Psychological Effect of Robot Interruption in Game
Mitsuharu Matsumoto
1
and Hiroyuki Yasuda
2
1
Department of Informatics, The University of Electro-communications, 1-5-1, Chofugaoka, Chofu-shi, Tokyo
2
Department of Human-communications, The University of Electro-communications, 1-5-1, Chofugaoka, Chofu-shi, Tokyo
Keywords: Human-robot Interaction, Psychological Study, Effect of Robot Interruption.
Abstract: In this paper, we report psychological effect of robot interruption on human. Although many robots are
developed to help people in daily life, such robots sometimes make users live a reactive life. On the contrary,
some researchers developed robots that depend on users. These types of robots require users’ assists to do
their tasks and users need to be active due to its dependence like children. Children not only require our help
to do their tasks but also interrupt us. In spite of their interruption, people come to like children and would
like to interact with children. To achieve long-term interaction between human and robot, we expect that
adequate interruption to users may have some merits rather than helping users at all times. To investigate
our hypothesis, we developed two types of robot and designed a simple game with the robots. Throughout
the experiments, users have stronger motivation to interact with robot that interrupted users than the robot
that did not interrupt them.
1 INTRODUCTION
Recently, a lot of robots are developed to support
our daily life such as housework and navigation
(Hiroi et al., 2003, Iwata and Sugano, 2009,
Matsuyama et al., 2010). Although many types of
robots have been developed, they have silimar
features, i.e., they are basically designed to do their
tasks perfectly without errors to help users.
However, in psychological area, some reports
describe that unlimited supports from humans and
robots have disadvantage rather than advantage. For
example, it is remarkable that physical and mental
decays of elderly people proceed rapidly when they
lost something to do and live a bleak life after
reaching the mandatory retirement age (Iguchi, 2002,
Takanaka et al., 2005). According to other
researchers, people who require nursing care feel
strong stress even if they receive good skilled care
because they feel loss of independence due to their
passive lives (Tanaka et al., 1997).
Some researchers noted the importance of works,
and recommend people to have some tasks to be
done in daily life. These types of tasks give people a
purpose in life (Kamiya, 2004).
Based on the above prospects, some authors have
developed the robots that require humans’ assists.
These robots cannot work without users’ help unlike
typical life supporting robots.
For example, a baby-like robot named Babyloid
was developed by Kanoh et al (Kanoh and Shimizu,
2011). It cannot do anything without users' help like
baby and wait users’ help by doing some actions
such as changing its facial expression and crying.
They expect that users intend to have active
interaction with Babyloid due to its ineffectuality.
Yoshida et al. have developed a trash box type
robot named Sociable Trash Box (Yoshida et al.,
2009). Although Social Trash Box can move and
bow its body, it cannot take garbage. Social Trash
Box only bows when users pick up the garbage and
dump it to Social Trash Box. Social Trash Box
requires users’ help and aim to encourage users’
active support by using the robot ineffectuality.
They labeled this concept power of weakness and
confirmed its availability.
We also developed a robot that sometimes makes
mistakes (Yasuda and Matsumoto, 2013). Although
it is considered that mistakes are not good for robot
design and should be avoided in common,
experimental results showed that some mistakes
cause users affection and interests to the robot.
However, it was still short-term interaction and more
study on long-term interaction should be done.
In this paper, we designed a simple game, and
developed two types of robots to investigate the
effect of the robot interruption to longer-term
133
Matsumoto M. and Yasuda H..
Psychological Effect of Robot Interruption in Game.
DOI: 10.5220/0004951001330137
In Proceedings of the 16th International Conference on Enterprise Information Systems (ICEIS-2014), pages 133-137
ISBN: 978-989-758-029-1
Copyright
c
2014 SCITEPRESS (Science and Technology Publications, Lda.)
interaction.
We expect that people live more active life not
by being helped by the robots any time but by
sometimes helping the robots.
In the next section, we summarize the
characteristics of the robot that sometimes interrupts
users, and its advantage compared to the typical
supporting robots. In Sec. III, we describe the
specification of our developed robots, and report
some experimental results of field experiments for
common people in our university. Discussion and
conclusion follow in Sec. IV.
2 CONCEPT OF
MISTAKE-BASED
INTERACTION
In our research, we propose the concept of mistake-
based interaction, that is, the interaction between
human and the incomplete robot that sometimes
interrupts users unlike typical robots that aim to
support human perfectly. In this section, we describe
the features of incomplete robots, and summarize the
aim of this approach.
To clarify the characteristics of the proposed
concept, we give two figures. Figure 1 shows
interaction between a user and a typical supporting
robot. As shown in Fig.1, the existing robot tries to
support users. The user expects the robot to do tasks
perfectly and robot is not allowed to make any
mistakes. Although this approach is very useful, as a
result, it may set the user passive position in their
interaction.
On the other hand, in mistake-based interaction,
the robot sometimes makes mistakes and interrupt
users as shown in Fig.2. In this approach, a user
needs to make active interaction from him/her to the
robot by its mistakes, and to behave as a caretaker
for the robot in some cases. Babies and infants seek
to do all things that they can do, whereas they bring
trouble to adults when they fail in things that they do.
Adults become more attached to them due to their
efforts and failures. In a similar fashion, we expect
that the robots become more human-like and users
may feel more humaneness and affection on it when
the robot is designed to make mistakes in some cases.
Figure 3 shows the orientation of this approach
compared to the other approaches.
Our aim is to encourage users to interact with the
robot actively by increasing the robot’s
incompleteness as shown in Fig.3.
Figure 1: Interaction between users and existing robots.
Figure 2: Interaction between users and incomplete robots.
Figure 3: Comparison of our approach with other
approaches.
3 EXPERIMENTS
3.1 Experimental Contents
The degree of interruption changes depending on the
tasks that users ask. The range of users’ allowance
on mistakes also varies depending on users. In this
study, we designed a game "ball in the basket".
ICEIS2014-16thInternationalConferenceonEnterpriseInformationSystems
134
Figure 4: Prepared robots (Rowdy and Goodboy).
Figure 5: Experimental setup.
Figure 6: Appearance of experimental environment.
The rule of ball in the basket is simple. Users
tried to put some balls into the basket on the robot
within given area. The limited time was 30 seconds.
We asked users to chase the robot and put balls into
the basket as possible.
We developed two types of robots.
One robot is Rowdy shown in the right in Fig.4.
It moves randomly and acts up to spill the ball out.
As the basket is fixed with movable stick, the balls
are spilled during the robot movement.
The other is Goodboy shown in the left in Fig.4.
It just moves randomly. As the basket was fixed, the
robot does not spilled balls.
As the robots' functions are different, maximal
balls are 12 in Rowdy, while are 6 in Goodboy.
We set the following three rules for the game.
1. Users are not allowed to catch or stop the robot.
2. Users should put the ball to the robot one by one
and should not throw it to the robot.
3. Users should continue to put the ball into the
basket during the game.
The robots were developed based on BeautoRover
(Vstone Co., Ltd.). They can be controlled through
remote computer via bluetooth connection.
We developed some programs to control the
robot. When we push the button on the programs,
the robot starts to move. The robot action was the
combination of "Move forward", "Move backward",
"Turn right", "Turn left" and "Stop". The actions
were selected randomly to run away from users.
To focus on the effect of robots' mistake, we did
not implement speech function to the robot, and
gave similar appearance to Rowdy and Goodboy.
We told subjects basic functions of them before the
experiments.
3.2 Experimental Condition
Figure 5 shows the experimental setup. As shown in
Fig.5, a coordinator joined the experiments. He
explained users about the game and controlled the
robot during the game. We demonstrated the
experiment as a part of our laboratory’s introduction
of open campus in our university. Visitors to our
laboratory were subjects in the experiments. 11
visitors joined the experiments. Figure 6 shows the
appearance of the experimental environment. The
outline of the experiment is summarized as follows:
We introduced the robot to subjects at our
laboratory and asked them to answer a questionnaire
about the robot behaviors after obtaining the
agreement from them. The detailed scenario of the
experiment and contents of the questionnaire are
summarized as follows:
Figure 7: Experimental result on Q1.
PsychologicalEffectofRobotInterruptioninGame
135
Figure 8: Experimental result on Q2.
Figure 9: Experimental result on Q3.
Figure 10: Experimental result on Q4.
** Scenario
1. Subjects visited our laboratory room.
2. The experimenter introduced the robot to
subjects and asked them to play with robots
when he explained about our laboratory.
Figure 11: Experimental result on Q5.
3. The experiment started the game and users
played with Rowdy and Goodboy.
4. The experimenter asked users to answer the
questionaires.
The contents of the questionnaire are as follows:
We ask you about your impression on the robots and
game.
- We first ask users the impression on the game.
Q1: How did you feel the game with Goodboy?
Q2: How did you feel the game with Rowdy?
Q3: Do you want to play the game again?
- We second ask users the impression on the robot.
Q4: How did you feel the robot, Goodboy?
Q5: How did you feel the robot, Rowdy?
The answer’s method of the questionnaire has 7-
point scale except Q3. The following is grading on
the answer.
-3: Unpleasant, 0: Neutral, 3: Pleasant
The answer of Q3 is as follows:
I would like play again with
1. Goodboy, 2. Rowdy, 3. Both, 4. Neither
3.3 Experimental Results
Figures 7, 8, 9, 10 and 11 show the summaries of
subjects' impression on the robot for Q1, Q2, Q3, Q4,
and Q5, respectively. Let us summarize the results
of questionnaire in order. In the answer for Q1, no
one felt neutral impression. About half of subjects (6
people) felt positive impression on Goodboy and
remains (5 people) felt negative impression as
shown in Fig.6. On the other hand, in the answer for
Q2, many people (9 people) felt positive impression
on Rowdy, while few people felt neutral and
negative impression.
Although Rowdy acted to interrupt users action
compared to Goodboy, users felt more positive
impression on the game with Rowdy than that with
Goodboy.
As shown in Fig.9, half of users would like to
play again with only Rowdy, and 3 users would like
to play again with both robots.
We think that these results gave us important
sights in human-robot interaction.
4 DISCUSSION AND
CONCLUSIONS
In this paper, we developed two types of robots,
Rowdy and Goodboy and investigated psychological
effect of interruption during games to users. For
achieving medium-term interaction, we designed a
ICEIS2014-16thInternationalConferenceonEnterpriseInformationSystems
136
simple game and evaluate the effect of interruption-
based interaction on medium term interaction.
Although the game was simple, we think that the
results give some insight about human-robot
interaction.
The system designer normally designs the robot
to be positive impression to users for creating good
relationship between human robot. However, some
interruptions attract users and give some motivation
to users to continue interaction with robot according
to the results of the experiments.
We consider that long-term interaction is a key
factor for robot to be common in our daily life, and
some interruptions may become important factors
for it.
For future works, we would like to design more
sophisticated game, and investigate the factor for
long-term interaction. We also aim to study the
effect of the robot voice, the robot appearance to the
users impression. Growth of the robot during the
game should also be considered.
ACKNOWLEDGEMENTS
This work was supported by JSPS KAKENHI Grant
Number 25730107, 2013. This work was also
supported by Japan Prize Foundation, Foundation
for the Fusion of Science and Technology and NS
promotion foundation for science of perception.
REFERENCES
Hiroi, Y., Nakano, E. et al, 2003. A Patient Care Service
Robot System Based on a State Transition
Architecture. In The 2nd International Conference on
Mechatronics and Information Technology.
Iwata, H., S. Sugano, 2009. Design of Human Symbiotic
Robot TWENDY-ONE. In IEEE International
Conference on Robotics and Automation, IEEE.
Matsuyama, Y., Fujie, S., et al, 2010, Psychological
Evaluation of A Group Communication Activation
Robot in A Party Game, In Interspeech2010.
Iguchi, A., 2002. Incoming geratology, The University of
Nagoya Press.
Takenaka, H. et al., 2005, Loss experience and rebirth of
elderly people, Seitosha Publishing.
Kamiya, K., 2004. On a purpose of life, Misuzu Shobo.
Tanaka K et al., 1997, Geratology isagoge, Nihon hyoron
sha.
Kanoh, M., Shimizu, T., 2011. Developing a robot
Babyloid that cannot do anything, In J. of the Robotics
Society of Japan.
Yoshida, Y., Yoshiike, Y., et al., 2009. Sociable Trash
Box: A robot collecting trashes with children, In J. of
Human Interface Society.
Yasuda, H., Matsumoto, M., 2013. Psychological impact
on human when a robot makes mistakes, In 2013
IEEE/SICE International Symposium on System
Integration.
PsychologicalEffectofRobotInterruptioninGame
137
