Training with a Pneumatic Assist Suit to Generate Lower-Body

Twisting during the Forehand Swing in Table Tennis

Akihiko Kashiwagi, Hayato Nakamura, Satoshi Nishikawa

and Kazuo Kiguchi

Faculty of Engineering, Kyushu University, 744 Motooka, Nishi-ku Fukuoka, Japan

Keywords: Pneumatic Actuator, Wearable Robot, Sports Support.

Abstract: Movement form is important in sports. However, self-learning of the form has a possibility that people will

acquire a form which puts a burden on one part of the body or a form which cannot exert sufficient power.

Although there is also a way to have an instructor, people do not always receive the instruction of the form.

Therefore, it is useful for sports training to develop a device which allows participants to acquire a form suited

for each sport. Existing research about support in sports with a racket does not pay much attention to the lower

body. In this research, we developed an assist suit that assists the lower body in executing a forehand swing

in table tennis with lower-body twisting. Using this suit for beginners of table tennis, we conducted

experiments under four conditions:(1) "No wear (before)", (2) "Without assist", (3) "With assist", and (4) "No

wear (after)". As a result of Tukey analysis within each participant, the range of movement of the lower body

is statistically increased by the assist suit and there are individual differences in whether to acquire a swing

with twisting the lower body.

1 INTRODUCTION

1.1 Background

Exercising in sports is one of the effective ways of

maintaining physical health and reducing stress.

However, a lot of learning and practice time is

required for beginners to master how to move their

bodies in each sport (Wulf & Shea, 2002).

As a training method for sports at present, two

typical examples are considered: self-learning like

using some services such as a pitching machine and

guidance by an experienced person. In the case of

self-learning, however, there is a possibility that a

beginner will acquire such as a form that puts a

burden on only one part of his/her body or a form that

is difficult to transmit force. Also, in the case of

guidance by an experienced person, an experienced

instructor is necessary. From the above, it is

considered that it is necessary to make a device that

can help a beginner acquire an ideal form. In this

research, we focused on table tennis, which is popular

with a wide range of ages, to investigate the

supporting method for the training of forehand swing.

https://orcid.org/0000-0003-0905-8615

https://orcid.org/0000-0003-4408-0420

1.2 Previous Research

To provide efficient sports support, it is necessary to

understand the parts of the body that greatly affect

sports movements. In previous studies of golf (Evans

& Tuttle, 2015) and tennis (Gordon & Dapena, 2006),

it is shown that waist twisting exercise is important.

Qian et al. investigated differences between advanced

and intermediate table tennis players in lower limb

movements during the forehand swing of table tennis.

In their research, it was found that intermediate

players rotated the hip joint and trunk lesser than

advanced players and the exercise skills of the lower

limbs were lower than advanced players (Qian et al.,

2016). Also, Zhang's study states that advanced

players have a greater range of motion during swings

and trunk twisting is important in table tennis (Zhang,

2017).

About sports support methods, it has been found

that tactile guidance is efficient to learn movement

more quickly (David & James,2009; Huang et al.,

2007). As a sport-aided device using tactile guidance,

there are robots that teach swing by controlling the

movement of tennis rackets and golf clubs (Kümmel

Kashiwagi, A., Nakamura, H., Nishikawa, S. and Kiguchi, K.

Training with a Pneumatic Assist Suit to Generate Lower-Body Twisting during the Forehand Swing in Table Tennis.

DOI: 10.5220/0011557000003321

In Proceedings of the 10th International Conference on Sport Sciences Research and Technology Support (icSPORTS 2022), pages 61-67

ISBN: 978-989-758-610-1; ISSN: 2184-3201

et al., 2014; Hirata et al., 2017). In addition, there are

studies that use wearable robots to assist in running

and swing movements (Miyazaki et al., 2021; Zhou

et al., 2021; Klein et al., 2012; Sakoda et al., 2018).

These studies supported a body part that directly

affects the action point of force. However, as

mentioned in the first paragraph of this subsection, it

is stated that the exercise of the part far from the point

of action is also an important factor in playing sports.

Therefore, it is expected that the performance will be

improved by supporting the training of the part far

from the point of action.

1.3 Purpose of Research

This study aims to achieve effective training by

developing an assist suit to help the wearer master the

form by inducing the twisting motion of the lower

body during the forehand swing of table tennis.

2 ASSIST SUIT DESIGN

Figure 1 shows the appearance of the developed assist

suit.

A B

Figure 1: A) The back of the assist suit. B) Position of

pneumatic muscles.

2.1 Actuator

An assist suit for sports needs to be lightweight, not

interfere with movement, and be able to safely control

the exerted force of the suit. Therefore, in this study,

we adopted pneumatic artificial muscle as an

actuator. Pneumatically driven artificial muscles have

a large exerting force relative to weight, flexible

materials, and the characteristic that the exerting

force decreases as the length of the artificial muscle

decreases from the start to the end of a contraction.

By the last feature, the artificial muscle does not give

excess power. So, there is little risk of injury to the

wearer. These features are consistent with the

requirements for assist suits.

2.2 Placement of Artificial Muscles

Figure 1.B shows the placement of the artificial

muscle of the assist suit. Three artificial muscles are

used for the assist suit. To get sufficient power for the

twist of the lower body, a longer artificial muscle is

needed. Also, by increasing the length of muscle

fixed parallel to the waist, the perpendicular force for

the axis of the body is increased. By this structure, the

force for causing the twisting motion of the lower

body is enough. Considering the above, the artificial

muscle #1 shown in Figure 1.B is fixed in a way that

makes one circumference of the thigh from the knee

to the waist. To fix the knee supporter as fixing the

position of the artificial muscles, the muscles #2 and

#3 are used. Artificial muscle is fabricated using the

braided tube (DENKA ELECTRON Co., NFL-19),

and the initial diameter is 12 mm. Table 1 shows the

initial length of each artificial muscle and the length

after contraction when the internal pressure is 0.6

MPa at no load.

Table 1: Length of each pneumatic muscle (mm).

1 2 3

Normal

(0MPa)

820 315 165

Maximum contraction

(

0.6MPa

)

600 220 115

2.3 Control Mechanisms

When the push-button (ELPA (Asahi Electric) Co.,

HK-PSS04H) is pressed, a signal to open the valve is

output from the microcontroller (Arduino Co.,

Arduino Uno R3). The signal is converted to a voltage

with a DA converter (ANALOG DEVICES Co.,

AD5308), and the valve (HOERBIGER Co., tecno

basic) changes the pressure in artificial muscles to a

specific value according to the voltage. These parts

are fixed to an acrylic plate (Kuraray Co.,

COMOGLAS). Then, this control unit is put in a bag.

The total weight of the bag is 1.52 kg. The total

weight of the assist suit is 2.12 kg. Since this suit is

for right-handers, so as not to interfere with the swing

of the right hand, it was placed over the left shoulder.

In addition, air from the compressor (JUN-AIR Co.,

6-4) is stored in a tank (JC Service Co., ECO JET E)

and this tank is carried on the back. By doing this, the

time of supplying air to the artificial muscles gets

shorter. In this experiment, the participants pressed a

button with their left hand to control the internal

icSPORTS 2022 - 10th International Conference on Sport Sciences Research and Technology Support

pressure of the artificial muscle to rise from 0 MPa to

0.6 MPa. After 0.3 s, the air let out automatically.

3 VERIFICATION EXPERIMENT

OF SUIT EFFECT

3.1 Purpose of the Experiment

The purpose of the experiment is evaluating the effect

of the suit on beginners in table tennis by conducting

experiments with the developed suit.

3.2 Experimental Method

Table 2 shows each participant information. All

participants are right-handed. In order to evaluate the

improvement of swing ability by the suit, we used two

parameters: the amount of rotation of the waist

around the vertical axis during a swing and the

maximum racket speed.

The following four experimental conditions were

set: condition (1) to examine the participant’s swing

before training, condition (2) to examine the effect of

wearing this suit, condition (3) to examine the support

effect, and condition (4) to examine the training effect.

(1) Not wear the assist suit.

(2) Wear the assist suit without assist.

(3) Wear the assist suit with assist.

(4) Not wear the assist suit

From here, each condition will be described as (1)

"No wear (before)", (2) "Without assist", (3) "With

assist", and (4) "No wear (after)".

In each experiment, a table tennis ball launcher

(Nittaku Co., NB-1150) provided a ball 20 times. In

addition, data on the participant's swing to strike the

ball were recorded. The trajectory of balls provided

by the table tennis ball launcher was set to bounce at

a position of 495 mm in the Z-axis direction (in

Figure 2) and -365 mm in the X-axis direction (in

Figure 2) on the participant side court from the center

of the table. The ball speed was set to 3.64 m/s, the

rotation speed was set to 26.7 round/s in the upward

rotation, and the ball launch interval was set to 1.83 s.

The experiment was conducted with the approval of

the Experimental Ethics Committee of the Faculty of

Engineering, Graduate School of Kyushu University

(approval number 2021-04).

Table 2: Each participant data.

Age Sex Height

[cm]

Body mass

[kg]

Participant1 22 Male 175 60

Participant2 23 Male 166 51

Partici

ant3 23 Male 170 50

Partici

ant4 21 Male 173 61

Partici

ant5 21 Male 173 58

Participant6 22 Male 167 67

Figure 2: The Arrangement of measuring equipment. This

is a side view of the experiment room. The blue triangles

(1-8) are motion capture cameras. The green rectangles are

normal cameras.

Figure 3: The Appearance of the experiment.

3.3 Measuring Instruments and

Analysis Methods

For measurement of two parameters, we use eight

motion capture cameras (NaturalPoint Inc., Prime

13W) installed at positions 1-8 as shown in Figure 2.

Also, Figure 3 shows the experiment appearance. For

participants 1-6, the three-dimensional position

coordinates of the center of the racket defined as rigid

bodies were measured. Also, for participants 1-3, the

three-dimensional position coordinates of the center

of the waist defined as rigid bodies were measured.

For participants 4-6, the three-dimensional position

coordinates of upper body skeleton were measured.

The sampling rate was 100 Hz.

The amount of rotation around the vertical axis of

the waist was calculated based on the measured value

of the Euler angle in the Y-axis direction of the rigid

Training with a Pneumatic Assist Suit to Generate Lower-Body Twisting during the Forehand Swing in Table Tennis

body. In addition, the racket speed was calculated by

dividing the moving distance of the racket between

frames by the time between frames. Regarding the

amount of rotation of the waist, the minimum value

of the angle around the Y axis was defined as the start

of the waist rotation and the maximum value of the

angle around the Y axis was defined as the end of the

waist rotation. Also, the angle difference between the

start and end of the waist rotation was calculated as

the rotation amount of the waist. About the racket

swing, the minimum value of the X coordinate was

defined as the start of the swing and the minimum

value of the Z coordinate was defined as the end of

the swing. Also, the maximum racket speed was

examined in the swing.

To investigate the effect of the suit, the amount of

waist rotation and the racket speed between the

conditions were compared. First, using the mean of

20 trials with each condition in each participant, four

conditions were statistically compared by paired t-test

with the Bonferroni correction. Second, the Tukey

method was used to statistically compare four

conditions within each participant for detailed

analysis. The significance level was set to

p < 0.05.

4 EXPERIMENTAL RESULTS

4.1 Analysis across Participants

Figure 4 shows the waist yaw angle of all participants.

Figure 5 shows the racket speed of all participants. In

the comparison of waist yaw angle between "With

assist" and "No wear (after)", there was a significant

decrease.

Figure 4: Waist yaw angle of each condition (Bonferroni

method *: p<0.05 **: p<0.01). Each plot shows the mean

of 20 trials with each condition in each participant.

Figure 5: Racket speed of each condition (Bonferroni

method *: p<0.05 **: p<0.01). Each plot shows the mean

of 20 trials with each condition in each participant.

4.2 The Amount of Waist Rotation

Figure 6 shows the results of the amount of waist

rotation within each participant. Although there was

no significant difference in the comparison between

"No wear (before)" and "Without assist" in all

participants except participant 6, there is a significant

increase in the comparison between "No wear

(before)" and "With assist", in the comparison

between "Without assist" and "With assist". In

addition, in participants 2 and 5, there is a significant

increase in the comparison between "No wear

(before)" and "No wear (after)".

4.3 Racket Speed

Figure 7 shows the results of the racket speed within

each participant. In the comparison between "No

wear (before)" and "Without assist", there was a

significant decrease in participants 1 and 6, no

significant difference in participants 2, 4 and 5, and a

significant increase in participant 3. In the

comparison between "Without assist" and "With

assist", there was a significant increase in participants

1, 2, 5 and 6. In addition, in participants 2 and 5, there

was a significant increase in the comparison between

"No wear (before)" and "No wear (after)".

icSPORTS 2022 - 10th International Conference on Sport Sciences Research and Technology Support

Figure 6: Waist yaw angle of each participant (Tukey method *: p<0.05 **: p<0.01).

Figure 7: Racket speed of each participant (Tukey method *: p<0.05 **: p<0.01).

Training with a Pneumatic Assist Suit to Generate Lower-Body Twisting during the Forehand Swing in Table Tennis

5 DISCUSSIONS

As shown in Figure 4 and 5, the analysis using the

Bonferroni method hardly showed significant change

across conditions. Therefore, Figure 6 and 7, which

shows the results of the Tukey test, is the subject of

the discussion.

5.1 Effect of Wearing the Assist Suit on

the Swing

The effect of wearing an assist suit on the swing is

examined by comparing the results of "No wear

(before)" and "Without assist". In all participants

except participant 6, there was no significant

difference in the amount of waist rotation between

"No wear (before)" and "Without assist". Therefore,

it can be said that wearing the assist suit had less

restriction on the lower body movement.

In addition, in participants 1 and 6, the racket

speed was significantly decreased in the comparison

between "Without assist" and "No wear (before)".

About this, participant 1 said there was a feeling that

the bag fixed on the back was misaligned. Based on

this opinion, in each experiment of participants 2-6,

the bag fixation was strengthened. In the result, a

significant decrease was not observed except

participant 6. Therefore, when the bag fixation was

strengthened, it is considered that wearing the assist

suit had little restriction of the upper body movement.

5.2 Support Effect for Waist Yaw

Angle

The support effect of the assist suit is examined by

comparing "With assist" and "No Wear (before)",

"With assist" and "Without assist".

In all participants except participant 6, the amount

of rotation of the waist increased significantly in the

comparison between "With assist" and "No wear

(before)". In addition, in all participants, the amount

of rotation of the waist increased significantly in the

comparison between "With assist" and "Without

assist". From this, it can be said that the purpose of

increasing the rotation of the lower body was

achieved with the developed suit.

Only in participant 6, there was no significant

difference in the amount of rotation of the waist with

the comparison between "With assist" and "No wear

(before)". In this regard, in participant 6, the amount

of rotation of the waist reduced in "No wear (after)".

So, there is a possibility that participant 6 generated

the opposite force against the support force of the

device.

5.3 Training Effect of Assist Suit

The training effect of the assist suit is examined by

comparing "No wear (before)" and "No wear (after)".

In the results of the comparison of waist rotation

amounts, there is no significant difference in

participants 1 and 4, but a significant increase in

participants 2 and 5, and a significant decrease in

participants 3 and 6.

In the results of the comparison of racket speed,

there is a significant decrease in participants 1 and 4,

a significant increase in participants 2 and 5, and no

significant difference in participants 3 and 6.

From the above, it can be said that the

improvement of the swing ability by training with the

assist suit was observed in participants 2 and 5, but

not observed in participants 1, 3, 4, and 6.

To identify whether there are any characteristics

specific to those who have benefited from the suit, the

difference between participants 2, 5 and other

participants is discussed. On considering about

difference, participant 1 whose bag on his back was

fixed weakly than other participants and participant 6

who was not able to get support with the assist suit

are excluded. In the comparison of the waist rotation

amounts of "Without assist" and "With assist", there

are significant increases in all four participants.

However, in the racket speed, there are significant

increases in participants 2 and 5. Therefore, it is

considered that participants 2 and 5 were able to

transmit the waist increase movement to the

racket.From the above, there is the possibility that this

assist suit gives a great effect to participants who are

able to coordinate the twisting motion of the waist and

the movement of the racket well.

Figure 8 shows each average of waist yaw angular

velocity in "No wear (before)" and "No wear (after)".

In participants 2 and 5 who had a training effect,

Waist yaw angular velocities of both "No wear

(before)" and "No wear (after)" were large. The same

character is also seen in participant 1. So, there is a

possibility that participant 1 got a positive training

effect if the method of fixing the bag was the same as

that of other participants. On the other hand, in

participants 1, 3 and 6, waist yaw angular velocities

of both "No wear (before)" and "No wear (after)"

were small. From this, training effect may be seen in

participants with large waist yaw angular velocities

before training. Also, increases of waist yaw angular

velocities is not seen for participants with small waist

yaw angular velocities before training.

From the above, there is a possibility that the

developed assist suit is effective by introducing a

mechanism which supports transmitting the increased

icSPORTS 2022 - 10th International Conference on Sport Sciences Research and Technology Support

movement of the lower body to the racket movement

and reproducing the movement with large waist yaw

angular velocity without the suit after training. In

addition, the total training time is short since

participants hit the ball 20 times for each condition.

So, it is necessary to confirm whether the training

effect will be produced even in participants with

small waist yaw angular velocity if the number and

duration of training increase.

Figure 8: Relation of the average of waist yaw angular

velocity between "No wear (before)" and "No wear (after)"

in all participants 1-6. The plots over the diagonal line

indicate increases in waist yaw angular velocities from "No

wear (before)" to "No wear (after)".

6 CONCLUSIONS

In this study, the assist suit that assists the twisting

exercise of the lower body during the forehand swing

of table tennis was developed. Then, the assistive and

training effect for table tennis beginners were verified.

As a result, it was found that the developed assist suit

can increase the twisting motion of the lower body

and there are individual differences in the

improvement of swing ability by training with the

assist suit.

In the future, the number of participants will be

increased and the changes in experimental conditions,

such as long-term training and expansion of the age

range of participants, will be conducted to evaluate

the performance of the suit more accurately.

Furthermore, it is necessary to develop a mechanism

for the wearer to be able to link the movements of the

lower and upper body.

ACKNOWLEDGEMENTS

This research was supported by JSPS Grants-in-Aid

for Scientific Research JP21K17830.

REFERENCES

Wulf, G., & Shea, C. (2002). Principles derived from the

study of simple skills do not generalize to complex skill

learning. Psychonomic Bulletin & Review. 19-4, 185-

211.

Evans, K., & Tuttle, N. (2015). Improving performance in

golf: current research and implications from a clinical

perspective. Braz J Phys Ther. 19-5, 381-389.

Gordon, B., & Dapena, J. (2006). Contributions of joint

rotations to racquet speed in the tennis serve. Journal of

Sports Sciences, 24-1, 31-49.

Qian, J., Zhang, Y., Baker, J., & Gu, Y. (2016). Effects of

performance level on lower limb kinematics during

table tennis forehand loop. Acta Bioeng. Biomech. 18,

149–155.

Zhang, Z. (2017). Biomechanical analysis and model

development applied to table tennis forehand strokes.

[Doctoral thesis, Nanyang Technological University]

David, R., James, P. (2009). Can Robots Help the Learning

of Skilled Actions?. Exercise and Sport Sciences

Reviews. 23, 1–7.

Huang, F., Patton, J., & Mussa-Ivaldi, F. (2007). Interactive

priming enhanced by negative damping aids learning of

an object manipulation task. Annu. Int. Conf. IEEE Eng.

Med. Biol. - Proc, C, 4011–4014.

Kümmel, J., Kramer, A., & Gruber, M. (2014). Robotic

guidance induces long-lasting changes in the movement

pattern of a novel sport-specific motor task. Hum. Mov.

Sci., Hum. Mov. Sci, 38, 23–33.

Hirata, Y., Shirai, R., & Kosuge, K. (2017). Position and

orientation control of passive wire-driven motion

support system using servo brakes. IEEE Int. Conf.

Robot, 3702–3707.

Miyazaki, T., Kawase, T., Kanno, T., Sogabe, M.,

Nakajima, Y., & Kawashima, K. (2021). Running

Motion Assistance Using a Soft Gait-Assistive Suit and

Its Experimental Validation. IEEE Access, 9, 94700–

94713.

Zhou, T., Xiong, C., Zhang, J., Hu, D., Chen, W., & Huang,

X. (2021). Reducing the metabolic energy of walking

and running using an unpowered hip exoskeleton. J.

Neuroeng. Rehabil. 18, 1–15.

Klein, J., Spencer, S., & Reinkensmeyer, D. (2012).

Breaking it down is better: Haptic decomposition of

complex movements aids in robot-assisted motor

learning. IEEE Trans. Neural Syst. Rehabil. Eng, 20,

268–275.

Sakoda, W., Ramirez, A., Ogawa, K., Tsuji, T., & Kurita,

Y. (2018). Reinforced suit using low pressure driven

artificial muscles for baseball bat swing. ACM Int. Conf.

Proceeding Ser, 8–9.

Training with a Pneumatic Assist Suit to Generate Lower-Body Twisting during the Forehand Swing in Table Tennis