14 Faculty of Sport, University of Ljubljana, ISSN 1318-2269  Kinesiologia Slovenica, 16, 3, 14–22 (2010)
IZVLEČEK
Cilj te študije je bilo poiskati način računanja časa in 
hitrosti iz meritev Globalnega Navigacijskega Satelitskega 
Sistema (GNSS) v sprintu za vnaprej določene položaje 
vzdolž atletske steze in pokazati, da lahko časi odsekov 
in hitrosti priskrbijo dodatne pomembne podatke za 
analizo uspešnosti na 100 m sprintu. Poskusi so bili 
izvedeni na dveh moških športnikih. Fotocelice so 
bile nameščene ob 100 m atletski stezi za neposredno 
merjenje časa. S predlaganim GNSS sistemom pa je 
bilo izračunanih 40 časov in iz tega 39 časov odsekov 
od katerih vsaka zajema 2,5 m. Glavne ugotovitve v tej 
študiji so: 1) predstavljen je bil novi način merjenja časa 
odsekov na 100 metrski atletski stezi z uporabo GNSS 
merilnega sistema, ki zagotavlja razmeroma majhne 
razlike v izmerjenih časih v primerjavi s fotocelicami; 
2) nova metoda merjenja časa omogoča neposredno 
primerjavo med tekači v krajevni odvisnosti, in 3) 
dodatno izmerjeni časi in hitrosti v sprintu s pomočjo 
nove metode merjenja priskrbijo sicer “skrite” razlike 
med sprinterji, ki lahko pomagajo strokovnjakom ali 
znanstvenikom podrobneje analizirati uspešnost. 
Metoda za izračun časa iz trajektorij visoko-natančnega 
GNSS sprejemnika zagotavlja zanesljive in točne 
rezultate meritev časa v sprintu. Z uporabnega stališča 
je nova metoda dobra alternativa za ustaljene meritve 
s fotocelicami, zlasti za zelo podrobne ali znanstvene 
analize.
Ključne besede: Atletika, GNSS RTK, GPS, čas, hitrost
ABSTR ACT
The aims of this study were to provide a method of 
computation time and velocity from Global Navigation 
Satellite System (GNSS) measurements in sprint 
running for predetermined positions along a sprint 
track and to show that section times and velocities can 
provide additional significant data in order to analyse 
performances in 100 m sprint running in greater detail. 
Experiments were performed on two male athletes. 
Photocells were set along a 100 m sprint track. Using the 
proposed GNSS system, 40 times were calculated for each 
runner, and afterwards 39 section times were derived, 
each covering 2.5 m. The main findings in this study 
are: 1) a new method of measuring time segments in a 
100 m sprint with the use of GNSS has been developed, 
providing relatively small differences of measured time 
in comparison to the photocell method; 2) a new method 
of measuring time enabled precise direct location 
comparisons between the runners; and 3) additional 
times and velocities in sprint running using the new 
method provide “hidden” data along the run which 
can help experts or scientists analyse performance. 
The method of computation time from the high-end 
GNSS surveyed trajectory provides reliable and accurate 
results for measurements in sprint running. From the 
usage point of view, it represents a good alternative to 
the established photocell measurement, especially for 
highly detailed or scientific analysis. 
Keywords: track and field, GNSS RTK, GPS, time, ve-
locity
University of Ljubljana, Faculty of Sport, Ljubljana, 
Slovenia
*Corresponding author:
Matej Supej
University of Ljubljana, Faculty of Sport,
Ljubljana, Slovenia
phone: +386 41395331
e-mail: matej.supej@fsp.uni-lj.si
THE USE OF A HIGH-END GLOBAL 
NA VIGATION SATELLITE SYSTEM IN A 100 M 
SPRINT
UPORABA VISOKO-LOČLJIVEGA 
GLOBALNEGA NA VIGACIJSKEGA 
SATLITESKEGA SISTEMA V SPRINTU NA 100 M
Matej Supej* 
Mitja Bračič 
Milan Čoh

GNSS in sprint running 15 Kinesiologia Slovenica, 16, 3, 14–22 (2010)
INTRODUCTION 
Results in sprints depend on the optimal integration of four phases: start, starting acceleration, 
maximum velocity and deceleration (Mero, Komi, & Gregor, 1992). Sprint running is a complex 
cyclical movement and is a product of stride frequency and stride length. Both parameters are 
mutually dependant and individually conditioned by the processes of the central regulation 
of movement, biomotor abilities, energy processes and morphological characteristics. Various 
phases influence the final result in a sprint race with different levels of importance. The study by 
Tellez and Doolittle (1984) revealed that the starting action contributes 6%, starting acceleration 
64%, maximum velocity 18% and deceleration 12% to the final time in a 100 m sprint race. 
Sprinting is, by definition, a cyclical movement with maximum possible velocity. However, 
the velocity changes significantly in individual segments. These changes are the result of the 
biomechanics of sprinting stride, the activation of various biomotor abilities, processes of neu-
romuscular control and functional factors (Coppenolle & Delecluse, 1989; Donatti, 1995; Muller 
& Hommel, 1997). The subject of this study was an examination of the control of dynamics in 
sprinting velocity in individual segments of a sprinting track in a 100 m race with the use of 
GPS technology. Previous research studies by several authors used a model of measuring time 
over 10 m segments, most frequently with the use of photocells. Such time-related information 
enabled researchers to precisely monitor the sprinting velocity. Nevertheless, the problem of such 
a method is its lack of precision when runners pass the beam as they can activate the photocell 
with different parts of their body; this factor cannot be controlled. When passing the beam with 
the hand first, at least a 1/100 second difference occurs than when passing the beam with the 
torso first. As a result, partial segment velocities over 10 m intervals provide only a partial insight 
into the dynamics of sprinters’ velocity. One alternative is the laser method which has been used 
in track and field since 2001. The laser works on the principle of the Doppler effect and provides 
more detailed information than the system of photocells. It enables the linear monitoring of 
sprinters’ movement in space. The key problem of the measurement precision is the reflection 
of beams from different sprinters’ body parts, i.e., the beam can reflect differently from a head, 
leg, torso etc., which can lead to a relatively large inaccuracy of results.
The latest technology is the Global Navigation Satellite System (GNSS) which enables the registra-
tion of a sprinter’s velocity in various time and position points on a 100 m race track. Examining 
instantaneous velocity at precisely defined moments of a race provides important information 
about the running biomechanics, motor control processes, the energy processes of runners and 
their degree of preparation. Further, monitoring the points of velocity along the track allows an 
indirect comparison of velocity parameters between sprinters of various levels.
The Global Navigation Satellite System is a term referring to different satellite navigation systems 
that provide autonomous global geo-spatial positioning; among these, the United States Global 
Positioning System (GPS) has recently been used in many different human locomotion and 
sport measurements (Ai & He, 1999; Skaloud & Limpach, 2003; Karboviak, 2005; Edgecomb & 
Norton, 2006; Tropedet al., 2008). GPS has been proven to be reliable when measuring distance 
and velocity in several outdoor activities (Ai & He, 1999; Schutz & Herren, 2000; Terrieret 
al., 2000; Larsson, 2003; Witte & Wilson, 2004; Rodriguezet al., 2005; Witte & Wilson, 2005; 
Townshendet al., 2008). However, GNSS has never been used to accurately measure performance 
or intermediate time, even though the system’s internal computation of positions relies on very 
accurate atomic clock time measurements (Parkinson & Spilker, 1996).

16 GNSS in sprint running Kinesiologia Slovenica, 16, 3, 14–22 (2010) 
The aims of this study were: 1) to provide a method of computation time and velocity from GNSS 
measurements in sprint running for predetermined positions along the sprint track; and 2) to 
show that section times and velocities can provide additional important data in order to analyse 
performances in 100 m sprint running in greater detail.
METHODS
Subjects
Experiments were performed on two male athletes (age: 28.0 and 23.0 years, body height: 183.0 
and 186.0 cm, body weight: 78.2 and 74.0 kg), who provided written informed consent. All 
procedures were approved by the Ethical Committee of the Faculty of Sport in Ljubljana, Slovenia 
and the study was conducted according to the Declaration of Helsinki.
Overall Design of the Study
The study started with a mathematical overview of a procedure for computation time and velocity 
from real-time kinematics (RTK) GNSS-surveyed trajectories on predetermined positions along 
a sprint track. Thereafter, a 100 m sprinting experiment was performed to show the functionality 
and advantages of the RTK GNSS system and methodology in sprint running. 
Time and velocity computation
All GNSS systems comprise three-dimensional movement trajectory in relation to time. In this 
study, the RTK GNSS system measured the trajectory sprinters’ movement. As the research 
intended to examine times and velocities for particular segments of the race track, first the marks 
of these segments had to be set along the track; in this study, every 2.5 m from the start to the 
finish. Mathematically speaking, these points were used to define referential planes which needed 
to be placed vertically and at a right angle to the direction of a race track (see Figure 1). Thus the 
calculated trajectory of a sprinter in the 100 m race crossed these referential planes. When the 
intersections of trajectory with each of the referential planes were calculated, running times from 
the start of the trajectory to the calculated intersections were derived. The calculated differences 
between individual segments represented the times of the segments. The desired precision of 
the time measurement is higher than the sampling rate of the GNSS device. Therefore, a “cubic 
spline” interpolation and an adequate “2-way” Kalman filter providing the time and place of the 
intersections more precisely were used.
 
20 m
2.5 m
Survey antenna
Intersections
Photocell
every 20 m
Reference plane
every 2.5 m
Sprint running
Figure 1: Diagram of a sprint runner passing the reference planes and photocell beams. Note 
that reference planes as used in combination with the GNSS were constructed every 2.5 m for 
timing purposes and were 10 times denser for velocity purposes. 

GNSS in sprint running 17 Kinesiologia Slovenica, 16, 3, 14–22 (2010)
The velocity of movement versus time was down-sampled to every 0.25 m from the measured 
trajectory with an intention of acquiring data about changes in velocity versus distance in order 
to enable a direct comparison on the race track between the measured subjects. To achieve this, 
first the velocities for the entire sprint race in relation to time were calculated with the use of a 
linear approximation; afterwards, its absolute values were calculated. A method identical to the 
calculation of the time segments was used to acquire the times of the race trajectory crossings 
with referential planes for every 0.25 m of the track. The times were then entered into the absolute 
velocity, related to time; thus calculating absolutevelocities for every 0.25 m of the race. As 
previously stated, a “cubic spline” interpolation was used to identify precise values. It must be 
noted that these values represent absolute velocities at precisely defined points of the race track 
and are not mere average velocities of the interval, which would be an equivalent parameter of 
time segments. In practice, the velocity-versus-time according to velocity-versus-distance was 
slightly down-sampled.
Instruments
A high-end RTK GNSS system with 99.99% reliability (according to the manufacturer; http://
www.leica-geosystems.com/uk/en/lgs_18994.htm) was used to measure the athletes’ trajectories. 
A rover and a reference station were built from the same hardware components: a dual frequency 
L1/L2, geodetic, Leica GX1230GG GNSS RTK receiver, a Leica GLONASS/GPS AX1202 GGsurvey 
antenna and Leica Satelline 3AS radio modems (Leica Geosystems AG, Heerbrugg, Switzerland) 
for real-time corrections. The system works in RTK mode at a maximum 20 Hz sampling rate s 
= 10 mm + 0.5 ppm and s = 20 mm + 0.5 ppm horizontal and vertical accuracy, respectively.
For all measurements, the reference station stood on a fixed tripod and was less than 100 m away 
from all surveyed points to transmit real-time corrections. To capture the athletes’ trajectories, 
the rover’s receiver, modem and antenna were placed in a small backpack worn by the athlete, 
with the antenna at the level of the upper thoracic spine (T2-T4). For the survey start and finish 
positions of the sprint track as well as to set the photocell positions, the antenna was attached 
to a 2 m high carbon geodetic pole and the rover was set to a static measurement mode with s 
= 5 mm + 0.5 ppm and s = 10 mm + 0.5 ppm horizontal and vertical accuracy, respectively. A 
local geodetic co-ordinate system was used for all measurements in order to achieve maximum 
accuracy (Parkinson & Spilker, 1996). At least eight satellites were visible during the measure-
ments with an elevation cut-off angle set to 15°. For validation purposes, six sets of Microgate 
Polyfemo photocells and Racetime 2 chronometers (Microgate S.r.l., Bolzan, Italy) were used, 
with a resolution of 3.47∙10
-5
 and 1.25∙10
-4
 s, respectively.
Experiment: 100 m sprint running
In the sprint running experiment, six photocells were set along a 100 m sprint track at 2.5, 20, 40, 
60, 80 and 100 m. Using surveyed start and finishpositions, each 2.5 m point was interpolated 
using a linear approximation. The first calculated time was at 2.5 m. The two athletes performed 
one 100 m run each. Using the proposed GNSS system, 40 times were calculated for each runner, 
and afterwards 39 section times were derived, each covering 2.5 m. In addition, five times were 
measured with the photocells and, from these, five section times were calculated. For validation 
purposes, the differences in time measured with GNSS and photocells were evaluated for the last 
four sections. To continue, using GNSS data absolute velocities were defined every 0.25 m with 

18 GNSS in sprint running Kinesiologia Slovenica, 16, 3, 14–22 (2010) 
an emphasis on 2.5 m and compared to average velocities that were derived from 20-m section 
times measured from photocells. 
Statistics
Where appropriate, the data were presented as the mean and standard deviation (s). Scatter and 
line plots were used to present section times and absolute velocities. All calculations and plots 
were programmed in Matlab R2007a.
R ESULTS
The mean differences in time between the GNSS and the photocell measurements over the last 
four 20-m sections were 0.0021 s (s = 0.0051) at a mean sprint velocity of 7.46 m . s
-1
 (s = 0.32), 
derived from 20-m sections. The mean time calculated for the last four 20-m sections for the 
two runners was 2.69 s (s = 0.12).
The scatter and the overlaid line plot in Figure 2 show the section times for 100 m runs measured 
with the GNSS and the photocell measurement systems. The resulting diagram is divided into 
three different phases: 1) an acceleration phase, with rapidly decreasing section times; 2) a middle 
phase with relatively constant section times; and 3) a deceleration phase with increasing section 
times. The results show that Runner B was faster over all sections and that the differences in the 
runners’ section times increased from the start to the end of the 100 m run. The larger number of 
section times derived from the GNSS compared to the photocell set-up (39 vs. 5) provided more 
detailed and different information. For example, the shortest section times measured with the 
GNSS system were in the sections between 37.5 and 40 m, and 82.5 and 85 m for runners A and 
B, respectively. However, when the section times from the photocell measurements were analysed, 
the shortest time was between 40 and 60 m for Runner A and between 60 and 80 m for Runner 
B. Further, the photocell times show that Runner B was accelerating up to the region between 40 
and 60 m with no possibility of providing a more detailed point. In contrast, the more detailed 
GNSS demonstrates that Runner B only accelerated up to 40 m, after which the runner’s speed 
was relatively constant for ~10 m (4 sections) and thereafter started to decrease. At the end of 
the run, the section times taken from the photocells for Runner B only show a small increase, 
whereas the section times measured by the GNSS system increased considerably. 
Similar to the times of the segments, the absolute velocities of the race were presented with the 
“line and overlaid scatter plot” and with three identical segments (see Figure 3). Changing the 
velocity from time- to space-related enabled a direct comparison of the runners according to 
their positions. Velocities acquired with the use of GNSS measurements represent instantaneous 
velocity at every part of the track. The curve of absolute velocities in relation to space showed the 
acceleration of both athletes in the first phase (“acceleration phase”), with clearly visible areas 
of first-foot take-off, changing of legs and the second-foot take-off. In the first strides, up to the 
7.5 m mark, no significant differences in velocities were noticed between the athletes. From that 
point onwards, the difference in the absolute velocity of the runners increased and remained 
unchanged between the 17.5 and 47.5 m marks (0.2–0.3 m/s). Further, the velocity curves for both 
runners have oscillations corresponding to inter-stride velocity changes. With Runner B, the 
oscillation of velocity during strides in the “middle phase” was higher than with Runner A. This 
difference was particularly noticeable in the “deceleration phase” as the oscillation in the velocity 

GNSS in sprint running 19 Kinesiologia Slovenica, 16, 3, 14–22 (2010)
of Runner B increased in comparison to the previous phase, whereas with Runner A it decreased. 
In addition to velocities for every 0.25 m, Figure 3 also clearly shows instantaneous velocities for 
every 2.5 m; the velocities varied between maximum and minimum values. Similarly to the time 
values (Fig. 2), also in this parameter (Fig. 3) the higher density of data (GNSS) in relation to the 
lower density (photocell) provides different results for the areas of acceleration, deceleration, 
maximums, minimums etc. 
Figure 3: Velocities (y axis) derived every 0.25 m on the sprint track (x axis) with the GNSS 
RTK system and every 20 m with photocells on a 100 m track for two different runners. Note: 
triangles and circles for GNSS-measured velocities are drawn every 2.5 m in order to emphasise 
the instantaneous velocity in the GNSS time segments.
Figure 2: Section times (y axis) measured every 2.5 m on the sprint track (x axis) with the GNSS 
RTK system and every 20 m with photocells on a 100 m track for two different runners. Note 
that the times measured by the photocells were divided by the ratio of the photocells and the 
GNSS RTK interval length to allow an easier comparison and now show the average times as 
they would be for 2.5 m intervals.

20 GNSS in sprint running Kinesiologia Slovenica, 16, 3, 14–22 (2010) 
DISCUSSION
The main findings in this study are: 1) a new method of measuring time segments in a 100 m 
sprint with the use of Global Navigation Satellite System measurements has been developed, 
providing relatively small differences in measured time in comparison to the photocell method 
0.0021 s (s = 0.0051); 2) the new method of measuring time enabled a precise direct location 
comparison between the runners; and 3) the additional times and velocities in sprint running 
using the new method provide “hidden” data along the run, which can help experts or scientists 
analyse performances.
The time computation method provided times from GNSS RTK surveyed trajectories in all 
required sections (reference planes) in the study. Further, the time differences between the 
photocells and GNSS derived times were small. These confirm the method’s robustness to small 
position errors of GNSS measurements as present in the current study. Thirty-nine sections from 
the GNSS show comparable and reliable results, even though the sections are only 2.5 m long and 
shorter in the constant velocity phase than one-third of a second. In addition to the five section 
times and five average section velocities from the photocells, these results reveal many hidden 
details in performance and inter-stride velocity changes. Although the photocells allowed the 
setting of several split times, they did not provide measurements of instantaneous velocity the 
way GNSS could, but mere average velocities of intervals. Moreover, it could be argued that the 
athletes’ posture when breaking the beam of the photocells plays an important role (Fig. 1). Differ-
ent parts of the body can break the beam at different times, which is a disadvantage of photocells 
and the reason more sophisticated and more expensive double photocells are often used, which 
start/stop the time only when beams from both of them are broken. Regardless of the fact that 
in most competitions the time is measured with photocells, an athlete’s performance is more 
accurately estimated by their centre of mass. Therefore, it might be speculated that the GNSS 
system with an antenna attached to the upper thoracic spine (T2-T4) representing the athlete’s 
trajectory of movement passing reference planes more accurately reflects the performance than 
the coincidence of a body part breaking the photocell beam. 
Can time computation based on RTK GNSS also be used in other sports?It provides reliable and 
accurate data; therefore, it can be used for scientific purposes and training at a high perform-
ance level in sports with strong technical demands, which has the benefit of providing detailed 
information in order to improve performance. It enables as many section times as needed with 
no additional costs; they can be even shifted or added after the measurements. Bearing our 
results from sprinting in mind, all these sections or combinations of sections may enable new 
important correlations between performance and physical abilities, as often studied in 100 m 
sprints (Cronin & Hansen, 2005; Young, 1995; Bret, Rahmani, Dofour, Mesonnier, & Lacour, 
2002). However, three major shortcomings have been identified: 1) the high price compared to the 
photocell method when only split times are required; 2) it is relatively bulky and heavy to wear 
even though the athletes did not complain about this; and 3) consequently it is impossible to use 
it during competitions. It still represents a very good alternative to conventional photocell timing, 
especially for larger amounts of intermediate times. It can be expected that future progress will 
minimise these shortcomings. 
The same computation method as presented in the current study may also be used in combination 
with other measuring technology, such as a local positioning system, laser guns, three-dimensional 
kinematics or other sensor-based systems that capture full body motion movement. What is 

GNSS in sprint running 21 Kinesiologia Slovenica, 16, 3, 14–22 (2010)
important is that they provide adequate accuracy of the representative trajectory in comparison 
to the desired time resolution. However, in the case of larger measuring errors, filter methods 
can be used to avoid computation problems of passing the same reference plane more than once, 
but this does not necessarily mean that the time accuracy will be improved. Further, using a low 
cost GPS that only has good relative accuracy, which is appropriate for velocity measurements, 
the whole trajectory or some points may only be shifted according to reference positions; this 
could lead to large systematic errors in time measurements. 
CONCLUSION
The method of computation time using a high-end GNSS RTK surveyed trajectory provides 
reliable and accurate results for measurements in sprint running. From the usage point of view, 
it represents a good alternative to the established photocell measurement, especially for a highly 
detailed or scientific analysis. In addition, due to changes in the body posture during perform-
ance it might be even more accurate than standard and established photocells. 
REFERENCES
Ai, K., & He, S. (1999). Development of the speed measuring system with GPS technology for rowing and 
canoeing. Sports Science/Tiyu Kexue, 19, 56–58.
Bret, C., Rahmani, A., Dofour, A.B., Mesonnier, L., & Lacour, J.R. (2002). Leg strength and stiffness as 
ability factors in 100m sprint running. Journal of Sports Medicine andPhysical Fitness, 42(3), 274‒281.
Coppenolle, H., & Delecluse, C. (1989). Technology and development of speed. Athletics Coach, 23(1), 
82–90.
Cronin, J.B., & Hansen, K.T. (2005). Strength and power predictors of sports speed. Journal of Strength 
and Conditioning Research, 19(2), 349‒357.
Donatti, A. (1995). The development of stride length and frequency in sprinting. New Studies in Athletics, 
1, 51–66.
Edgecomb, S., & Norton, K. (2006). Comparison of global positioning and computer-based tracking sys -
tems for measuring player movement distance during Australian football. Journal of Science & Medicine 
in Sport, 9, 25–32.
Karboviak, R.J. (2005). Using GPS Technology to Monitor Intensity, Speed, and Training Volume in 
Outdoor Athletes. Strength & Conditioning Journal, 27, 24–25.
Larsson, P. (2003). Global positioning system and sport-specific testing. Sports Medicine, 33, 1093–1101.
Mero, A., Komi, P., & Gregor R. (1992). Biomechanics of sprint running. Sports Medicine, 13(6), 376–
392.
Muller, H., & Hommel, H. (1997). Biomechanical Research Project at the VIth World Championship in 
Athletics, Athens 1997. New Studies in Athletics, 2–3, 43–56.
Parkinson, B.W., & Spilker, J.J., (eds.) (1996). Global Positioning System: Theory & Applications, Volume I. 
American Institute of Aeronautics & Astronautics, Inc., Washington.
Rodriguez, D.A., Brown, A.L., & Troped, P.J. (2005). Portable global positioning units to complement 
accelerometry-based physical activity monitors. Medicine & Science in Sports & Exercise, 37, 572–581.
Schutz, Y., & Herren, R. (2000). Assessment of speed of human locomotion using a differential satellite 
global positioning system. Medicine & Science in Sports & Exercise, 32, 642–646.

22 GNSS in sprint running Kinesiologia Slovenica, 16, 3, 14–22 (2010) 
Skaloud, J., & Limpach, P. (2003). Synergy of CP-DGPS, Accelerometry and Magnetic Sensors for Precise 
Trajectography in Ski Racing. In ION GPS GNSS. Portland, Oregon.
Terrier, P., Ladetto, Q., Merminod, B., & Schutz, Y. (2000). High-precision satellite positioning system as a 
new tool to study the biomechanics of human locomotion. Journal of Biomechanics, 33, 1717–1722.
Tellez, T., & Doolittle, D. (1984) Sprinting from start to finish. Track Technique, 88, 2802–2805.
Townshend, A.D., Worringham, C.J., & Stewart, I.B. (2008). Assessment of speed and position during 
human locomotion using nondifferential GPS. Medicine & Science in Sports & Exercise, 40, 124–132.
Troped, P.J., Oliveira, M.S., Matthews, C.E., Cromley, E.K., Melly, S.J., & Craig, B.A. (2008). Prediction 
of activity mode with global positioning system and accelerometer data. Medicine & Science in Sports & 
Exercise, 40, 972–978.
Witte, T.H., & Wilson, A.M. (2004). Accuracy of non-differential GPS for the determination of speed over 
ground. Journal of Biomechanics, 37, 1891–1898.
Witte, T.H., & Wilson, A.M. (2005). Accuracy of WAAS-enabled GPS for the determination of position 
and speed over ground. Journal of Biomechanics, 38, 1717–1722.
Young, W. (1995). Laboratory strength assessment of athletes. New Studies in Athletics, 10 (1), 89–96.