Design of Health Advice System for Elderly People
by Communication Robot
Takahito Tamai
1
, Yukie Majima
1
, Hideki Suga
2
, Syuuki Inoue
2
and Kotoka Murashima
1
1
Graduate School of Humanities and Sustainable System Sciences, Osaka Prefecture University, Japan
2
Kansai Electric Power Company Research and Development Department R&D Center, Japan
Keywords: Elderly People, Communication Robot, RoBoHoN, Health Advice System.
Abstract: Recently in Japan, as the declining birth rate and the aging of the population progress, increased medical
expenditures and nursing care burdens are presenting great social difficulty. To mitigate that difficulty it is
necessary for many elderly people to live independently and healthily. Therefore, this study develops a
health advice system to support health promotion for elderly people. We conducted verification experiments
to assess the operability and impressions of tablet PCs and robots and to elicit impressions against health
advice presented from different devices. Furthermore, we examined the usefulness of the health advice
system for elderly people who used the robots.
1 INTRODUCTION
Recently, in Japan, declining birth rate and aging of
the population are progressing. It is estimated that
1.2 people of 20–64 years will be compelled to
support one elderly person aged 65 or older in 2050
(Cabinet Office, 2016). Increased medical expenses
and nursing care burdens are cited as looming
difficulties (Ministry of Health, 2016). To mitigate
those and related problems will require not merely
life span extension, it will be necessary for many
elderly people to stay independent for as long as
possible and to have daily health care. Therefore,
this research is aimed at developing a health advice
system that monitors the daily health condition of
elderly people and which promotes light exercise.
In terms of operability, an earlier study (Suga et
al. 2016) showed that elderly people had difficulty
tapping a tablet PC. Recently, voice guidance for
operations of general products is used (Narita et al.
2011). Therefore, we developed a system that can be
operated by using a robot. Because a robot is
presented the possibility that as an information
presentation terminal suitable for use by elderly
people, a robot is more effective than a tablet PC in
previous study (Inoue et al. 2014) (Nihei et al. 2013).
The novelty of this research is verifying the
usefulness of a system that utilized a robot by
comparing impressions for health advice presented
from different devices by the same voice operation
method. We conducted comparative experiments to
verify the operability and impressions of tablet PCs
and robots, and impressions of health advice
presented by devices. We also assessed the health
advice system used by elderly people who used the
robots.
2 HEALTH ADVICE SYSTEM
2.1 Communication Robot
In recent years, robot development has advanced
according to their use in various roles. Humanoid
type robots are based mainly on motion mechanisms
aimed at supporting people's lives. Pet type robots
resembling dogs have been designed to heal people.
Research assessing utilization of robots and dialog
agents aimed at reducing burdens on health care
workers and at preventing dementia is underway,
especially at medical sites (Ibuki et al. 2005) ( Ono
et al. 2015).
2.2 Utilization of RoBoHoN
We adopted RoBoHoN (Sharp Corp.), a mobile
robot phone, because of its high speech recognition
rate. RoBoHoN, with gender set as male, is a small
486
Tamai, T., Majima, Y., Suga, H., Inoue, S. and Murashima, K.
Design of Health Advice System for Elderly People by Communication Robot.
DOI: 10.5220/0006637904860491
In Proceedings of the 11th International Joint Conference on Biomedical Engineering Systems and Technologies (BIOSTEC 2018) - Volume 5: HEALTHINF, pages 486-491
ISBN: 978-989-758-281-3
Copyright © 2018 by SCITEPRESS – Science and Technology Publications, Lda. All r ights reserved
humanoid robot (20 cm height, 400 g weight) that is
easy for elderly people to carry and easy for elderly
people to use easily. Cameras, GPS, projectors, etc.
are mounted as other functions (Fig. 1).
Figure 1: Appearance of RoBoHoN.
3 APPLICATION
DEVELOPMENT
3.1 Application Overview
The overall flow of the developed application is
presented in Fig. 2. The difference of the application
between the tablet PC and the robot is whether or
not tapping is necessary before inputting vital sighs
data (blood pressure, pulse rate, body temperature)
by voice. In the tablet PC, some operations are
required to tap the buttons on the screen before the
putting by voice, while the robot is designed to
operate all by voice. Now, we show the procedure of
the application. First, (1) the user inputs the
measured vital sighs data by voice. For example, the
voice guidance is "What was your body temperature
today? Please enter by voice." Next, (2) the physical
data for the past 30 days are displayed as a graph.
(3) The physical condition is judged from the
inputted data. Based on past data, health advice such
as meal guidance and promotion of outdoor
activities is provided to the user. For instance, when
the blood pressure value is high compared to past
data, more rest and lower salt intake are advised.
When one’s health is good, the robot proposes
places to go and encourages daily exercise habits.
Figure 2: Overall flow of developed application.
3.2 Physical Condition Determination
and Health Advice
The health state is ascertained by comparing the
measured physical data of the day with the average
value of the prior 30 days. The physical data of
elderly people have the following features (The
Japanese Society of Hypertension 2009) (Arita et al.
2016).
(1) Many people have high blood pressure
(2) The pulse rate fluctuates greatly, but lower
values are better
(3) Elderly people have higher than normal body
temperature, but have difficulty noticing it
themselves
We set criteria based on these considerations (Table
1). We chose the physical condition level (Table 2)
according to the number of items to be satisfied.
Then we gave health advice.
If the physical condition level is good, then people
are advised to do outdoor activities. At that time,
advice is based on weather information. For
example, when the weather is good, proper
hydration is encouraged ("Outing advice (A)" in Fig.
2) and outdoor activities are proposed such as park
visits ("Outside destination proposal (A)" in Fig. 2).
By contrast, when the weather is rainy, temperature
Design of Health Advice System for Elderly People by Communication Robot
487
regulation is advised ("Outing advice (B)" in Fig. 2)
along with staying indoors and the like ("Outside
destination proposal (B)" in Fig. 2). Moreover, in an
earlier study [3], as information necessary for
elderly people to venture outside, we described the
presence or absence of toilets on the go and barrier-
free information, among others. However, if a
person’s physical condition is bad, then advice is
given in light of the physical condition such as
refraining from outdoor activities (Table 3).
Table 1: Criteria for physical condition.
Type
Conditions
Good
Bad
blood
pressure
Difference
from the
average value
over the past
30 days is less
than 10
mmHg
Difference
from the
average value
over the past
30 days is 10
mmHg or more
pulse
Difference
from the
average value
over the last
30 days is less
than 10 beats
per minute
Difference
from the
average value
over the last 30
days is 10 and
more beats per
minute
body
temperature
Difference
from the
average value
over the past
30 days is less
than 1°C
Difference
from the
average value
over the past
30 days is 1°C
or more
Table 2: Physical condition judgment I.
Conditions
Physical
condition
Blood pressure · pulse rate · body
temperature is good 3
3
good
Blood pressure · pulse rate · body
temperature are good 2
2
Blood pressure · pulse rate · body
temperature are good 1
1
bad
Blood pressure · pulse rate · body
temperature are all bad
0
Table 3: Physical condition judgment II.
Advice
· Be careful not to catch a cold sweat.
· Wear a mask because it is dry.
· Refrain from going out for a long
time.
· Devote attention to salt intake.
(e.g., when blood pressure is high)
3.3 Implementation of Robohon
In application development, we create a graph-based
conversation scenario using software (yEd graph
Editor; yWorks Co.). This graphical conversation
scenario is implemented in RoBoHoN by conversion
it into an hvml file using tools created in Python
published by Sharp Corp. and running in Android
Studio™.
Creating a conversation scenario for RoBoHoN
differs from the original flow chart format. Because
the graph notation method is special, a method to
modify the sample scenario (Sharp Corp.) and add a
necessary graph was developed. A conversation
scenario development example is shown in Fig. 3.
As described earlier, RoBoHoN has male gender,
which is reflected in conversation. Compared with
tablet PCs and robots without gender, it gives a more
human-friendly and familiar impression. In addition,
RoBoHoN has many words and remarks that are
pre-set, assuming responses in advance. By creating
scenarios that use them, flexible and natural
conversation can be conducted (Table 4).
Figure 3: Conversation scenario development.
Table 4: Examples of words to which RoBoHoN reacts.
Text example
Positive reply
Yes, good, ok, that's right,
okay, Please do,
Negative reply
No, I can not, I do not want to,
do not do, I do not need it,
I will stop it, Different
HEALTHINF 2018 - 11th International Conference on Health Informatics
488
4 PRACTICE
4.1 Evaluation Method
To assess the effectiveness in the developed system,
we recorded the experiences of 10 men and women
in their 50s to 80s who agreed to cooperate. We
compared their questionnaire and interview
responses to those obtained using tablet PCs
equipped with a similarly developed application.
The experiment conditions are portrayed in Fig. 4.
The questionnaire was evaluated in four stages: 1.
Good; 2. Somewhat good; 3. Somewhat bad; 4. Bad.
Practices and timing are shown in Table 5. In this
experiment, to present advice to promote outdoor
activities when physical condition shows a change,
the subject's response is set in advance: "when there
is no physical schedule level 3 and no going out"
and "there is no physical condition level 0 and no
going out case" in the two scenarios were verified.
This step eliminates differences among subjects
according to the contents of presentation. Evaluation
items for the system are presented in Table 6.
This research was conducted with the approval of
the Research Ethics Committee of Osaka Prefecture
University.
Figure 4: Experimental situation.
Table 5: Practice target and practical place · timing.
Target
10 adults in their 50s to 80s
(6 men, 4 women)
Time
January 30, 2017 to February 1
Table 6: System evaluation item.
Voice
operation
Was it easier to operate by voice?
Impressions
Did you want to use it in the future?
Health
advice
Did you want to follow the suggested
health advice?
4.2 Results
The average value of the questionnaire result is
presented in Table 7. The speech manipulation was
highly evaluated overall, but women evaluated it
more highly than men did. Additionally, high
evaluations were given for impressions of robots
from participants of all ages.
Table 7: Questionnaire results.
Attribute
Overall
Men
Women
Number of people
10
6
4
Voice
operation
Tablet
PC
3.1
2.7
3.8
Robot
3.5
3.3
3.8
Impressions
Tablet
PC
3.4
3.2
3.8
Robot
3.7
3.7
3.8
Health
advice
Tablet
PC
3.5
3.3
3.8
Robot
3.4
3.0
4.0
Evaluation values are average (4.0 is the maximum)
5 DISCUSSION
Regarding operability by voice, the tablet PC was
highly evaluated overall with 3.1 out of 4, robot 3.5.
However, by sex, evaluation was 3.8 among women,
but 2.7 among men. According to an interview
survey, many men reported that they used electronic
devices such as smartphones and tablet PCs on a
daily basis. For those who are accustomed to tap
operation, an inconvenience arose by which it can
only be input after waiting for the voice guidance
because it cannot be advanced at their own pace.
However, one opinion holds that it is easier than the
tap operation for a person who is not familiar with
the tap operation. A system must be developed that
allows a user to select the waiting time and guidance
setting until input. Additionally, an opinion arose
that the timing to address the robot is difficult
because there are times when the robot reaction is
slow and conversations cannot be done at a certain
speed. Future improvements can be considered, such
as shortening words to be recognized, thereby
hastening the system operations.
As for impressions, the robot was given a very
high rating of 3.7 overall. Reasons such as robot's
"good appearance" and "good character" were
described. RoBoHoN is small and can be picked up.
Design of Health Advice System for Elderly People by Communication Robot
489
Because it speaks in first person voice, calling itself
"boku," it might be perceived as a small boy, such as
a child or a grandchild. The robot was therefore very
accessible, with a friendly feel. Moreover, ambitious
opinions were found related to the use of the
developed system because "conversation is not
working but fun." Responses of the tablet PC were
monotonous responses such as "yes–no", whereas in
the dialogue with the robot, responses were natural
and conversational.
However, regarding the presented health advice,
the tablet PC was evaluated as 3.5 vs. the robot
evaluation of 3.4. Whereas a concrete example can
be shown on the tablet PC, the robot can present
only sound. For example, with meal advice, when
presenting advice such as "Devote attention to salt
intake," the tablet PC shows the amount of salt
contained in food in the table.
However, evaluation of health advice from
robots was high for women, probably because of the
fact that the appearance is cute, so attachment is
easy. It was easy to attract interest. To give effective
health advice, we must consider visual as well as
audio presentation using the display on the back and
the projector function of RoBoHoN. Furthermore,
regarding the health advice contents, in addition to
proposals made outside, we would like to examine
other points that were raised: "proposals for exercise
for promoting health," "knowledge related to
diseases," and "how to respond when symptoms of
onset illness occur." Health advice must incorporate
the age and sex of the user and the state of the
individual.
This study used only conversation with
RoBoHoN on the desk, but contact interaction with
the robot might be effective for communication
(Kanda et al. 2003) (Ogawa et al. 2011) (Nakagawa
et al. 2012), so construction of a system to do them
is also listed as a future task. In addition, this time,
the subjects who consented to cooperate were few.
Statistical analysis was difficult, only the evaluation
by the average could be done. Therefore, it is a
future subject to increase the number of subjects and
conduct experiments.
The description presented above indicates that
the health advice system using the robot is useful. It
increases ones willingness to use it because of
enjoyment of the conversation so that they talk with
humans than speaking to a tablet PC, and the
character nature compared. People are likely to
listen to the health advice. Furthermore, since the
evaluation was based on short interaction in this
experiment, we would like to evaluate by long-term
interaction similar to the actual situation in the
future.
6 CONCLUSION
For this study, we developed a health advice system
that uses a robot operated by voice. Elderly people
can use it easily for promoting health management
and light exercise (outdoors). Furthermore, this
study compared the system with a tablet PC based
system.
Speech manipulation was easy for those who
were not accustomed to conventional tap
manipulation. Evaluations were high, but it was
reportedly difficult to use because it responded at its
own pace. Therefore, it is necessary to design a
system that can arbitrarily change according to user
characteristics. Furthermore, flexible conversation
capability that made use of the robot character is
apparently fun because users can human
communicate without monotonous input work and
users tend to heed the suggested health advice. The
system using a robot was demonstrated to be useful.
We plan to improve the system using of the
RoBoHoN’s functions of camera and projector to
promote behaviours modification.
REFERENCES
Arita, K. Ishii, T. Imai, H. et al. 2016. Basic nursing skills
for systematic nursing skills, Igaku Shoin, 140-147
Cabinet Office, Japan, Annual Report on the Aging
Society [Summary] FY 2016
Ibuki, S. et al. 2005. Living Assistance System by
Communication Robot for Elderly People, Journal of
the Society of Mechanical Engineers, Vol. 108 No.
1038
Inoue, T. 2014. The development of a robotic support
system for the day-to-day needs of elderly people with
memory and cognitive decline, Transactions of the
Virtual Reality Society of Japan. Vol. 19, No. 3
Kanda, T. Imai, M. Ono, T. et al. 2003. Numerical
Analysis of Body Movement on Human-Robot
Interaction, Transactions of Information Processing
Society of Japan, Vol. 44, No. 11, pp. 2699-2709
Ministry of Health, Labour and Welfare. 2016. Overview
of national medical expenses
Nakagawa, K. Shiomi, M. Shinozawa, K. et al. 2012. Does
Robot’s Active Touch Improve People’s Motivation? ,
IEICE Transactions on Fundamentals of Electronics,
Communications and Computer Sciences, Vol. J95–A,
No. 1, 136-145
Narita, T., Nihei, M., Shino, M. et al. 2011. A Study of
Ease of Comprehension of Synthesized Speech for the
HEALTHINF 2018 - 11th International Conference on Health Informatics
490
Elderly, Assistive Technology, Bio Medical
Engineering and Life Support, O1-3
Nihei, M. et al. 2013. Comparing Robots with Robots
Agents as Information-Support Systems for the Elderly
and Those with Mild Dementia, Assistive Technology
Research Series, 33, Assistive Technology: From
Research to Practice, 880-886
Ogawa, T. Morita, K. Fuketa, M. et al. 2011. A study of a
dialogue robot with haptic interaction, Forum on
Information Technology, 2011, 547-548
Ono, S. Woo, J. Matsuo, Y. et al. 2015. A Health
Promotion Support System for Increasing Motivation
Using a Robot Partner, Transactions of the Institute of
Systems, Control and Information Engineers, Vol. 28,
No. 4, 161-171
Suga, H., Majima, Y., Takahashi, H. 2016. Development of
HEMS Equipment for Senior Citizen with Function of
Security and Safety (Part 2) Senior Citizen's Go Out
Promotion and the Functional Verification Test Results
of HEMS In Summer, The Institute of Electronics,
Information and Communication Engineers Technical
Report, MICT2016-49(2016-11), 7-11
The Japanese Society of Hypertension, 2009,
Hypertension treatment guidelines 2009，34～44
Design of Health Advice System for Elderly People by Communication Robot
491