COSEY: Computer Supported Enhancement of Young Children’s

Cooperation

Toward a Multiple-player Cooperative Full-body Interaction Game

Tsugunosuke Sakai

, Haruya Tamaki

, Ryuichi Yoshida

, Ryohei Egusa

2,3

, Shigenori Inagaki

Etsuji Yamaguchi

, Fusako Kusunoki

, Miki Namatame

Masanori Sugimoto

and Hiroshi Mizoguchi

Department of Mechanical Engineering, Tokyo University of Science, 2641 Yamazaki, Noda-shi, Chiba-ken, Japan

JSPS Research Fellow, Tokyo, Japan

Graduate School of Human Development and Environment, Kobe University, Hyogo, Japan

Department of Computing, Tama Art University, Tokyo, Japan

Tsukuba University of Technology, Ibaraki, Japan

Hokkaido University, Hokkaido, Japan

Keywords: Children, Cooperation, Full-body Interaction Game, Kinect Sensor.

Abstract: We have developed a full-body interaction game that allows children to cooperate and interact with other

children in small groups. The intent of the full-body interaction game developed in this study is to

encourage cooperation between children. The game requires two children to jump together with the same

timing. We let children experience the game, and observed the children using several strategies to

coordinate the timing of their jumps. These included shouting time, watching each other, and jumping in a

constant rhythm as if they were skipping rope. In this manner, we observed the children playing the game

while cooperating with each other. Therefore, the game enabled children to experience cooperation with

others by using body movements.

1 INTRODUCTION

The importance of play in early education is

recognized by education systems around the world

(Samuelsson, I. P., 2008). Studies show that as

children play, they create knowledge (Dau, E., 1999,

Levin, D., 1996). Children learn well by being active

(Samuelsson, I. P., 2008). Kinesthetic learning

facilitates cognitive development (Tscholl, M.,

2013), particularly for younger children (Engelkamp,

J., 1991, Goldin-Meadow, S. 2009). When children

use gestures and movements, the learning

environment becomes more natural (Grandhi, S. A.,

2011, Nielsen M., 2004, Villaroman, N., 2011) and

children are able to retain more of the knowledge

being taught (Edge, D., 2013, Antle, A., 2009).

Based on these research efforts, many full-body

interaction games have been developed as learning

support systems for children. These full-body

interaction games allow children to acquire

knowledge by interacting with the system though

body movements.

In addition to knowledge, sociality and

cooperation are important for children’s growth.

Sociality and cooperation are acquired from actual

experiences gained when interacting with others to

achieve some goal. Group size was the only factor to

significantly affect cooperation, with children in

small groups cooperating significantly more than

children in larger groups (Anuska, 2008). Gender

and subject matter had no effect on cooperation

(Anuska, 2008). However, although many full-body

interaction games have been developed as learning

support systems for children, very few of them allow

children to cooperate with others in small groups to

enhance cooperation skills. Previous learning

support systems for children that used full-body

interaction games tended to emphasize interaction

between the children and the system. There was

almost no emphasis on interaction between the

children.

Therefore, this study aims to develop a full-body

Sakai, T., Tamaki, H., Yoshida, R., Egusa, R., Inagaki, S., Yamaguchi, E., Kusunoki, F., Namatame, M., Sugimoto, M. and Mizoguchi, H.

COSEY: Computer Supported Enhancement of Young Children’s Cooperation - Toward a Multiple-player Cooperative Full-body Interaction Game.

In Proceedings of the 8th International Conference on Computer Supported Education (CSEDU 2016) - Volume 2, pages 175-180

ISBN: 978-989-758-179-3

175

interaction game that lets children interact with each

other. To that end, we developed a full-body

interaction game that allows children to cooperate

with others in small groups. Based on findings in

previous studies, the game is designed to enhance

cooperation between children by allowing them to

cooperate with others in small groups by using body

movements. As the first step in our study, we

developed a full-body interaction game that allows

two children to cooperate with each other. This

paper provides an overview of the current

implemented game, and describes an experiment in

which children were observed played the game.

2 CURRENT IMPLEMENTED

GAME

2.1 Concept

The game developed in this study enables children

to interact with each other. Figure 1 shows an

overview of the game. A certain device, a meter, and

lamps expressing the “electricity” of the device are

projected onto the screen. In this game, two players

are asked to cooperate to “make electricity.” The

device’s electricity is generated through the

cooperation of two players, who must jump together

with the same timing. Two players cannot make

electricity without adjusting their reciprocal timing.

Figure 1: Overview of the game.

CSEDU 2016 - 8th International Conference on Computer Supported Education

176

If two players jump together with the same timing,

the red meter on the left side of the screen increases.

When all lights on the red meter are lit, the lamp on

the top of the screen is also lit. When all lamps are

lit, electricity has been generated for the device and

the game finishes. To achieve this, two players must

jump together with the same timing; thus, it is likely

that the two players will communicate with each

other. For this reason, the game developed in this

study produces interaction between children.

2.2 Overview and Configuration of

System

Figure 2 shows an overview of the system. The

system consists of a screen, a short focus projector, a

notebook computer, and a Kinect sensor. The

notebook computer operates the Kinect sensor and

controls the game screen. These operations and

controls were implemented using a C# program that

we developed in Visual Studio 2013. The playing

area recognized by the sensor is set freely by the

program. The Kinect sensor is a range image sensor,

originally developed as a home video game device.

Although inexpensive, the sensor can capture

complex measurements that accurately represent the

user’s location. In addition, by using a library such

as KINECT for Windows SDK., this sensor can

recognize humans and the human skeleton. It can

measure the locations of human body parts such as

hands and legs. By using these functions, the Kinect

sensor can recognize the gestures of a person. Jump

gestures are used to progress the game developed in

this study. Figure 3 shows the configuration of the

system. The notebook computer processes jump

information recognized by the Kinect sensor. The

game screen, which is projected by a short focus

projector, changes according to the processing

results. Jump recognition achieved with the Kinect

sensor is discussed in the following section.

2.3 Jump Recognition by Kinect Sensor

2.3.1 Tools

Kinect Studio and Visual Gesture Builder are used

to recognize and process jumps captured by the

Kinect sensor.

Kinect Studio can record the 3D position

information of a human body captured with a Kinect

sensor. By using this tool, a player’s 3D position

information can be recorded when they make a

gesture that must be recognized by the system. This

data can be used to train the gesture reader in Visual

Gesture Builder, by means of a machine learning

process. The gestures the system needs to recognize

are recorded by Kinect Studio in a clip file, which is

used by Visual Gesture Builder as training data. In

this file, frames that contain gestures that must be

recognized by the system are marked. The gesture

reader learns the 3D position information of players

while building its gesture database after the frames

are marked. In this manner, the gesture reader can

learn the gestures that must be recognized in Visual

Gesture Builder. A program that recognizes gestures

captured by the Kinect sensor can be created by

accessing the gesture reader that was trained by

Visual Gesture Builder.

Figure 2: Overview of the system.

Figure 3: System configuration.

2.3.2 Recording Jumps of Children

The full-body interaction game developed in this

study is for children. Because of this, the training

data used for learning in Visual Gesture Builder

should be collected from children, not adults. The

Kinect sensor recognizes the jumps of children in

the game developed in this study. Therefore,

COSEY: Computer Supported Enhancement of Young Children’s Cooperation - Toward a Multiple-player Cooperative Full-body Interaction

Game

177

manners of jumping specific to children must be

recorded by Kinect Studio. A jump reader that can

recognize the jumps of children must then be built.

Figure 4 shows the flowchart for building a jump

reader that can recognize children’s jumps. First, we

recorded our own jumps in Kinect Studio and built a

simplified version of a jump reader in Visual

Gesture Builder. This reader cannot effectively

recognize the jumps of children. Second, a

simplified version of the game that can be played by

one player was developed. Third, we let children

play the simplified version of the game. On that

occasion, the children’s manners of jumping were

recorded in Kinect Studio. Thus, the training data

used for learning in Visual Gesture Builder were

collected from children. Finally, the jump reader was

Figure 4: Flowchart for building jump reader that can

recognize jumps of children.

Figure 5: Children’s manners of jumping are recorded.

rebuilt using the training data collected from

children. As a result, the jump reader was improved

to recognize the jumps of children. Figure 5 shows

the children’s manners of jumping being recorded.

During this process, we recorded the jumping

manners of 51 11- and 12-year-old children.

3 DEMONSTRATION

EXPERIMENT

3.1 Study Participants and

Environment

Twenty 6

-grade elementary school children played

the game developed in this study. The children were

arranged into 10 groups, with two players in each

group. Figures 6 and 7 show the demonstration

environment. The dimensions of the play space were

provided within the demonstration environment. We

told only one player in each pair how to play the

game. The other player waited outside the

demonstration room while his/her partner was told

how to play the game. The other player then entered

the room and was told how to play the game by

his/her partner. The game started after game playing

instructions were given from one partner to the other.

Game play was captured by a video camera.

Figure 6: Demonstration environment #1.

Figure 7: Demonstration environment #2.

3.2 Observing Children Playing the

Game

Figures 8, 9, and 10 show the children playing the

CSEDU 2016 - 8th International Conference on Computer Supported Education

178

game.

Figure 8 shows the children who were told how

to play the game telling their partners how to play

the game. The following explains were voiced:

 “We must jump with the same timing.”

 “If two players jump together, the meter

increase.”

 “Let’s jump after saying “One, Two, Three.””

In nine out of the 10 groups, the children who were

told how to play the game by their partner asked

their partner to confirm the rules of the game. In

these cases, we observed children assuring one

another about uncertain matters.

Figure 9 shows instances in which two children

were able to coordinate the timing of their jumps.

Instances in which two children coordinated their

jumping by watching each other were observed in all

groups. Further, we observed instances in which

children shouted cues to coordinate their timing.

These states were often observed in groups that were

previously unable to coordinate their jumps. In these

cases, we observed children trying to cooperate with

each other to coordinate their jumps. Further, one

group jumped in a constant rhythm, as if skipping

rope. This group could play the game more

efficiently.

Figure 10 shows instances in which two children

jumped together with the same timing. In this

experiment, children thought out ways to play the

game more efficiently. As these results show, we

developed a full-body interaction game that allowed

two children to cooperate with each other to play a

game.

4 CONCLUSIONS

We have developed a full-body interaction game that

allows children to cooperate with others in small

groups. Based on findings in previous studies, the

game was designed to enhance cooperation between

children by allowing them to experience cooperation

with others in small groups through body

Figure 8: Children who were told how to play the game first needed to tell their partners how to play the game.

Figure 9: Children strategizing to coordinate their jump timing.

Figure 10: Two children jumping together with the same timing.

COSEY: Computer Supported Enhancement of Young Children’s Cooperation - Toward a Multiple-player Cooperative Full-body Interaction

Game

179

movements. This paper provided an overview of the

current implemented game, and described the results

of experiments in which children actually played the

game. The game requires two children to jump

together with the same timing to achieve a goal. We

let children experience the game developed in this

study. As a result, the children used several

strategies to coordinate the timing of their jumps;

these included shouting time, watching each other,

and jumping in a constant rhythm as if they were

skipping rope. Thus, we observed children

cooperating with each other while playing the game

developed in this study. Therefore, the game let

children experience cooperation with each other

through body movements. In future work, we plan to

increase the number of players, in order to analyze

how players might cooperate with each other in

larger groups. Further, evaluation and improvement

will be performed to develop a game that allows

children to experience greater cooperation with

others.

ACKNOWLEDGEMENTS

This work was supported by JSPS KAKENHI

numbers 15H02936 and 26560129. The

demonstration experiment was supported by Kobe

University Elementary School.

REFERENCES

Samuelsson, I. P., & Carlsson, M. A., 2008. The Playing

Learning Child: Towards a pedagogy of early

childhood. Scandinavian Journal of Educational

Research, 52(6), pages 623-641.

Dau, E., & Jones, E., 1999. Child's Play: Revisiting Play

in Early Childhood Settings. Brookes Publishing,

Maple Press.

Levin, D., 1996. Endangered play, endangered

development: A constructivist view of the role of play

in development and learning. In A. Phillips (Ed.),

Topics in early childhood education 2: Playing for

keeps. St. Paul, MI: Inter-Institutional Early

Childhood Consortium, Redleaf Press.

Tscholl, M., Lindgren, R., & Johnson, E., 2013. .Enacting

orbits: refining the design of a full-body learning

simulation. Interaction Design and Children, pages

451-454.

Engelkamp, J. & Cohen, R.L., 1991. Current issues in

memory of action events. Psych. Research, 53, pages

175-182.

Goldin-Meadow, S., Cook, S. W., & Mitchell, Z. A., 2009.

Gesturing gives children new ideas about math.

Psychological Science, 20(3), pages 267-272.

Grandhi, S. A., Joue, G., Mittelberg, I., 2011.

Understanding naturalness and intuitiveness in gesture

production: insights for touchless gestural interfaces.

Proc. CHI 2011, pages 821-824.

Nielsen M., Störring, M., Moeslund, T., Granum E., 2004.

A procedure for developing intuitive and ergonomic

gesture interfaces for HCI. Gesture-Based

Communication in HCI, pages 105-106.

Villaroman, N., Rowe, D., Swan, B., 2011. Teaching

natural user interaction using OpenNI and the

Microsoft Kinect Sensor. SIGITE 2011, pages 227-232.

Edge, D., Cheng, K. Y., & Whitney, M., 2013.

SpatialEase: learning language through body motion.

CHI, pages 469-472..

Antle, A., Kynigos, C., Lyons, L., Marshall, P., Moher, T.,

and Roussou, M., 2009. Manifesting embodiment:

Designers’ variations on a theme. In Proceedings of

CSCL (Rhodes, Greece), Panel, pages 15-17.

Anuska, I. A., Juse, O. S., Maria, E. Y., 2008. Does group

size matter? Cheating and cooperation in Brazilian

school children, Evolution and Human Behavior,

pages 42-48.

CSEDU 2016 - 8th International Conference on Computer Supported Education

180