Content uploaded by Yasin Yuzbasioglu
Author content
All content in this area was uploaded by Yasin Yuzbasioglu on Aug 17, 2021
Content may be subject to copyright.
Computer Simulatıon Model For Biomechanical Analysis Of Overhead Throwing In
Handball: A Pilot Study
1
Yasin Yüzbaşıoğlu, 2Nihat Şükrü Özgören
1Department of Coaching Education, Faculty of Sport Sciences, Ege University, İzmir, Turkey
2Biomechanics Research Group, Faculty of Sport Sciences, Hacettepe University, Ankara,
Turkey
E-mail: yasinyuzbasioglu@gmail.com
Öz
Atış hareketi hentbolda temel oyun becerilerinden biridir. Bu çalışmanın amacı atış
hareketinin biyomekanik analizi için bir bilgisayar benzeşim modeli oluşturmaktır.
Araştırmaya bir erkek hentbolcu katıldı. Katılımcının gövdesine ve topun üzerine toplam 24
adet yansıtıcı işaret yerleştirildi. Katılımcının farklı kütledeki (0.350, 0.450, 1 kg) toplarla
gerçekleştirdiği temel atışlar kaydedildi. Noktaların üç boyutlu uzaysal konum değerleri ve
üyelerin dönme açıları hesaplanarak dışa aktarıldı. Ters dinamik analizle eklem momentlerini
hesaplamak için Simulink/MATLAB ortamında Simscape kütüphaneleri kullanılarak insan
vücudunun, sekiz üyeli katı cisim mekanik modeli oluşturuldu. Üyelerin eylemsizlik
momentleri Dempster parametreleri kullanılarak hesaplandı. Modelin girdileri topun zamana
bağlı doğrusal yer değiştirmesi, kütlesi ve eylemsizlik momenti, üye kütle değerleri, kütle
merkezi konumları, eylemsizlik momentleri ve eklemlerin zamana bağlı açı, açısal hız ve
açısal ivme değerleridir. Hareketlerin ters dinamik benzeşimi sabit adım aralığında (0.0042 s),
ode4 denklemleri kullanılarak gerçekleştirildi. Modelin çıktıları olan dirsek eklem
momentlerinin ve topun elden çıkış hızlarının, topun ağırlığına bağlı değişimleri incelendi.
Sırasıyla 0.350, 0.450 ve 1 kg kütleli toplarla yapılan atışlarda topun elden çıkış hızı 10.35,
9.50 ve 8.57 m/s’dir. Topun elden çıkış anında elde edilen net dirsek momentleri 29.5, 27.5,
8.4 N.m (pronasyon), -18.0, -12.8, -1.7 N.m (dış rotasyon) ve 2.3, -6.8 ve -13.7 N.m’dir
(ekstansiyon/fleksiyon). Topun kütlesi arttıkça topun elden çıkış hızı düşmektedir; dirsek
eklemindeki pronasyon ve dış rotasyon momentleri azalırken fleksiyon momenti artmaktadır.
Anahtar Kelimeler: Hentbol, Temel Atış, Modelleme, Benzeşim, Eklem Kinetiği
Abstract
Overhead throwing is a fundamental skill in handball. This study aims to build a computer
simulation model for the biomechanical analysis of overhead throwing. One male handball
player attended the study. Twenty-four markers were attached at specific anatomical
landmarks and ball. The participant's overhead throws with weighted balls (0.350, 0.450, 1
kg) were captured using eight infrared cameras. Locations of the markers and segment
rotation angles in three-dimensional space were calculated and exported. An eight-segment
mechanical of the rigid body was build using Simscape libraries in Simulink/MATLAB. The
segment inertial parameters were calculated using Dempster parameters. Input to the model
comprised time histories of ball translations, ball mass, segment mass and inertia matrix, time
histories of joint angle, angular velocities and accelerations. The inverse dynamics
simulations were performed using the ode4 solver with a fixed step size (0.0042s). Outputs
from the model comprised elbow joint torques and the ball release velocity. The ball release
velocities were 10.35, 9.50 and 8.57 m/s at overhead throwings with 0.350, 0.450 and 1 kg
balls, respectively. The net elbow torques were calculated as 29.5, 27.5, 8.4 N.m for
pronation, -18.0, -12.8, -1.7 N.m for external rotation, and 2.3, -6.8, -13.7 N.m for
extension/flexion. The release velocity of the ball decreased as the ball mass increased. The
lower elbow pronation and external rotation torques, and higher flexion torques were obtained
while throwing a heavier ball.
1
Keywords: Team Handball, Overhead Throwing, Modelling, Simulation, Joint Kinetics
Introduction:
The overhead throw is a fundamental skill for scoring and passing the ball among the players
in team handball. It is an example of a complex movement and an essential factor for
performance [1]. One of the key factors is the ball release velocity for overhead throwing
performance. Based upon principles of different load, various training methods are used either
by strength or velocity of the throwing [2]. The studies with overweighted and underweighted
equipment could raise the throwing velocity [3, 4]. Van den Tillaar and Ettema [5] pointed
that the better throwers have a higher internal rotation of the shoulder and elbow extension
during throws. Generally, in team handball, some studies reported the linear velocity of
segments or angular velocities of the joints during throwing [6, 7]. Consequently, the major
factor contributing to this change in ball velocity is the work done by the muscles, which
alters the joint torques.
From this point of view, this study aimed to build a biomechanical simulation model for
investigating the kinetics of the upper extremity during overhead throwing.
Methods:
One male handball player (age 37 years, 90 kg, 1.80 m) participated in this study. Eight
infrared cameras (Vicon, Bonita, U.K.) were placed at proper heights using tripods around the
field of motion (Figure 1). The cameras and floor plane were calibrated sequentially using the
Vicon Blade software (version 2.6.1) calibration tool. A table was placed in the middle of the
capture field to let the participant sat on it with legs hang. Then the pelvis fixed to the table
with a belt to minimize the rotational or translational movement of the pelvis.
Figure 1. The view of the eight-camera setup from directly overhead. The arrow indicates the
throw direction.
A custom eight-segment body model was generated in Blade software comprising pelvis,
abdomen, thorax, clavicula (left-right), head, right upper arm and forearm (Figure 2). This
2
model requires a total of 23 markers. Thus, we attached 23 markers at the specific anatomical
landmarks and one extra marker at the top of the ball. The distances between anatomical
markers defining segment lengths and widths were measured as well as the segment
circumferences. The participant made a general warm-up of 15 min consisted of jogging and
throwing drills for the throwing arm. First, the range of motion performance of the subject
was captured. Second, the participant performed three overhead throws to a catcher standing 6
m away. The participant was asked to achieve the personal maximum ball release velocity
during throws with a standard ball (0.450 kg), light ball (0.350 kg) and heavy ball (1 kg).
Third, the subject's motions were captured and recorded using Blade software. The sampling
frequency was 240 frames per second. The steps of data processing using Blade software are
as follows:
1. Labelling markers in the range of motion scene,
2. Calibration of the created skeleton to the range of motion data,
3. Labelling the unlabeled markers in overhead throw scenes using the calibrated
skeleton,
4. Solving the motions using inverse kinematics solver of Blade software.
5. Exporting three-dimensional labelled marker trajectories as .trc files,
6. Exporting three-dimensional orientation of the segments (Euler angles) as .bvh files.
Figure 2. Frontal view of the labelled markers and solving bone created using Blade software.
The exported BVH (skeleton angles) and TRC (marker trajectories) files were imported into
MATLAB (R2021a) using custom parser scripts. The raw marker trajectories and bone angles
were filtered using a 6th order low-pass Butterworth digital filter with a cut-off frequency of 8
Hz to eliminate the fluctuations. The angular velocity and angular acceleration of the
segments were calculated from filtered angular displacement data. The ball release frame was
detected by a custom script, which takes the distance between the wrist and ball markers into
account.
An eight-segment mechanical model was built (Figure 3) by using Simscape libraries in
Simulink (version 10.3). The root body of the model is the pelvis bone, connected to the World
3
frame with a Bushing joint that has six degrees of freedom (three rotational and three
translational). Other segments within the body were connected with Gimbal joint with three
rotational degrees of freedom. The ball was connected to the wrist with a Cartesian joint that
has three translational degrees of freedom. The segments were assumed rigid rectangular
prisms. The location of the mass centre, mass, and the inertia of the segments were calculated
using Dempster [8] parameters (Table 1).
Table 1. Segmental inertia parameters of the simulation model.
Segment Mass (kg) Length (m) Moment of inertia (kg m2) [X,Y,Z]
Pelvis 12.78 0.200 [0.045, 0.045, 0.005]
Abdomen 12.51 0.225 [0.055, 0.055, 0.005]
Thorax 19.44 0.178 [0.055, 0.055, 0.008]
Clavicula 1.4220 0.185 [0.004, 0.004, 0.000]
Upper arm 2.52 0.295 [0.018, 0.018, 0.001]
Forearm 1.44 0.231 [0.006, 0.006, 0.000]
Head 7.29 0.154 [0.016, 0.016, 0.003]
Inputs of the model were:
1. Translation-time and angle-time histories of the pelvis,
2. Joint angle-time histories of the abdomen, thorax, neck, shoulder (right and left) and
right elbow,
3. The initial position of the joints,
4. Translation-time histories of the ball,
5. Segmental mass and inertia.
Figure 3. The inverse dynamics simulation of the overhead throw in Simulink.
4
A total of 9 simulations (3 for each weighted ball) were performed using the ode4 (Runge-
Kutta) solver of the Simulink with fixed step size, 0.0042 s (same as the sampling frequency).
Outputs of the simulations were:
1. Elbow joint torques
2. Ball release velocity
Only the simulations that resulted in the maximum ball release velocity among three throws
were interpreted.
Results:
The ball release velocities and elbow joint torques can be seen in Table 2.
Table 2. Simulation outputs.
Ball mass
(kg)
Ball release
velocity (m/s)
Elbow torque (Nm)
Supination(-)
Pronation(+)
External (-)
rotation
Internal (+)
rotation
Flexion(-)
Extension(+)
0.350 10.35 29.5 -18.0 2.3
0.450 9.50 27.5 -12.8 -6.8
1 8.57 8.4 -1.7 -13.7
Conclusion:
The ball velocity increased 8.95% in the light ball and decreased 9.80% in the heavy ball
compared to the standard (Table 2). Lower elbow pronation and external rotation torques, and
higher flexion torques were obtained while throwing a heavier ball. Throwing with the heavy
ball increased external rotation torques, suggesting that this will be an effective exercise for
safely building biceps strength.
5
References:
[1] Werner, S.L., et al., Biomechanics of the elbow during baseball pitching. J Orthop
Sports Phys Ther, 1993. 17(6): p. 274-8.
[2] Ettema, G., T. Glosen, and R. van den Tillaar, Effect of specific resistance training on
overarm throwing performance. Int J Sports Physiol Perform, 2008. 3(2): p. 164-75.
[3] Escamilla, R.F., et al., Effects of throwing overweight and underweight baseballs on
throwing velocity and accuracy. Sports Med, 2000. 29(4): p. 259-72.
[4] Caldwell, J.E., F.J. Alexander, and C.S. Ahmad, Weighted-Ball Velocity Enhancement
Programs for Baseball Pitchers: A Systematic Review. Orthop J Sports Med, 2019.
7(2): p. 2325967118825469.
[5] van den Tillaar, R. and G. Ettema, A force-velocity relationship and coordination
patterns in overarm throwing. J Sports Sci Med, 2004. 3(4): p. 211-9.
[6] Fradet, L., et al., Do handball throws always exhibit a proximal-to-distal segmental
sequence? J Sports Sci, 2004. 22(5): p. 439-47.
[7] Joris, H.J., et al., Force, velocity and energy flow during the overarm throw in female
handball players. J Biomech, 1985. 18(6): p. 409-14.
[8] Dempster, W.T., Space requirements of the seated operator, geometrical, kinematic,
and mechanical aspects of the body with special reference to the limbs. 1955,
Michigan State Univ East Lansing: Wright-Patterson Air Force Base, Ohio.
6