectivity Diagram
The aforementioned components in our system work in
tandem in order to provide a reliable workout schedule to A. INTERFACE
the user. Whenever a user registers for a gym membership, The gym equipment modules, as well as the fitness device
his/her fitness device is provided a unique ID and a
relationship is established between his/her fitness device module, provide the machine and user interface
and the gym equipment. A new database entry is made for respectively. The fitness device gives specific instructions to
the user which contains the users age, height, weight, BMI, the user regarding the exercises he/she needs to perform
Body fat% and other relevant information. Initially, the user along with the recommended repetitions and sets. If any
is asked to perform a set of predefined exercises to identify workout machine on which user needs to perform exercise
his/her capabilities. The system contains various flexible according to the schedule is already engaged, then in that
levels ranging from complete beginner to advance which case the user is not concerned about the occupation of the
are managed using the level management module. The
assessment of the user ability identifies the appropriate level machine and our system recommends another machine
in which the user must be placed. Every level has a based on the workout schedule. This recommendation is
predefined weekly schedule which comprises of a set of made keeping in mind any precedence constraints related
exercises to be performed on the daily basis. There are with the machine. The occupancy related information is
various constraints on the performance of exercises such as; accessed by the occupant recognition module from the gym
an isolation exercise must be preceded by a compound equipment module which acts as the general interface
exercise which in turn must be preceded by a full body between the system and the gym machines. The gym
warm-up. These constraints, if any, are explicitly mentioned
equipment module provides other information regarding the
in the predefined weekly schedule. The history information
of the user, as well as the machine feedback, is utilized to user such as the machine feedback about the user
make the workout schedule much more user specific. The performance, the load used etc.
sequence of the exercises to be performed is not completely
rigid and the schedule is more or less dynamic. Whether a
particular machine is occupied by another user and if so,
since when, is used in order to predict the future availability

193
B. Server C. Client
The server part of the implementation deals with the The client part of the implementation consists of the user as
management and various computations required to well as the gym machines. The Gym machines which are
formulate a user-centric workout. First of all, the user is equipped with Motion detection sensors and time detection
identified by the identity management module which sensors among other things count the number of repetitions
manages distinct user entries in the database. These entries performed by the user, force exerted by the user, and the
contain user specific information regarding the user history, time elapsed and provide this information to the machine
user level etc. Level management module manages various management module. This is utilized by the feedback
levels defined in the system which are very dynamic and generator and the occupant recognition to calculate their
flexible in nature. It assigns an appropriate level to the user respective outputs. The user is provided the
recommendation regarding the sets of exercise he/she needs
based on his/her past performance. The user level changes
to perform next. Entire workout schedule is not provided to
with any change in user performance parameters. The
the user in advance. Rather the user is informed about the
workout management module formulates the user workout exercise that he/she needs to perform next so that he/she
schedule based on various inputs from other modules. Any doesnt need to bother about future exercises.
history regarding information is obtained from the history
management module. The user workout formulated by the
workout management module is not completely rigid in V. EXPERIMENTATION
nature. Rather it takes into account machine occupation, To validate the above-proposed framework 15 participants
user related machine feedback and other machine related with varying degree of experience performed 10-12
inputs in order to incorporate various changes in the user repetitions on different loads on a squat machine (Table 1).
workout. The scheduler helps in machine scheduling so that According to the training data, so collected we divided the
in a multi-user multi-machine environment all the resources individuals into 6 different levels based on their age,
are utilized to their optimum level. The main module experience, sex, height, weight, and performance, assigning
integrates various processing related system components similar individuals to the same level.
while the machine management module integrates various a) Beginner male: Subjects 7, 8, 9.
workout machine related components. All these modules b) Beginner female: Subjects 12, 14, 15
working in tandem formulate a comprehensive workout and c) Intermediate male: Subjects 13,10,5
accordingly provide recommendations to the user about the d) Intermediate female Subjects 1, 3, 6
next exercise to be performed. e) Advanced male: Subjects 2, 11
f) Advanced female: Subject 4

Table 1: Detailed biographical characteristics of the participants, experience level and used load for each set

194
Based on his age, sex, height, weight and other performance
parameters the system assigns him to the beginner male
level. The workout on any particular day for Legs exercise
for a beginner male is 3 sets of squats of 10-12 reps in the
load range 60-80 kg and 3 sets of leg presses of 8 reps each
in the load range 70-85 kg. There is no constraint on the
precedence of the exercises and both the workout machines
are free so we do not need to worry about the scheduling
part. The Squat machine and the Leg press machine are
equipped with sensors as shown in figure 3. The user is
informed about the current set of exercise along with the
number of repetitions: both remaining and to be performed
(Fig 4). The sensors embedded on the machines calculate
the displacement, force and time-elapsed and provide these
parameters to the system at regular intervals with the help of
gym equipment module which acts as a communication
interface between the system and the workout machines.
Using these parametric values, we draw a Displacement-
Force-Time curve. Load Range is automatically set
according to the feedbacks provided by the machine at
regular time intervals, based on the user performance. Fig 4:The user interface
While managing the load range we try to make sure that the
Force-Displacement-Time curve remains as smooth as
possible.
We see that whenever an individual performs the exercises VI. CONCLUSION AND PRACTICAL CHALLENGES
based on the recommendations of a human trainer then in
In this paper, the theory for IOT based smart gym has been
that case curve is similar to the improper execution
proposed to fulfill the said objectives in section 3.The major
curve(Fig 5a).This is because the trainer is not able to judge
advantage of this implementation is the creation of a
the exact capability of the person and thus is not able to
personalized workout schedule for the user, for which a
adjust the load and the repetitions accordingly. The
tremendous amount of manpower is needlessly being used
improper execution curve exhibits variable displacement
these days. We can save a lot of man hours and money in a
paths and bigger force fluctuations. On the other hand, in
clear juxtaposition to the meager initial investment. The
case of system recommended exercise routine, the curve has
framework can be successfully applied at any well-
very little variations in force and displacement and is overall
established gym where the user and machine data are
smooth and consistent (Fig 5b).This shows that the system
already available. The framework requires highly
recommends a highly user-specific schedule which the user
sophisticated machines which can be equipped with IOT
is able to perform smoothly. This reduces the chances of any
sensors. Like other IOT implementations, the success rate of
injury to the user. In addition to this; we found out upon
a smart gym depends in the accuracy and efficiency of the
medical examination of the user, that the system
IOT devices and sensors.
recommended workout was much more efficient in burning The major challenges that can be faced during the actual
calories in comparison to the human trainer recommended implementation of our framework in a full-fledged gym can be
workout. multifaceted. The availability of the initial data required to train
the framework poses a great challenge here. Exception
handling is also one of the major concerns in this framework as
the interaction between man and the machine is highly
unpredictable. To attain and maintain the efficiency and
accuracy of the scheduler not only at the user level but also at
runtime is a gargantuan task. A good scheduling policy along
with an efficient synchronization technique is the foremost
requirement for the success of our framework. Also generating
a user-centric schedule, keeping in mind know-how in the field
of medical science together with that of physical education
imposes a tough challenge.

FIG 3: SENSOR EQUIPPED SQUAT AND LEG PRESS MACHINE

195
Fig 5: The Execution Curve

7) JosL.Hernndez-Ramos, M.Victoria Moreno, JorgeBernal


REFENCES Bernab, Dan GarcaCarrillo, Antonio F.Skarmeta.SAFIR:
1) J. D. Garca Prez Ubiquitous low-cost sports training system Secure access framework for IoT-enabled services on smart
for athletes EATIS 12 May 23 - 25, 2012, Valencia, Spain buildings. http://dx.doi.org/10.1016/j.jcss.2014.12.021
2011 ACM 0-12345-67-8/90/01. 8) MuddasserAlam, Alex Rogers, and Sarvapali D. Ramchurn.
2) Hristo Novatchkov at. el. Fuzzy Logic in Sports: A Review Demo Abstract: Towards a Smart Home Framework
and an Illustrative Case Study in the Field of Strength Training .Copyright 2013 ACM 978-1-4503-2431-1/13/11.
International Journal of Computer Applications (0975 8887) 9) Ashim Kumar Debnath, HoongChorChin , Md. MazharulHaque,
3) Myung-kyungSuh, Ani Nahapetian, Jonathan Woodbridge, Belinda Yuen. A methodological framework for benchmarking
Mahsan Rofouei, Majid Sarrafzadeh, Machine Learning-Based smart transport
Adaptive Wireless Interval Training Guidance System DOI citieshttp://dx.doi.org/10.1016/j.cities.2013.11.004
10.1007/s11036-011-0331-5 10) Hafed hChourabi at.el Understanding Smart Cities: An
4) Ryan Burchfield, S. Venkatesan, A Framework for Integrative Framework2012 45th Hawaii International
Golf Training Using Low-Cost Inertial Sensors DOI Conference on System Sciences
10.1109/BSN.2010.46 11) Sasa Savic, Hao Shi An Intelligent Object Framework for
5) Hassan Ghasemzadeh, Vitali Loseu Roozbeh Jafari, Wearable Smart Living doi:10.1016/j.procs.2011.07.05
coach for sport training: A quantitative model to evaluate
wrist-rotation in golf, DOI 10.3233/AIS-2009-0021
6) A.Ahmadi, E. Mitchell, C. Richter, F.Destelle, M.Gowing, N.E.
OConnor,K.Moran.TowardsAutomaticActivity
Classification and Movement Assessment During a Sports
Training Session, 10.1109/JIOT.2014.2377238, IEEE Internet
of Things Journal

196

Vous aimerez peut-être aussi