Content uploaded by Hristo Novatchkov
Author content
All content in this area was uploaded by Hristo Novatchkov
Content may be subject to copyright.
©Journal of Sports Science and Medicine (2013) 12, 27-37
http://www.jssm.org
Received: 29 June 2012 / Accepted: 02 November 2012 / Published (online): 01 March 2013
Artificial Intelligence in Sports on the Example of Weight Training
Hristo Novatchkov and Arnold Baca
University of Vienna, Centre for Sport Science and University Sports, Auf der Schmelz 6A, 1150 Vienna, Austria
Abstract
The overall goal of the present study was to illustrate the poten-
tial of artificial intelligence (AI) techniques in sports on the
example of weight training. The research focused in particular
on the implementation of pattern recognition methods for the
evaluation of performed exercises on training machines. The
data acquisition was carried out using way and cable force sen-
sors attached to various weight machines, thereby enabling the
measurement of essential displacement and force determinants
during training. On the basis of the gathered data, it was conse-
quently possible to deduce other significant characteristics like
time periods or movement velocities. These parameters were
applied for the development of intelligent methods adapted from
conventional machine learning concepts, allowing an automatic
assessment of the exercise technique and providing individuals
with appropriate feedback. In practice, the implementation of
such techniques could be crucial for the investigation of the
quality of the execution, the assistance of athletes but also
coaches, the training optimization and for prevention purposes.
For the current study, the data was based on measurements from
15 rather inexperienced participants, performing 3-5 sets of 10-
12 repetitions on a leg press machine. The initially preprocessed
data was used for the extraction of significant features, on which
supervised modeling methods were applied. Professional train-
ers were involved in the assessment and classification processes
by analyzing the video recorded executions. The so far obtained
modeling results showed good performance and prediction
outcomes, indicating the feasibility and potency of AI tech-
niques in assessing performances on weight training equipment
automatically and providing sportsmen with prompt advice.
Key words: Artificial intelligence, machine learning, pattern
recognition, weight training, feedback.
Introduction
The design and implementation of innovative systems on
the basis of state-of-the-art information and communica-
tion technologies in combination with sophisticated proc-
essing methods are getting increasingly important for the
instant collection, transfer, storage as well as analysis of
sensor data in sports. Moreover, the integration of ma-
chine-aided intelligence into the development of modern
sport information systems enables a prompt and automatic
evaluation of sport-specific parameter values, thereby
allowing the establishment of computer-based feedback
and intervention routines (Baca et al., 2009; 2012).
In general, artificial intelligence (AI) is derived
from imitating human actions and abilities such as think-
ing and learning. It involves the idea of designing so-
called intelligent agents or machines that are similarly
able to acquire, simulate and employ knowledge, analyti-
cal capabilities and professional skills for the overall
purpose of problem solving (Poole et al., 1998). While AI
techniques experienced a boom with the rise of expert
systems in the 1980s, such methods are meanwhile main-
ly applied for rather specific and isolated research topics.
Chess and particularly the first win of a computer against
a world champion (Deep Blue vs. Garry Kasparov) in
1997 (Newborn, 1997; Campbell, 2002) is an example
illustrating the high potential of AI. It has to be consid-
ered, however, that such achievements are strongly re-
lated to the constant increase of computer power – a main
feature and benefit of today’s information technology
environment.
Weight training fundamentals and research goals
The current paper focuses on the implementation of AI
routines for the automatic evaluation of exercises in
weight training. Weight training is commonly described
as a specific type of strength training where lifting
weights causes an overload of a certain muscle or muscle
group to trigger adaptive reactions of the organism (su-
percompensation). Also known as resistance training,
weight training is nowadays among the most popular
stabilizing, invigorating or even prime sport activities at
professional and amateur level. Positive effects of this
type of training include the overall strengthening as well
as the improvement of the physical condition, fitness and
performance levels. Thus, not only athletes but also non-
professional sportsmen can benefit from resistance train-
ing. Other known advantages of weight training include
prevention (e.g. in case of back pain or osteoporosis),
maintenance of muscular functional abilities, fat loss and
promotion of a healthy cardiovascular system (Winett and
Carpinelli, 2001). Therefore, weight training is often
recommended by specialists and professional organiza-
tions (Westcott, 2009), as it decreases the risk of injuries
and serves as effective prophylaxis compared to other
exercises.
Nowadays, plenty of diverse weight training ma-
chines exist, aiming at a controlled and hence less risky
execution. Particularly inexperienced individuals initially
prefer exercising on such equipment, due to their conven-
ience, easier use and supporting purpose in comparison to
free weights and in order to get used to the movement.
However, it is still important that experts like fitness
coaches assist and provide advice to beginners, inexperi-
enced and elderly people for the purpose of adjusting and
correcting the execution, preventing health and injury
risks as well as adapting and improving the overall train-
ing. Today’s weight training machines usually include
directions on how to use the equipment, visually illustrat-
Research article
Artificial intelligence in sports
28
ing and describing the proper technique (for example:
performing each repetition in a slow and constant man-
ner). Also in the literature, it is reported that the flexion
and extension phases should be executed smoothly and
completely with preferable time durations of 2-3 seconds
(e.g. Evans, 1999). Similarly, the velocity (in terms of a
constant and consistent movement) plays a crucial role for
a correct and low-impact execution (Rana et al., 2008).
Figure 1 represents an example of a typical begin-
ner’s mistake. It shows the measured parameters (cable
force and weight displacement) on an incline bench press
machine, illustrating inconstant and incorrect characteris-
tics with noticeable force fluctuations at the turning point
of a single repetition. The visible oscillation in the cable
force graph was most probably caused by a sudden re-
lease of stress and a consecutive loading, which is a
common error among beginners.
Based on such abnormalities, it can be concluded
that determinants like force, displacement, velocity and
duration are essential for the analysis of the quality of the
technique. In particular, these features may be applied for
the automatic evaluation of weight training exercises with
the help of sophisticated modeling methods. Moreover,
such routines could be integrated in an automated coach-
ing system, allowing real-time analysis of the quality of
the movement and returning prompt feedback informa-
tion. The notifications for the mistake in Figure 1 might
for instance alert on the occurred fluctuation point and
provide directions for error correction.
In the area of weight training, only few ap-
proaches integrating computer-based evaluation routines
have been presented in the literature so far. Although
Ariel (1984) suggested first ideas for the design of intelli-
gent weight training machines on the basis of AI methods
already in 1984, no effective realizations have been con-
structed up to now. The author proposed the implementa-
tion of a feedback-based system integrating factors like
duration, displacement and force characteristics of the
movement, thereby suggesting the most suitable exercise.
A more recent study (Chang et al., 2007) concentrated on
the recognition of various free-weight exercises by apply-
ing specific classifiers and models to measured accelera-
tion characteristics. The purpose of the developed meth-
ods, though, was to determine which but not how the
exercise was executed.
The research objectives of the present paper were
to confirm and demonstrate the high capability and poten-
tial of AI and, in particular, of machine learning methods
in the field of sports by the practical and still not well-
investigated example of weight training. A specific aim
included the assessment of measured way, force and fur-
ther derived characteristics on the basis of common pat-
tern recognition methodologies including automatic clas-
sification algorithms. First results involving rather basic
data analysis provided thereby a basis for the improve-
ment, optimization and extension of the developed ma-
chine learning routines (Novatchkov and Baca, 2012).
The current research, however, focused on the implemen-
tation of more enhanced, exact and in-depth modeling
techniques and outcomes by applying multi-scale feature
spaces and further classification types.
Based on the analysis conducted so far, the ulti-
mate goal is to integrate machine-aided techniques into a
mobile coaching system (Baca et al., 2010), providing
athletes with automated and instant evaluation and feed-
back notifications. This system integration would enable a
real-time data transfer – for instance via an Internet en-
abled portable device such as a handheld or tablet PC – to
a server component (see also section “Data acquisition
and equipment” in the chapter “Methods”), where the
measured parameters would be analyzed and assessed by
the developed models. Crucial notifications could be then
sent to the exercising person, giving feedback on the
quality of the execution as well as providing appropriate
Figure 1. Filtered cable force and weight displacement data of an execution by an inexperienced individual on a sen-
sor-equipped incline bench press machine collected at 200 Hz with a load of 30 kg. The red circles label the turning
point of the repetition, indicating the discrepancy of the measured channels with appearing force fluctuations.
Novatchkov and Baca
29
advices. This information could be presented to the per-
forming individual via a mobile device, indicating oc-
curred mistakes, suggesting corrective measures and in
this way reducing the risk of injuries. In an alternative
design of the mobile coaching system, the portable de-
vices could be replaced by an in-built computer device
including a screen, used for the instant transfer of the
measured information to the server component and the
prompt display of feedback alerts.
In the following, the overall method including rel-
evant AI fundamentals, main application fields as well as
the underlying study design and applied procedure are
presented. The rest of the article includes current results,
an outlook and final conclusions.
Methods
AI techniques in sports
Practical concepts for the realization of AI-based method-
ologies for sport science disciplines like biomechanics or
kinesiology have been already discussed and reviewed
earlier (e.g. Lapham and Bartlett, 1995; Bartlett 2006). A
commonly used technique involves the development of
methods on the basis of AI for the assessment of different
sport-related data measurements or game analysis. The
so-called TESSY (tennis simulation system) framework,
for instance, is one of the first knowledge-based decision-
making implementations, aiming at the supervision, proc-
essing and interpretation of results and tactical behavior
as well as the subsequent transformation of conclusions
into tennis practice (Lames et al., 1990). Other, more
recent, approaches also suggest the implementation of
expert systems integrating fuzzy logic procedures for
diverse purposes like the evaluation of the fast bowling
technique in cricket (Bartlett, 2006; Curtis, 2010) or for
the identification of sport talents (Papić et al., 2009).
Ratiu et al. (2010) provide an overview on the overall
application of AI in sports biomechanics, giving examples
of diagnostic tools for the evaluation of movements in
different sports.
Specific investigations, on the other hand, focus
on the design of machine learning methods for the cluster-
ing, classification, pattern recognition and prediction of
sport-specific data such as movement sequences. Today,
particularly performance analyses by means of self-
learning algorithms like artificial neural networks (ANNs)
are increasingly discussed as promising application areas
in the mathematics and computer science related sports
literature and fields of activities (Perl, 2004a; 2004b;
McCullagh, 2010). Successful implementations include
also analytical studies for different movement evaluations
in sports such as golf (Ghasemzadeh et al., 2009), base-
ball (Ghasemzadeh et al., 2011), and soccer or basketball
(Lamb et al., 2010; Bartlett and Lamb, 2011). As another
example, Silva et al. (2007) present predictive solutions
for the dynamic system modeling and talent identification
in swimming. Furthermore, in (Baca and Kornfeind,
2012) a self organizing map is trained for the purpose of
clustering stability of the aiming process of elite biathlon
athletes.
But also other classifiers like the k-nearest
neighbor (k-NN) algorithm or Support Vector Machines
(SVMs) are commonly applied modeling tools, providing
good opportunities for the analysis and recognition of
sport-specific data patterns. Acikkar et al. (2009), for
instance, use SVMs in their approach in order to predict
the aerobic fitness of athletes. A number of further studies
are related to running, aiming either at the in-built classi-
fication of track inclination and speed parameters
(Eskofier et al., 2010) or the identification of differentia-
tion of kinematic characteristics (Fischer at al., 2011).
Design and procedure
Based on the above described methodologies, the current
study was built on a typical machine learning approach
including the following distinct and successive phases:
data acquisition, preprocessing, feature extraction and
classification. Figure 2 shows in detail the connections,
purpose and significance of each step.
Figure 2. Applied machine learning approach.
In the following, all stages are described individu-
ally, giving an overview on used equipment, participants,
applied data processing as well as analysis and modeling
techniques.
Data acquisition and equipment
The data acquisition procedure was based on sensors
attached to various exercise equipment, allowing the
collection of crucial characteristics during the workout.
For the current study, a weight leg press machine was
equipped with a load cell (PW10A or PW12C3, Hottinger
Baldwin) and a rotary encoder (DP18, Altmann). In this
way, significant force and displacement parameters could
be measured directly and thereupon used for the detection
of single repetitions and extraction of further determinants
such as time periods, velocity, acceleration or power. The
data was acquired at a sampling of 100 Hz for each chan-
nel. In addition, a special sensor construction (NEON,
Spantec) with an integrated microSD card served as
collection point of the measured values. As the sensor
platform supports wireless sensor transmission on the
basis of the so-called ANT+ (Dynastream) protocol, one
future aim is to immediately forward the gathered
information to a handheld PC such as already available
smartphones with in-built ANT technology or a more
powerful laptop including an USB reception hardware.
Figure 3 illustrates the university sport’s hall with the
Artificial intelligence in sports
30
Figure 3. Sensor-equipped weight training machines and attached force (load cell) and way (rotary encoder)
sensors to a leg press machine.
installed weight machines and embedded sensors.
Participants
The current study examined executions by 15 individuals
with a rather inexperienced background in weight train-
ing, performing 3-5 sets of 10-12 repetitions on a leg
press machine. Descriptive details regarding the partici-
pants are shown in Table 1.
Table 1. General biographical characteristics of the study
participants.
Men (n) 8
Women (n) 7
Mean age (±SD) [years] 24.6 (2.7)
Mean height (±SD) [m] 1.73 (.10)
Mean body mass (±SD) [kg] 63.6 (13.8)
All subjects gave written informed consent for the
study procedures, which were reviewed and approved by
the research Ethics Committee of the Medical University
of Vienna.
Table 2 gives an in-depth description of the sub-
ject’s biographical characteristics, experience level and
used load for each of the performed sets. The participants
in this study were primarily inexperienced and slightly
experienced individuals, since the overall goals were to
assess the quality of the movement focusing on beginners
and to identify significant characteristics regarding the
performances of unskilled people. The initial intention
was neither to investigate the maximum load and force
potentials of the participants nor to examine the condition
and improvement in performance, power and muscle gain.
The actual research idea was rather to evaluate and clas-
sify the performed exercises according to crucial criteria
such as duration, constancy and completeness and not to
analyze the effects of different exercise methods. There-
fore, commonly recommended exercise methods with a
reasonable amount of sets, repetitions and loads for basic
resistance training were chosen to be followed (Garber et
al., 2011). The variable selection was chosen in accor-
dance with the literature-based importance of the men-
tioned criteria as well as today’s effective possibilities to
instantly measure, derive and assess significant parame-
ters by integrating modern sensors into the equipment.
Finally, another major aspect refers to the applicability of
the measured data to AI-based modeling techniques for
automatic classification purposes.
Data preprocessing
Once the raw sensor output was collected, the subsequent
step included the preprocessing of the acquired
Table 2. Detailed biographical characteristics of the participants, experience level and used load for each set.
Subject
Number Age Sex
Height
(m)
Mass
(kg) Experience
Load
Set 1
(kg)
Load
Set 2
(kg)
Load
Set 3
(kg)
Load
Set 4
(kg)
Load
Set 5
(kg)
1 22 f 1.62 49 Yes 30 40 40 50 60
2 30 m 1.72 66 Yes 110 120 110 100
3 22 f 1.63 46 Yes 40 50 60 70
4 27 f 1.68 55 Yes 50 70 80 90
5 21 m 1.80 73 Yes 120 100 90 90
6 21 f 1.80 63 Yes 60 60 60 60
7 23 m 1.73 80 No 40 60 60 80
8 24 m 1.76 72 No 40 60 90 120
9 27 m 1.78 71 No 40 60 60 70
10 24 m 1.72 79 Yes 50 100 110 120
11 25 m 1.90 72 No 50 70 100 120 140
12 31 f 1.62 43 No 30 30 30 30
13 27 m 1.93 85 No 40 60 70 80 100
14 25 f 1.64 51 No 40 40 40
15 20 f 1.62 49 No 40 40 40
Novatchkov and Baca
31
measurements. This procedure was necessary for the
preparation of the data without any loss of significant
information, the improvement of its quality and, conse-
quently, the final outcome and performance of the applied
machine learning method on the refined training set. In
particular, this comprised the processes of cleaning and
filtering the measured parameters. The cleaning routine
involved procedures such as detecting, correcting and
removing unreliable, incorrect and irrelevant data. Filter-
ing, on the other hand, had the goal of smoothing the time
series by the reduction of the effect of noise. A detailed
survey regarding the most useful preprocessing methods
is described by Kotsiantis et al. (2006). In addition, also
the segmentation of the gathered data was closely con-
nected with the preparation step, as it was essential in
forming the basis for the fragmentation (particularly into
single repetitions).
As it was necessary to preprocess the time series
before analyzing and classifying them, after several tries,
the measured displacement values were refined by a
strong low-pass filter (with a passband frequency of 0.1π
radians/sample, stopband frequency of 0.3π radi-
ans/sample, 10 dB of allowable passband ripple and a
stopband attenuation of 20 dB) in order to simplify the
process of segmenting the data. On the other hand, the
force input was essential for identifying possibly occur-
ring fluctuations (relevant for the classification of the
data) and was therefore smoothed by applying an average
digital low-pass Butterworth filter with normalized cutoff
frequency of 0.02 radians/sample, which is equivalent to 1
Hz. Figure 4 shows the measured and subsequently fil-
tered force and displacement data of a rather well-
performed and, in comparison, a poor execution (sta-
ble/complete vs. instable/incomplete movement range and
constant vs. inconstant time and force characteristics) of
the same participant.
Figure 5, on the other hand, illustrates the acquired
signals of an inexperienced in relation to a slightly ex-
perienced female individual, both with similar biographi-
cal data and identical load set-up (subject 1, set 4 and
subject 15, set 1 from Table 2). Obviously, the execution
of the complete beginner is characterized by variable
properties including bigger force fluctuations and variable
displacement paths, compared to the rather smooth com-
pletion of the slightly experienced participant.
The last task of the preprocessing step involved the seg-
mentation of the data into single repetitions. This proce-
dure was accomplished on the basis of the filtered dis-
placement measurements by detecting peak regions. In
particular, the filtered force characteristics were parti-
tioned into individual cycles by identifying extrema and
turning points within the entire data sets. Since some of
the first and last repetitions appeared to be interrupted
(e.g. by correcting the feet position just after the initial
extension or abandoning the final flexion phase) causing,
for instance, “incorrect” time intervals, these sequences
were not included in the classification process. Further-
more, the time series were divided into single stages (ex-
tension, flexion and holding phases), allowing precise
data analyses in the subsequent transformation procedure.
Figure 4. Comparison of the time series of improper (a) and proper (b) executions of an entire set by the same subject.
Artificial intelligence in sports
32
Figure 5. Comparison of the time series of an inexperienced (a) and a slightly experienced (b)
female participant with similar biographical characteristics, using the same load.
Table 3. Gathered parameters, main criteria and exemplary
features derived for the applied pattern recognition proce-
dure.
Parameters Criteria Features
Displacement
Cable force
Velocity
Acceleration
Power
Time
Completeness
Constancy
Extension/flexion/
reversal:
Durations
Maxima
Minima
Ranges
Relations
Fluctuations
Amplitudes
Inclines
Declines
Data transformation
In the following step, the initially cleaned, filtered and
segmented sensor measurements were applied to further
data analyses. This mainly included the identification and
deduction of relevant information, describing the gathered
time series. This feature extraction procedure aimed at the
data characterization by dimension reduction, detecting
and deriving crucial attributes (e.g. durations, amplitudes,
fluctuations or ranges) within the segmented periods. The
features were transformed into a vector representing the
data in feature space. A selected subset was thereby ap-
plied for the classification regarding various criteria in-
cluding time, completeness and constancy (see also Table
3).
The preprocessing, data analysis and transforma-
tion stages were carried out using MATLAB® version
R2010b for Windows. The programming environment
includes the so-called Neural Network Toolbox™, pro-
viding functions for the design, realization, visualization
and simulation of various ANNs. The practical applica-
tion of the routines is described in the following sections.
Modeling and supervised classification (ANNs)
In the overall machine learning theory, the specific area of
ANNs is generally divided into the so-called supervised
and unsupervised learning methods depending on the
labeling of the data. While supervised techniques require
labeled input data, the goal of unsupervised procedures is
to find significant patterns from the given examples. A
typical ANN structure with input, hidden and output lay-
ers is illustrated in Figure 6.
For the current study, the use of supervised learn-
ing methods, mapping input objects to desired output
values, appeared to be a suitable modeling technique,
considering the inclusion of the measured time series and
the experts’ evaluations of the executions. These assess-
ments in respect to pre-defined indicators and specifica-
tions were carried out on the basis of video recordings
with the help of professional coaches. In particular, the
chosen evaluation process was based on the available
literature discussed earlier and common recommendations
stating that factors like time, velocity, constancy and
completeness are significant determinants for the execu-
tion and the quality of the movement.
The appraisements were furthermore used
for training and classification purposes by labeling the
Novatchkov and Baca
33
Figure 6. The typical design of an ANN with an input, hid-
den and output layer, respectively consisting of 3, 4 and 2
neurons.
extracted feature information in respect to the evaluated
exercises. Consequently, with regard to the presented
weight training approach, the application of supervised
procedures aimed at the classification of the executions
into rather good and bad categories in terms of the men-
tioned factors. In this way, particularly inexperienced
individuals could benefit from the realization of automatic
algorithms by optimizing their technique and hence their
training.
In the present research, the modeling of the meas-
ured signals included the design and realization of con-
ventional ANNs. More precisely, various multilayer pat-
tern recognition networks (special type of feedforward
networks) were set up for learning and classification pur-
poses. The first step included the assignment of the ex-
tracted information to chosen labels or classes, which
were thereupon applied together as training sets for the
development of data models. Thereby, the extracted fea-
tures were combined into input vectors characterizing
each repetition and thus defining the shape of the training
data and number of dimensions. The output, on the other
hand, consisted of the expert assessments, also represent-
ing the number of neurons and the respective layer size.
Traditionally, the mapping of the data was accomplished
by the insertion of hidden layers. The used training func-
tion was based on the Levenberg-Marquardt algorithm,
which is among the fastest techniques for feedforward
networks due to its high efficiency in minimizing a func-
tion by applying curve-fitting methods.
In order to enhance and evaluate the learning pro-
cess, the computed feature vectors were divided into 3
subsets. This fragmentation was particularly needed for
identifying a model fitting the seen set, representing the
actual training process. Afterwards the model was pruned
based on different techniques such as the estimation of
prediction error (also known as validation) and finally
evaluated by unseen data (often referred to as testing
stage). The division ratio was fixed to 70:15:15.
This split-up was furthermore important for im-
proving generalization, whereas the performances of the
created models were measured on the basis of the error
rate (where error rate is defined as the number of incor-
rectly classified instances). For these purposes, the so-
called early stopping method was applied, aborting the
learning process at the point of minimal validation set
error, where the networks usually generalize the best. In
this way, the performance was verified and controlled
after each iteration and an overfitting or overtraining
could be avoided. A practical example including the clas-
sification and performance outcome of the designed ANN
is presented in the following section.
Results
Data segmentation and feature extraction
The implemented algorithm was able to segment all per-
formed sets (more than 60), detecting all executed repeti-
tions (over 750). The outcome of the segmentation proce-
dure on the example of continuous executions is illus-
trated in Figure 7.
As shown, the determined peaks (visualized by as-
terisk markers) were used to detect two subsequent repeti-
tions including the durations of the reversals and the con-
centric and eccentric actions. In addition, due to further
data transformation needs and simplification purposes, the
data was not only divided into single repetitions but inter-
nally also into extension and flexion phases as well as
holding times in between both movements. This division
was accomplished by taking into consideration the meas-
ured time series and particularly the changes of the dis-
placement values. Moreover, when looking at the force
characteristics, similar patterns can be recognized, based
on which the application of supervised machine learning
techniques appears to be a suitable classification method.
In particular, various specifications including time, force,
velocity, consistency, range and completeness factors
were defined for feature extraction and selection purposes
(see also Table 3).
Modeling
Figure 8 shows the classification results of the applied
pattern recognition network in respect to the overall exe-
cution based on the specified constancy, time and com-
pleteness criteria, considering duration, force and velocity
dependent features. The layer sizes of the networks were
fixed as follows: 7 (input), 10 (hidden) and 3 (output).
The selected labels are illustrated in Table 4.
Table 4. Definition of labels for classification purposes re-
garding the overall stability of the executions.
Label/class Definition
1
2
3
Stable execution
Instable eccentric stage
Instable concentric stage
Sub-chart (a) reflects the classification outcome in
form of a confusion matrix, indicating the correlations of
the labels. Apparently, the agreement is quite high for all
3 labels, demonstrating the high potential of the applied
pattern recognition method. Sub-diagram (b), on the other
hand, highlights the performance development of the
ANN throughout the training stage, illustrating also the
Artificial intelligence in sports
34
Figure 7. Illustration of the segmentation process for an improper (a) and proper (b) execution applied on the
same data as in Figure 4. The displacement data was segmented into single repetitions based on peak detection
techniques (marked by asterisk). The red and green lines indicate the starting and ending points in time of the
holding phases after the concentric and eccentric actions.
application of the used stopping methodology. Thereby,
the best performance was reached at epoch 70 (with a
maximum failing rate fixed to 5). As the displayed train-
ing, validation and test curves represent quite similar
characteristics with likewise and constant slopes, the
performance outcome appears to be an adequate founda-
tion for the developed techniques and further applications.
Discussion
Nowadays, due to the progress of information and com-
munication technologies including simplified and conven-
ient implementations of wireless sensor networks for data
acquisition and mobile devices for processing purposes,
the integration of intelligent methods becomes increas-
ingly important for the automatic analysis of measured
parameters and the realization of prompt intervention
routines.
AI concepts appear to be particularly suitable for
the design of effective evaluation and feedback frame-
works in sport. After the initial boom in the 1970s and
1980s, the use of AI techniques is meanwhile limited to
rather specific application fields including also sport, as
their application gets essential for the assessment of sports
data. Recent examples include the development of mobile
monitoring systems integrating classification algorithms
for the real-time analysis and feedback generation in
sports like, for instance, running (Kugler et al., 2011) or
golf (Eskofier et al., 2011). Similarly, the current study
investigated the application of AI methods in combination
with novel measuring instruments in the field of weight
training.
Today, due to the advances in measuring tech-
nologies, effective hardware implementations exist that
enable the integration of modern sensors into the fitness
equipment itself. For example, it is possible to attach load
cells or rotary encoders directly to weight training ma-
chines, allowing the measurement of relevant force and
displacement characteristics. The gathered data can there-
by be used for the implementation of sophisticated rou-
tines by means of machine learning techniques, automati-
cally analyzing the exercises. Particularly supervised
ANNs appear to be promising classifiers, as they offer
effective techniques for mapping input data into already
labeled output information.
In light of the current popularity of fitness studios
and the broader usage of weight training machines, also
the accuracy and correctness of the performances and
executions on the offered equipment has become crucial
(particularly for inexperienced or elderly individuals).
Practically, the quality of the movement plays a signifi-
cant role, as it contributes to the efficiency and value of
the workout. Therefore, a particular focus of the illus-
trated approach was to provide automatic analysis on the
technique as well as appropriate interventions and sugges-
tions. The development and integration of such models
Novatchkov and Baca
35
Figure 8. Classification results illustrating the outcome of the confusion matrix (a) and performance
curves (b) for the trained ANN in respect to characteristics regarding the constancy of the executions.
and routines thus might enable new facilities for the sup-
port of sportsmen and injury prevention.
The target user group includes in the first instance
inexperienced and elderly individuals, who can benefit
from the modern sensor equipment and measurements in
combination with the realization of the developed assess-
ment routines. Their systematic integration in intelligent
weight training machines would allow an automatic anal-
ysis of the quality of the execution on the basis of the pre-
defined criteria, thereby providing appropriate feedback
during or just after the performed exercise.
The most suitable field of application of such de-
velopments including the intended design of automated
resistance training equipment can be seen in the increas-
ing amount of available regular but also exclusive fitness
clubs. In these facilities the personal and individual care,
assistance and mentoring of the members play major
factors. Sportsmen can improve their technique on the
Artificial intelligence in sports
36
basis of the automated evaluation routines and the return
of instant notifications regarding occurred mistakes, by
receiving appropriate corrective advices and improvement
suggestions. Based on this feedback information also the
risk of injuries can be reduced, which is another big future
objective of the approach.
At the same time, the application of the developed
approach would bring valuable advantages to professional
sportspeople. In this context, the aim of monitoring veloc-
ity characteristics and action forces of exercising move-
ments might be to determine the contraction force speci-
ficity. Well-trained athletes need to optimize their training
to their sport-specific requirements and to improve their
functional ability more specifically. In order to achieve
the desired adaptations it is therefore advisable to choose
exercises that specifically meet the force-velocity needs
of the sport. Hence, the presented approach would allow
professionals but also their coaches to analyze in detail
the athletes’ executions and improve their performances
by looking in real time at the measured force and dis-
placement time series or also calculated acceleration,
velocity and power properties. Consequently, the possibil-
ity of immediate control and comparison of the results
could lead to a considerable training enhancement for
elite sportsmen.
Limitations
The restrictions of the presented approach are that, for the
moment, the implemented data analysis and feedback
methods are narrowed down to the mentioned criteria
such as time, constancy, velocity and completeness. In
particular, it is not possible to observe other pre-
conditions like, for example, the correct sitting position or
placement of the feet, which would be important parame-
ters for the application of the implemented models in
conjunction with the sensor-equipped leg press machine.
Furthermore, the developed routines are not able to di-
rectly monitor the posture of the body throughout the
movement including, for instance, the knee or upper body
and particularly the lower back motions. Such factors,
however, might be for example detected and assessed on
the basis of the integration of other measuring devices
such as goniometers, torsiometers or pressure sensors.
Outlook
The future work of the presented research will concentrate
on the adaptation of the developed models for different
sensor-equipped machines including lat biceps curl, lat
pull-down and shoulder press machine. At the same time,
another particular aim will focus on the development of
further models and solutions for the computer-based as-
sessment of weight training data. One promising approach
would, for instance, involve the implementation of fuzzy
logic concepts, which are also commonly applied in the
area of AI. As part of the probabilistic logic, such meth-
ods might be suitable for the realization of other effective
possibilities for the automatic evaluation of weight train-
ing exercises.
It is intended to include the implemented proce-
dures into the already mentioned mobile coaching system
by integrating the ANT technology as well as computer
devices into the measurement process. Thus, a bidirec-
tional data transfer between the weight training machines
and the implemented evaluation routines on the server
could be accomplished. This system integration would
contribute significantly to the idea of instant data acquisi-
tion, automated analysis and prompt feedback routines.
Conclusion
Computer-based feedback frameworks involving sophisti-
cated assessment techniques become increasingly essen-
tial for the instant analysis and appropriate intervention
during workouts. The present study suggests a novel
evaluation approach integrating AI methods for the ma-
chine-aided appraisement of weight training exercises.
The implementation involved the use of modern sensor
technologies attached to the training equipment, allowing
an effective acquisition and collection of sport-specific
data. The gathered parameter values were applied for the
automatic analysis of the performed exercises. The mod-
eling of the data was based on supervised learning proce-
dures integrating ANNs. The pre-processed sensor input
was used for the classification and autonomous appraisal
of the executions. The developed techniques showed good
results and performance outcome, raising promise for
their practical application in integrated feedback systems.
Further research, optimizations and hardware realizations
would then allow the intended implementation of a sup-
portive coaching system that provides professional and
hobby athletes as well as coaches with prompt assessment
and feedback tools.
References
Acikkar, M., Akay, M. F., Ozgunen, K. T., Aydin, K. and Kurdak, S. S.
(2009) Support vector machines for aerobic fitness prediction of
athletes. Expert Systems with Applications 36, 3596-3602.
Ariel, G. (1984) Resistive exercise machines. In: Biomechanics. Eds:
Terauds J. et al. Eugene, Oregon, 21-26.
Baca A., Dabnichki P., Heller M. and Kornfeind P. (2009) Ubiquitous
computing in sports: A review and analysis. Journal of Sports
Sciences 27(12), 1335-1346.
Baca, A., Kornfeind, P., Preuschl, E., Bichler, S., Tampier, M. and
Novatchkov, H. (2010) A server-based Mobile Coaching sys-
tem. Sensors 10, 10640-10662.
Baca, A. (2012) Methods for Recognition and Classification of Human
Motion Patterns – A Prerequisite for Intelligent Devices Assist-
ing in Sports Activities. MATHMOD 2012, February 15-17,
Vienna, Austria. Abstract Volume 14.
Baca, A. and Kornfeind, P. (2012) Stability analysis of motion pattern in
biathlon shooting. Human Movement Science 31(2), 295-302.
Bartlett, R. (2006) Artificial intelligence in sports biomechanics: New
dawn or false hope? Journal of Sports Science and Medicine 5,
474-479.
Bartlett, R. and Lamb, P. (2011) The use of artificial neural networks in
studying movement coordination. In: Proceedings of 8th Inter-
national Symposium of the International Association of Com-
puter Science in Sports. September 21-24, Shanghai, China. 22-
33.
Campbell, M. (2002) Deep Blue. Artificial Intelligence 134, 57-59.
Chang, K.-H., Chen, M. Y. and Canny, J. (2007) Tracking free-weight
exercises. In: Proceedings of UbiComp 2007: 9th International
Conference on Ubiquitous Computing. September 16-19, Inns-
bruck, Austria. 19-37.
Curtis, K. M. (2010) Cricket batting technique analyser/trainer: a pro-
posed solution using fuzzy set theory to assist West Indies
cricket. In: Proceedings of the 9th WSEAS international confer-
Novatchkov and Baca
37
ence on Artificial intelligence, knowledge engineering and data
base. February 20-22, Cambridge, UK. 71-76.
Evans W. (1999) Exercise training guidelines for the elderly. Medicine
and Science in Sports and Exercise 31(1), 12-17.
Eskofier, B., Wagner, M., Munson, I. and Oleson, M. (2010) Embedded
classification of speed and inclination during running. Interna-
tional Journal of Computer Science in Sport 9(1), 4-19.
Eskofier, B., Tuexen, S., Kugler, P., Jensen, U. and Wright, I. (2011)
Development of Pattern Recognition Methods for Golf Swing
Motion Analysis. In: Proceedings of 8th International Sympo-
sium of the International Association of Computer Science in
Sports. September 21-24, Shanghai, China. 71-75.
Fischer, A., Do, M., Stein, T., Asfour, T., Dillmann, R. and Schwa-
meder, H. (2011) Recognition of individual kinematic patterns
during walking and running – A comparison of artificial neural
networks and support vector machines. International Journal of
Computer Science in Sport 10(1), 63-67.
Garber, C. E., Blissmer, B., Deschenes, M. R., Franklin, B. A., Lamonte,
M. J., Lee, I. M., Nieman, D. C. and Swain, D. P. (2011) Amer-
ican College of Sports Medicine position stand. Quantity and
quality of exercise for developing and maintaining cardiorespi-
ratory, musculoskeletal, and neuromotor fitness in apparently
healthy adults: guidance for prescribing exercise. Medicine and
Science in Sports and Exercise 4(7), 1334-1359.
Ghasemzadeh, H. and Jafari, R. (2011) Coordination analysis of human
movements with body sensor networks: A signal processing
model to evaluate baseball swings. IEEE Sensors Journal 11(3),
603-610.
Ghasemzadeh, H., Loseu, V. and Jafari, R. (2009) Wearable coach for
sport training: A quantitative model to evaluate wrist-rotation in
golf. Journal of Ambient Intelligence and Smart Environments
1, 1-12.
Kotsiantis, S., Kanellopoulos, D. and Pintelas, P. (2006) Data Preproc-
essing for Supervised Leaning. International Journal of Com-
puter Science 1(2), 111-117.
Kugler, P., Schuldhaus, D., Jensen, U. and Eskofier, B. (2011) Mobile
Recording System for Sport Applications. In: Proceedings of
8th International Symposium of the International Association of
Computer Science in Sports. September 21-24, Shanghai, Chi-
na. 67-70.
Lamb, P., Bartlett, R. and Robins A. (2010) Self-Organising Maps: An
Objective Method for Clustering Complex Human Movement.
International Journal of Computer Science in Sport 9(1), 20-29.
Lames, M., Perl, J., Schroder, H.J. and Uthmann T. (1990) Tennis
Expert System – TESSY. Leistungssport 4, 49-54. (In German).
Lapham, A. and Bartlett, R. (1995) The use of artificial intelligence in
the analysis of sports performance: A review of applications in
human gait analysis and future directions for sports biomechan-
ics. Journal of Sports Sciences 13, 229-237.
McCullagh, J. (2010) Data Mining in Sport: A Neural Network Ap-
proach. International Journal of Sports Science and Engineer-
ing 4(3), 131-138.
Newborn, M. (1997) Kasparov versus Deep Blue: Computer Chess
Comes of Age. Springer.
Novatchkov, H. and Baca, A. (2012) Machine learning methods for the
automatic evaluation of exercises on sensor-equipped weight
training machines. In: Procedia Engineering 9, Engineering of
Sport Conference 2012 34, 562-567.
Papić, V., Rogulj, N. and Pleština V. (2009) Identification of sport
talents using a web-oriented expert system with a fuzzy module,
Expert Systems with Applications 36(5), 8830-8838.
Poole, D., Mackworth, A. and Goebel, R. (1998) Computational Intelli-
gence: A Logical Approach. New York: Oxford University
Press.
Perl, J. (2004a) Artificial neural networks in motor control research.
Clinical Biomechanics 19(9), 873-875.
Perl, J. (2004b) A neural network approach to movement pattern analy-
sis. Human Movement Science 23, 605–620.
Rana, S., Chleboun, G., Gilders, R., Hagerman, F., Herman, J., Hikida,
R., Kushnick, M., Staron, R., and Toma, K. (2008) Comparison
of early phase adaptations for traditional strength and endur-
ance, and low velocity resistance training programs in college-
aged women. Journal of Strength and Conditioning Research
22(1), 119-127.
Ratiu, O.G., Badau, D., Claudia Georgeta Carstea, C.G., Badau, A. and
Paraschiv, F. (2010) Artificial intelligence (AI) in sports. In:
Proceedings of AIKED'10: 9th WSEAS International Confer-
ence on Artificial Intelligence, Knowledge Engineering and Da-
ta Bases. February 20-22, Cambridge, UK. 93-97.
Silva, A. J., Costa, A. M., Oliveira, P. M., Reis, V. M., Saavedra, J.,
Perl, J., Rouboa, A. and Marinho, D. A. (2007). The use of neu-
ral network technology to model swimming performance. Jour-
nal of Sports Science and Medicine 6, 117-125.
Westcott, W. (2009) ACSM strength training guidelines: Role in body
composition and health enhancement. ACSM's Health & Fitness
Journal 13, 14-22.
Winett, R. A. and Carpinelli, R. N. (2001) Potential Health-Related
Benefits of Resistance Training. Preventive Medicine 33, 503-
513.
Key points
• Artificial intelligence is a promising field for sport-
related analysis.
• Implementations integrating pattern recognition
techniques enable the automatic evaluation of data
measurements.
• Artificial neural networks applied for the analysis of
weight training data show good performance and
high classification rates.
AUTHORS BIOGRAPHY
Hristo NOVATCHKOV
Employment
Research Assistant, University of Vienna,
Centre for Sport Science and University
Sports, Vienna, Austria
Degree
Dipl.-Ing.
Research interests
Machine learning, mobile and pervasive
computing in sports
E-mail: hristo.novatchkov@univie.ac.at
Arnold BACA
Employment
Full Professor and Head of Centre, University of
Vienna, Centre for Sport Science and University
Sports, Vienna, Austria
Degree
Dipl.-Ing. Dr.
Research interests
Motion analysis, feedback systems, mobile coach-
ing, pervasive computing in sports
E-mail: arnold.baca@univie.ac.at
Hristo Novatchkov
University of Vienna, Centre for Sport Science and Uni-
versity Sports, Auf der Schmelz 6A, 1150 Vienna, Austria
