Submitting Institution
Sheffield Hallam UniversityUnit of Assessment
Sport and Exercise Sciences, Leisure and TourismSummary Impact Type
TechnologicalResearch Subject Area(s)
Mathematical Sciences: Numerical and Computational Mathematics
Information and Computing Sciences: Artificial Intelligence and Image Processing
Medical and Health Sciences: Neurosciences
Summary of the impact
The Centre for Sports Engineering Research (CESR) has developed new
methods to allow accurate 3D measurement. These methods have evolved from
lab-based to in-situ systems allowing real world measurements
using multi-camera systems, object tracking, signal processing and planar
calibration. This research has had four main types of impact:
- the implementation of camera-based analysis systems embedded within
Olympic teams in preparation for London 2012 and Rio 2016;
- the implementation of systems at the International Tennis Federation
(ITF) to allow that organisation to monitor the game of tennis and set
its rules;
- the transfer of knowledge and systems to the commercial and health
sectors;
- dissemination to the research community.
Underpinning research
Between 2008 and 2010, Allen, Choppin, Goodwill and Haake (references 1,
2) carried out industrially sponsored research (grant a) to develop an
understanding of dynamic tennis shots using finite element or Newtonian
models. These models were validated experimentally using a bespoke impact
rig with 3D photogrammetry to measure the complex motion of the racket and
ball. The finite element model in reference 1 was the first to simulate
accurately oblique-spinning off- centre impacts with a racket.
Allen used his model to look at the effect of tennis racket parameters on
simulated ground-strokes showing that mass distribution rather than racket
stiffness was the most important parameter. The work also highlighted the
need to measure what happens during real play on a tennis court and
Choppin et al. (2010; reference 2) used two high speed cameras and
the planar calibration method to determine racket and ball motion at
Wimbledon qualifiers. The work required the development of in-situ
techniques with minimal interaction with the player to produce detail 3D
information on the motion and dynamics of the racket and ball during play.
In parallel to this between 2006 and 2010, research was sponsored by the
ITF (grant b) to create a method of automatically measuring ball spin
during play. The work was used to develop laboratory systems at the ITF to
determine the overall physics of spin generation in tennis (references 4,
5) using the knowledge of real play conditions. This work determined the
effect of impact spin, speed and angle on rackets with 6 different string
types with a wide range of friction coefficients. Multiple regression
analysis showed that rebound was extremely complex and depended upon
whether strings slipped throughout impact. Thus, friction between strings
was much more important than friction between the strings and the ball.
The work described here marks a transition from lab-based studies to
measurement in real environments. References 1, 2 and 3, between 2010 and
2013, showed that capture of useful 2D and 3D information in the field
relies on a robust and practical camera calibration system, for which the
planar calibration system was developed. The research demonstrated that
planar calibration was superior to Direct Linear Transformation (DLT) in
many real-life scenarios because reconstruction errors were smaller, the
calibration object was simpler to use and errors scaled with the
environment.
The work has moved into the research of low-cost off-the-shelf systems
using depth cameras (such as the Kinect). In 2013, Choppin and Wheat
(reference 6) explored the potential of these systems to measure body
segment parameters, body segment angles and player tracking over a large
volume. They showed that the centre mass position over a principal moments
of inertia could be measured with root mean square error to ±3% using a
Kinect (compared to a 12-camera motion capture system). While the overall
accuracy was not good enough for all biomechanical analysis, it was shown
that systems are capable of useful field measurements which will become
more accurate as hardware improves.
The status of the researchers at Sheffield Hallam University is:
- Allen T, Senior Sports Engineer, 1/6/09-1/9/11; Senior Lecturer 1/9/11
onwards;
- Choppin, S, Sports Engineer, 1/6/09 onwards;
- Goodwill, SR, Senior sports engineer, 1/3/06 onwards;
- Haake SJ, Professor of Sports Engineering, 1/3/06 onwards;
- Kelley J, Sports Engineer 1/8/10 onwards
- Wheat J, Senior Research Fellow 1/8/10-31/8/12, Principal Research
Fellow 1/9/12 onwards.
References to the research
All outputs were published in peer-reviewed journals. References 5 and 6
are the only ones not submitted to the REF; however, both are published in
international journals.
Allen T, Haake SJ & Goodwill SR (2010) "Effect of tennis
racket parameters on a simulated groundstroke", Journal of Sports
Sciences, 29(3), 311-325. Scopus SJR 0.909. (Allen REF submission
1);
Choppin, S.B., Goodwill, S.R. & Haake, S.J. (2010)
"Investigations into the effect of grip tightness on off-centre forehand
strikes in tennis", Proc. IMechE Part P: Journal of Sports Engineering
and Technology, 224, pp. 249-257. Scopus SJR 0.274 - 0.616. (Choppin
REF submission 1);
Choppin, S.B., Goodwill, S.R. & Haake, S.J. (2011) "Impact
characteristics of the ball and racket during play at the Wimbledon
qualifying tournament", Sports Engineering,13 (4), pp. 163-170.
Scopus SJR 0.283. Cited 7 times. (Goodwill REF submission 4);
Haake, S.J., Allen, T., Jones, A., Spurr, J. & Goodwill,
S.R. (2012) "Effect of inter-string friction on tennis ball
rebound", Proc. IMechE, Part J: 226 (7), pp. 626-635. Scopus SJR
0.274 - 0.616. (Goodwill REF submission 2);
Allen, T., Ibbotson, J. and Haake, S.J. (2012) "Spin
generation during an oblique impact of a compliant ball on a non-compliant
surface", Proc. IMechE, Part P: 226 (2), pp. 86-95. Scopus SJR
0.274 - 0.616;
Choppin, S. and Wheat, J.S. (2013). The potential of the
Microsoft Kinect in sports analysis and biomechanics. Sports
Technology, 6, 78-85.
Grants associated with the impact
a. SJ Haake: Prince Racquets, £30,000, 1st Jun 2006 to 31st
May 2009, PhD sponsorship of Tom Allen "Finite element model of a tennis
ball impact with a racket";
b. SJ Haake: International Tennis Federation, £30,000, 31st
Oct 2006 to 30th September 2010, PhD sponsorship of John Kelley
"Measuring ball spin rates in match play tennis";
c. SJ Haake: UK Sport, £209,181, 1st Apr 2008 - 31st
Mar 2012, UK Sport Innovation Partnership;
d. J Kelley: UK Sport/English Institute of Sport (EIS), £60,808, 1st
Aug 2011 to 31st July 2012, Elite Sport Training Tools;
e. SJ Haake: UK Sport/EIS Research and Innovation Team, £260,000, 1st
Apr 2012 to 31st Mar 2014, Innovation Partnership;
f. J Kelley: UK Sport, £25,000, 1st Dec 2011 to 31st
December 2012, UK Sport Ideas for Innovation Competition - iSwim automated
tracking of swimmers;
g. J Wheat: GB Cycling/EIS, £6,780, 1/3/13-1/7/13. Monitoring the
morphology of athletes using a depth camera based scanning system;
h. J Wheat, The Royal Derby Hospital, £10,000, 1/1/2012 to 31/12/2012,
Development of a portable measurement system for breast clinics.
Details of the impact
Supporting UK Sport and Team GB
In 2008 UK Sport made CSER a UK Sport Innovation Partner to implement
performance analysis technologies using its research expertise (grant c)
in the lead up to London 2012. CESR used the camera and photogrammetric
expertise developed in tennis to create calibrated measurement systems for
use in day-to-day training and coaching. These systems used the planar
calibration technique (references 2, 3) to allow measurement in large
volumes and tracking algorithms (reference 5) developed in tennis to track
athletes. Systems generally had the following characteristics: coaches
found them simple to use; they employed automated video analysis; they
provided storage and retrieval; and the analysis they generated gave a
performance advantage.
Examples are as follows:
- GB Boxing required CSER to instrument the new GB Boxing Gym at the EIS
in Sheffield with a 5-camera system called iBoxer. The system
allowed boxing analytics to be developed over and above tracking across
the ring. This included punch rates during a round and the strategic
analysis of opponents. The GB Boxing performance analyst said "the
data suggests that this kind of preparation increases the boxer's
success rate by around 25%" (http://www.bbc.co.uk/news/technology-18735629).
GB Boxing won 5 medals at London 2012.
- GB Taekwondo installed iTaekwondo at its gym in Manchester and was a
derivative of iBoxer. The EIS performance analyst said of the system, "It
has been a pleasure collaborating with you on the project to produce a
very innovative, and truly world leading product that I can
confidently say puts GB Taekwondo and myself at the cutting edge of
the sport in the run up to 2012" (by email, 11th Feb
2010). Taekwondo won 3 medals at London 2012.
- British Gymnastics required a camera and review system for its gym at
Lilleshall. CSER developed a system with calibrated cameras and simple
player tracking to allow coaches to review technique and actions of its
gymnasts. The EIS performance analyst said, "iGym has enabled video
feedback during training for all squad gymnasts providing them with a
tool to aid skill development and polishing of routine work. The high
speed automatic capture system on two vaults is the most impressive
part of the system which serves as such a powerful tool replaying the
fast movement to both gymnast and coach...A very successful project."
Gymnastics won 3 medals at London 2012.
Over 30 systems have been developed for UK Sport (source 1) and this
success has led to further UK Sport/EIS funding (grants d, e) and a UK
Sport Innovation Prize (grant f). Following London 2012, Dr Goodwill
received a personal letter from the CEO of UK Sport (1/11/2012: source 2)
which said, "Without a doubt, the projects delivered have contributed
to the medal success we have seen in London, Vancouver, Beijing, Turin
and as far back as Athens in 2004. In the current funding review process
as part of the Rio vision, many of the sports have highlighted the
support and the benefits of the numerous projects in partnership with
Sheffield Hallam University and the team of scientists and engineers. It
is no coincidence that your name is continually linked to these
projects."
In March 2013, Wheat was commissioned by EIS Manchester to design and
build a bespoke apparatus using low-cost depth cameras to measure the body
shape of Olympic athletes (grant g). The first prototype was tested in
July at EIS Manchester and is being used to track the development of elite
athletes. Additionally, CSER was awarded the following peer-reviewed
prizes for this work: (1) Sports Analytics Innovation Awards, Sports
Analytics Research Institution of the Year Award (March 2013); (2) PODIUM
Awards, RCUK Award for Exceptional Research Contribution. Silver Medal
(May 2012).
Support for the International Tennis Federation
The ITF has implemented much of CSER's research within its systems for
monitoring the nature of the sport. The work on the measurement of spin
(references 4, 5) has been implemented in the software SpinDoctor
which has been used by the ITF to monitor the game: it was used at the
Davis Cup (2010, 2011, 2012), the ATP Masters in London (2012, 2013) and
at Wimbledon (2011, 2012, 2013) with a total of 3,437 shots analysed. The
spin rig used in references 4 and 5 has been coupled to a database to
allow the ITF to assess the effect of new strings and stringing methods on
spin generation by specific rackets. The results of this work were
presented to the ITF Technical Commission (source 3) from 2010 (on which
Prof Haake is a voting member) whose role it is to monitor the game and
assess the appropriateness of the rules of tennis to technology. The use
of SpinDoctor and the spin rig showed that spin generated by
professional players is significantly less than that of the `spaghetti
racket' which was banned in 1980; the ITF will continually monitor the
amount of spin in the game, using Rule 4 (The Racket) to temper excessive
spin. Prof Haake's expertise in performance analysis systems and
experience from UK Sport were used to inform the Technical Commission and
the introduction of a new rule to tennis (Rule 31) at its AGM on 12th
July 2013. This rule allows the use of performance analysis technology in
tournament play for the first time and will come into force on 1st
January 2014.
Supporting commercial and health organisations
The understanding of dynamics developed in references 1 and 2 and funded
by Prince (grant a) was used by Prince to improve its understanding of
tennis racket performance from 2006 onwards and supported its growth to
number one racket manufacturer by 2010. The improvements in measurement
techniques described in references 2, 3, and 5 were developed by CSER
staff into tools that could be used by our commercial partners.
SpinDoctor was sold to Prince, IsoSport and adidas to use in their
labs (with the ITF receiving 25% of royalties) to allow the measurement of
scalar spin in the field. adidas requested a further system for its
football testing with the ability to give 3D spin vectors and used the
camera calibration and tracking systems outlined in references 3 and 5.
Since 2011, footballs have been tested using the system, enhancing adidas'
ability to develop its products (source 4). Tracking and signal processing
techniques (reference 3) were also used to develop an iPhone application
for adidas (Snapshot) to automatically measure the speed, angle and
trajectory of a kick. The application had been downloaded around 300,000
times (31/7/2013) and is the first analysis App developed by adidas.
CSER is also working with Labosport to test the Goal-Line Technology
systems being installed at Premier League clubs around the UK. The work
uses techniques from references 2, 4 and 5 to measure simulated goals
according to FIFA protocols. CSER staff have been accredited by FIFA to
carry out the work and all 2012-13 Premier League clubs have been assessed
and passed by CSER according to the FIFA specifications.
The depth-camera work has also been transferred into the commercial and
health sector. Wheat (grant h) has been commissioned to build a low cost
scanning device for Oxylane (Decathlon) for installation in its stores and
also for breast clinics in collaboration with a Senior Breast Surgeon at
the Royal Derby Hospital (grant h; source 5). The significance of the
latter work allowed it to win 3rd prize in the 2012 Health
Enterprise East Awards, Medical Technology & Software category. Heller
used depth cameras to allow older people to enhance their physical
capabilities by interacting with virtual worlds ("2nd Lives for
the 3rd Age). This work has been used as a case study to inform
the research community on the future of research through a Parliamentary
report on `Assistive Technology 2010-11' and a RCUK report (6/2011) on `UK
Research that will have a Profound Effect on our Future'.
Dissemination of research findings to the research community
The work of Goodwill, Choppin, Haake and Kelley showed that planar
calibration was superior to the more commonly used DLT for in situ
environments. A full suite of applications was made available as open
source software (http://www.check3d.co.uk/default.html)
to allow the research community to benefit from the enhanced accuracy of
the planar method. Since the site went live in Sept 2012, the research has
had more than 600 views and by 31/7/2013 the software had been downloaded
over 200 times by those seeking to benefit from the enhanced accuracy it
brings.
A further website (http://www.depthbiomechanics.co.uk/)
made the 3D software available to the research community to allow others
to benefit from the research. From 8/11/2012 to 31/7/2013, the software
has been downloaded 46 times by researchers in North America (40%), Europe
(44%) and Asia and Australasia (17%); the majority are university
researchers but also benefiting are companies such as Casio. The
researchers come from a range of disciplines including software
engineering (53%), biomechanics (30%), physiotherapy (10%), and psychology
(7%).
Sources to corroborate the impact
- Senior Manager, Performance Solutions, English Institute of Sport;
- Chief Executive Officer, UK Sport;
- Executive Director, Science and Technical, The International Tennis
Federation;
- Senior Manager, Football Development, adidas Innovation Team;
- Senior Breast Surgeon, The Royal Derby Hospital.