3d-cg based musculoskeletal simulation for a swimmer ... · 12/3/2014 · on swimmers taking a...

Procedia Engineering 60 ( 2013 ) 367 – 372

1877-7058 © 2013 The Authors. Published by Elsevier Ltd.Selection and peer-review under responsibility of the School of Aerospace, Mechanical and Manufacturing Engineering, RMIT Universitydoi: 10.1016/j.proeng.2013.07.071

Available online at www.sciencedirect.com

6th Asia-Pacific Congress on Sports Technology (APCST)

3D-CG based musculoskeletal simulation for a swimmer wearing competitive swimwear

Motomu Nakashimaa*, Takahiro Hasegawab, Akihiro Matsudac, Takatsugu Shimanad, Kazuhiro Omorid

aGraduate school of engineering, Tokyo Institute of Technology, 2-12-1 Ookayama, Meguro, Tokyo 152-8552, Japan bGraduate school of information science and engineering, Tokyo Intitute of Technology, Tokyo, Japan

cFaculty of engineering, Information and systems, University of Tsukuba, Ibaraki, Japan dResearch and development department, Mizuno Corporation, Osaka, Japan

Received 15 April 2013; revised 20 May 2013; accepted 27 May 2013

Abstract

The objective of this study was to develop a musculoskeletal simulation for a swimmer wearing competitive swimwear. For the simulation, the body geometry and joint motion were put into the swimming human simulation model SWUM in order to calculate the distributed fluid forces acting on the whole swimmer’s body. The distributed fluid forces were put into the musculoskeletal model. In addition to the fluid forces, the forces induced by the tension of the swimwear were also considered. As an example, the simulation for the crawl stroke was conducted. The results showed the time-varying effect of the swimwear on the muscle activity during swimming. © 2013 Published by Elsevier Ltd. Selection and peer-review under responsibility of RMIT University Keywords: Swimwear; musculoskeletal simulation; swimming; crawl stroke; 3D-CG

1. Introduction

Swimwear has been a key issue in competitive swimming since it was known that its function certainly affects racing performance. Its fluid dynamics properties have been main issue, such as special surface processing for the reduction in frictional drag as well as the insertion of hard panels for the suppression in swimmer’s flesh vibrations. However, the effect of swimwear on the muscle activity has not been studied

* Corresponding author. Tel.: +81-3-5734-2586. E-mail address: [email protected].

© 2013 The Authors. Published by Elsevier Ltd.Selection and peer-review under responsibility of the School of Aerospace, Mechanical and Manufacturing Engineering, RMIT University

368 Motomu Nakashima et al. / Procedia Engineering 60 ( 2013 ) 367 – 372

to date. The objective of this study was to develop the method of musculoskeletal simulation for a swimmer wearing competitive swimwear.

Although musculoskeletal simulations are becoming a popular method in the research field of biomechanics to estimate muscle activity from body motion [1], no study had been conducted for the swimming motion. As a pioneering work, Nakashima and Motegi [2] developed the musculoskeletal simulator for swimming. This simulator was used in the present study. In addition, the tension property of the swimwear was modeled and incorporated into the simulation model. The developed simulation method is called “the 3D-CG based musculoskeletal simulation” since the tension property of the swimwear was calculated from the 3D-CG based model for a swimmer.

2. Simulation method

2.1. Overview

The flow of the simulation in the present study is shown in Fig. 1. The body data (body geometry) of the swimmer, the joint angles (relative body motion) during the swimming motion, and the absolute motion of the whole body were put into the swimming human simulation model SWUM, which was developed for the mechanical simulation of swimming [3], [4]. The fluid force acting on the swimmer was computed and output by SWUM. The other input data were also converted into the format for the musculoskeletal simulation. The output data from SWUM were put into AnyBody Modeling System (AnyBody Technology Inc., Aalborg), which is a commercial software for musculoskeletal simulation [5]. The tension property of the swimwear was also put into the musculoskeletal model. Finally, the muscle activity of the swimmer was obtained. The details of each part are described in the following sub-sections.

Fig. 1. Flow of the simulation

2.2. Swimming human simulation model SWUM

The simulation model SWUM was designed to solve the six degrees-of-freedom absolute movement of the whole swimmer’s body as a single rigid body by time integration, using the inputs of the swimmer’s body geometry and relative joint motion. Therefore, the swimming speed, roll, pitch and yaw motions, propulsive efficiency, joint torques, etc. are computed as the output data. The swimmer’s body is represented by a series of 21 rigid body segments as follows: lower and upper waist, lower and upper chest, shoulder, neck, head, upper and lower hip, thighs, shanks, feet, upper arms, forearms, and hands.


Fig. 2. (a) Body representation in SWUM; (b) An example of a breaststroke simulation.

Each body segment is represented by a truncated elliptical cone. The unsteady fluid force and gravitational force are taken into account as external forces acting on the whole body. The unsteady fluid force is assumed to be the sum of the inertial force due to the added mass of the fluid, normal and tangential drag forces and buoyancy. These components are assumed to be computable, without solving for the flow, from the local position, velocity, acceleration, direction, angular velocity, and angular acceleration for each part of the human body at each time step. The coefficients in this fluid force model were identified using the results of an experiment with a limb model and measurements of the drag acting on swimmers taking a glide position. The body representation in SWUM and a simulation example simulation are shown in Fig. 2. Further details about SWUM are described in the literature [3].

2.3. Musculoskeletal model

The musculoskeletal model is shown in Fig. 3(a). The human body is modeled as a rigid body link model with the muscles. The muscles are modeled as 581 wires (294 for the upper limbs, 84 for the lower limbs and 203 for the trunk). The distributed fluid forces acting on the swimmer as well as the tension of the swimwear are input into AnyBody as external forces. The joint torques are calculated in AnyBody, taking the external forces and inertia of the human body itself into account. The muscle forces are

1 1.2 1.4 1.60

20

40

60

80

100

120BHBLFHFL

Nominal stretch

Ten

sion

[N]

Fig. 3. (a) Musculoskeletal model; (b) The representation of seams of the swimwear in the musculoskeletal model (the red, blue and green lines); (c) Results of uniaxial tensile test for four types of seams


obtained by solving the equilibrium equations for the joint torques. In this calculation, an algorithm to determine the distribution of the muscle forces is necessary since the musculoskeletal system is highly redundant, that is, the distribution cannot be determined only from the equilibrium equations. In the present study, the Min/Max Criterion [6] was employed for this algorithm. In this criterion, the maximum among all values of the muscle activation is minimized. Rasmussen et al. [6] claimed that this algorithm is reasonable from the physiological viewpoint and efficient in computation.

2.4. Modeling of competitive swimwear

In the present study, the tension of the swimwear was considered to be the sum of the tension of the seams and that of the uniform cloth, since the tension of the seams are much larger than that of the uniform cloth. The seams were modeled as additional ‘virtual ligaments’ in the musculoskeletal model. The modeled seams are shown in Fig. 3(b). The red, blue and green lines represent the seams. The hyper-elastic tension property of the seams were represented by adjusting five parameters in the ligament model so that the simulated tension property became sufficiently consistent with those obtained in the uniaxial tensile test, whose results for four types of the seams are shown in Fig. 3(c).

Since the tension of the uniform cloth was difficult to represent in the musculoskeletal model itself, it was represented as external torques at the hip joint. The original tension property of the uniform cloth was calculated by the anisotropic hyperelastic model. The data were obtained as the distributed forces at the mesh points, which were defined on the surface of the swimwear. An example of the tension distribution is shown in Fig. 4(a). In order to calculate the hip joint torque from the tension distribution, the ‘attaching ratio’ for each mesh point was newly defined in this study. If a mesh point is completely attached to the thigh segment, the attaching ratio becomes 1:0 for the thigh and pelvis segments, respectively. In this case, the distributed force acting on the mesh point is supposed to completely act on the thigh segment. If another mesh point is located at a position closer to the hip joint, the attaching ratio of this point becomes 0.7:0.3, for example. In this case, the distributed force multiplied by 0.4 (0.7 − 0.3) is supposed to act on the thigh segment. The distribution of the attaching ratio for a swimsuit was determined as shown in Fig. 4(b). It was determined based on the experiment in which the deformation of the swimwear was measured.

Fig. 4. (a) An example of tension distribution obtained from anisotropic hyperelastic model; (b) Colour map of the attaching rates of the mesh points.


In this experiment, a large number of markers were attached on the swimwear worn by a subject. The hip flexion and extension were performed by the subject. The deformation of the swimwear was then measured by a motion analysis system. The attaching ratio was calculated from the displacement of the marker in the experiment.

3. Results and discussion

The musculoskeletal simulation for the crawl stroke was carried out. As the seam condition, the tension properties of BH, FL and BL in Fig. 3(c) were used for the virtual ligaments of green, red and blue in Fig. 3(b), respectively. For the uniform cloth, the tension property calculated in previous work [7] was used. The results of simulated swimming motion are shown in Fig. 5. The left figures are the original 3D-CG animation images. The center figures are the results of SWUM. The right figures are the results of the

Fig. 5. Simulated swimming motion of the crawl stroke (left, CG model; center, SWUM; right, musculoskeletal simulation)

musculoskeletal simulation. It was confirmed that the swimming motion of the original 3D-CG was successfully converted and put into SWUM as well as the musculoskeletal simulation. The simulated results of the hip joint torque in the swimwear in the extension/flexion direction are shown in Fig. 6(a).


The abscissa is the nondimensional time which is normalized by the stroke cycle. The period of one flutter kick is shown in this figure although three kicks were beaten in one stroke cycle. The black line represents the hip joint angle in the flexion/extension direction. It was found that the joint torque opposing the joint angle was produced by the swimwear. That is, when the hip joint was in an extended position, the joint torque was produced in the flexing direction, and vice versa. The muscle forces of flexors and extensors are shown in Fig. 6(b). The red and blue lines represent the sums of the muscle forces of the flexors and extensors, respectively. It was found that the muscle force of the extensors had a large peak when the nondimensional time was 0.6, which corresponded to the beginning of the extending motion. The difference in the muscle force from the no tension condition is shown in Fig. 6(c). It was found that the muscle force of the flexors increased and that of the extensors decreased. It suggests that the swimwear basically acts in the extending direction.

0.4 0.5 0.67

8

9

10

11

12

13

-25

-20

-15

-10

-5

0

Nondimensional time

Join

t ang

le [

deg]

Hip

join

t tor

que

[Nm

] Extension

FlexionTorque by swimwearExtension/Flexion angle

0.4 0.5 0.60

2000

4000

6000

-20

-15

-10

-5

0

Nondimensional time

Join

t ang

le[d

eg]

Mus

cle

forc

e[N

]Extension

Flexion

ExtensorsFlexorsExtension/Flexion angle

0.4 0.5 0.6-400

-300

-200

-100

0

100

200

300

Nondimensional time

Dif

fere

nce

in m

uscl

e fo

rce

[N]

ExtensorsFlexors

Fig. 6. (a) Hip joint torque by swimwear in the direction of extension/flexion; (b) Muscle forces of flexors and extensors; (c) Differences in muscle forces from the no tension condition

4. Conclusion

The method of musculoskeletal simulation for a swimmer wearing competitive swimwear was developed in the present study. An example of the simulation for the crawl stroke was shown and reasonable results were obtained. More detailed analyses for other swimming motions as well as other conditions of the swimwear are future tasks. In addition, the EMG measurement to validate the method is also a future task.

References

[1] Rasmussen J. Challenges in human body mechanics simulation. Proceedings of 2011 Symposium on Human Body Dynamics

2011:176-185.

[2] Nakashima M, Motegi Y. Development of a full-body musculo-skeletal simulator for swimming. Proceedings of the

Eleventh International Symposium on Computer Simulation in Biomechanics 2007:59–60.

[3] Nakashima M, Miura Y, Satou K. Development of swimming human simulation model considering rigid body dynamics and

unsteady fluid force for whole body, J Fluid Sci & Tech 2007;2:56–67.

[4] SWUM developer team. http://www.swum.org/

[5] AnyBody Technology. http://www.anybodytech.com/

[6] Rasmussen J, Damsgaard M, Voigt M, Muscle recruitment by the min/max criterion – a comparative numerical study, J

Biomech 2001;34:409–415.

[7] Matsuda A, Tanabe H, Nagaoka T, Nakashima M, Shimana T, Omori K. Stress calculation of competitive swimwear using

3D-CG based anisotropic hyperelastic model. Procedia Engineering (Proceedings of the 6th Asia-Pacific Congress on Sports

Technology (APCST) 2013.

3d-cg based musculoskeletal simulation for a swimmer ... · 12/3/2014 · on swimmers taking a...

Documents