Partner Robots for Mental Health Care

parisfawnAI and Robotics

Nov 17, 2013 (3 years and 6 months ago)

59 views

Partner Robots for Mental Health Care


-

J a p a n e s e C h a l l e n g e

t o w a r d
P r a c t i c a l R o b o t s
-

G o r o

O b i n a t a



E c o T o p i a

S c i e n c e I n s t i t u t e, N a g o y a U n i v e r s i t y,
N a g o y a, A i c h i 4 6 4
-
8 6 0 3,
J a p a n

E
-
m a i l: o b i n a t a @ m e c h.n a g o y a
-
u.a c.j p

Japanese

Challenge in Service Robots




Paro

( seal robot )


Wakamaru


Muu


ifbot


Robot for tea service

One craftsman still does

his work on making

these kinds of robots.

AIBO


350,000
\

= 2,700 UK pounds


2.7

kg, 9 W


Speaker,

Microphones (3),


Sensors Tactile (12), Orientation (3D), Temperature,


Photo


Actuators (7) Eyelid, Head, Limbs


Operating time


1.5 hours

per one charge




good selling

(more than 500 Paros)


Paro

“Most successful

therapeutic robot”

by Dr. Shibata

http://paro.jp

Alternatives for dogs in animal therapy


Big effects
can be seen in
Paro
.


for patients of Alzheimer.


for assisting care
-
givers,



Some trials have been conducted


for children of autism or Down syndrome.


・・・

results in reducing the care costs,


reducing the quantity of


consumed medicine.


However,

it is only the assists to care
-
givers

or
more
???

H
ealing
by pets
Therapeutic effect
Japan
Korea
U.S.A
Brunei
Sweden
Italy
U.K.
Paro

-

culture gap
-

Principle components analysis on questionnaire surveys

Summary on
Paro


Non
-
verbal communication of
Paro

has a big effect with
patients of Alzheimer, like animal therapy.



Paro

also works good with people in emergency
evacuation area.



Paro

gives frequently topics of conversation.



Paro

plays good with care
-
givers as the interposition
between the patients and care
-
givers.



The cost merit of
Paro

in such situations is quite big.


Wakamaru


-
Mitsubishi Heavy Industry with Prof. Ishiguro
-

Geminoid

http://www.irc.atr.jp/Geminoid/

Wakamaru


It can be a guide in hospital, or museum, etc.



It can be a cast mate to human actors.



However, it does not seem to be real
communication.

Summary on
Wakamaru


Wakamaru

can work for guiding visitors in museums.



The conversation strategy is programmed by a
scenario. People can not feel the reality.



Wakamaru

can serve if the situation fits.



Muu

-

Research on
metamessage

or another
topics of communication
-

by Prof. Okada (Toyohashi)

It is a good tool for research
.

ifbot


-

Our trials for elderly and autistic children
-


Undergraduate

students



Elderly people who live alone



Autistic children


ifbot

Microphone

2

Sensor

(distance)


Sensor

(tactile)

Sensor

(distance)

Sensor

(impact)

CCD


camera

Speaker

Wheel

(foot)



Microphone

1



アアー眠い

恥ずかし



ワーイ

悲しい!

知らない!

しょんぼり

500,000 Yen = 3,900 UK pounds

Sale 100

This

robot

can

talk

with

individuals

using

the

voice

recognition

system
.

It

can

express

simple

emotions

with

LED

on

a

head,

eyes,

ears,

and

mouth

or

movement

of

arms
.

It

also

can

respond

toward

individuals

automatically
.

Young undergraduate students with
ifbot

-

Getting the impression from a robot
-


Compare the emotional intensity when a young student
use
ifbot

or computer.


The given task is to search required information.


64 students ( 49 male, 13 female)


Search by Q&A with
ifbot

Search by using web search engine



Blues and Anxiety



Hostile feeling



Boredom



Active pleasurable sensation



Passive pleasurable sensation



Affinity



Concentration



Surprisal

①抑鬱・不安
②敵意
③倦怠
④活動的快
⑤非活動的快
⑥親和
⑦集中
⑧驚愕
Questionnaire after the tasks

The subject selected one from four degrees in each question

The questionnaire consists of 40 questions, which are

categorized by the following

The subjects feel more something


in robot than in web browser



Blues and Anxiety



Blues and Anxiety



Hostile feeling



Hostile feeling



Boredom



Boredom



Activepleasurable



sensation



Activepleasurable



sensation



Passive pleasurable


sensation



Passive pleasurable


sensation



Affinity



Affinity



Concentration



Concentration



Surprisal



Surprisal

ifbot

Web browser

Monitoring daily living of two weeks with
ifbot
.

The comparison in cases of elderly people who live
alone and live with their family.


6 elderly people spent two weeks with
ifbot
, and 90 min
talk per day in average.


3 persons thought it is a toy.


: Elderly people with family


2 persons treated


as a conversation partner.


1 person feel


it is a kind of child or grandchild
.

Monitoring daily lives of two weeks with
ifbot
.



Twp elderly people who live alone answered higher rate (90%)
than the correct rate (57%) to the question for the recognition
of words in
ifbot
.

Autistic children with
ifbot


Study 1 explores a relationship between the feature
of toys which autistic children usually play with and
the development of the children’s social skills
through a questionnaire answered by their parents.



Study 2 investigates the effect of verbal
communication which the robot conveys on autistic
children’s attention and motivation to join training.


* The experimenter manipulated the response of
ifbot

in this
study.





Study 1


Feature of Toy and Development of Social Skills


Procedure:
19 parents of autistic children (age 6 to
17) participate in this survey. The parents are asked
to indicate the degree of social skill which their child
is capable and features of a toy which their child
prefer to play with in daily life

.


Study 1


Feature of Toy and Development of Social Skills


Questionnaire:
In section 1
, the parents respond to 4
items which are based on Social Responsiveness Scale
(SRS) , using 4 point scale (1= not true, 2 = sometimes
true, 3 = often true, 4 = almost always true);


eye contact
(“does he or she make eye contact?”),


gesture
(“does he or she use gestures to express his or
her feeling?”),


facial expression
(“does expression on his or her face
match what he or she is saying?”),


language

(“Does he or she use language to express his or
her feeling?”)


Study 1


Feature of Toy and Development of Social Skills


Questionnaire:
In section 2,
the parents are asked
to refer the feature of their children’s favorite toy.
Then, they judged whether the toy contains
following 4 features or not; face, movement of limb,
sound function, and wheel.


FACE

MOVEMENT
of LIMB

SOUND
FUNCTION

WHEEL


ANSWER


Yes or No


Yes or No


Yes or No


Yes or No

Results of Study 1

FACE

LIMB

SOUND

WHEEL

HAVE

NONE

MOVE

NONE

HAVE

NONE

HAVE

NONE

Eye
Contact

2.25

2.21

2.00

2.17

2.00

2.31

2.33

2.17

(0.50)

(0.80)

(0.00)

(0.72)

(0.00)

(0.85)

(1.03)

(0.58)

Gesture

1.75

1.50

1.40

1.50

1.75

1.54

1.50

1.58

(0.50)

(0.94)

(0.55)

(0.90)

(0.50)

(0.97

(0.84)

(0.90)

Facial
Expression

3.00

2.36

3.00

2.33

2.75

2.38

2.33

2.58

(0.00)

(0.63)

(0.00)

(0.65)

(0.50)

(0.65)

(0.52)

(0.67)

Language

2.50

2.14

2.00

2.17

1.50

2.46

2.50

2.08

(0.58)

(1.17)

(0.71)

(1.11)

(0.58)

(1.13)

(1.05)

(1.08)

Results of Study 1


Study2:


Autistic Children and Verbal Communication Robot


Procedure:
Participants were asked to engage in two interaction trials. In
a trial, the participants played a facial expression game with communication robot,
which played “instructor” of the game. In another trial, they engaged in the same
game with the experimenter, who played same role.


The facial expression game
designed to be a tool to determine whether their behavior was identical with
verbal information. At the beginning of the trial, we gave time to him or her to
adapt new environment. Then,
ifbot

asked him or her to express the emotion
which it inquired by verbal information and
ifbot

also made the correct facial
expression.



ifbot


Study2:


Autistic Children and Verbal Communication Robot



Measurements:


(1) The number of times making eye
-
contact toward the
instructor


(2) The number of times responding in verbal communication
by instructor


(3) Score of facial expression game



The

number

of

times

making

eye
-
contact

and

responding

in

verbal

communication

are

measured

when

those

performance

happen

during

the

time

we

give

the

children

before

the

facial

expression

game

for

adapting

to

new

environment

including

robot
.

The

score

of

eye

contact

and

response

in

verbal

communication

are

evaluated

by

experimenter

using

4

point

scale



Study2:


Autistic Children and Verbal Communication Robot


Measurements:


(1) The number of times making eye
-
contact toward the
instructor


(2) The number of times responding in verbal communication
by instructor


(3) Score of facial expression game






Case

1

.

H

is

a

girl,

12

years

old,

extremely

quiet

and

reserved,

who

usually

avoids

eye

contact

with

others
.

In

a

trial

with

the

experimenter,

at

first

time

the

experimenter

talked

to

her,

she

seemed

distracted

and

didn’t

even

look

at

the

experimenter
.

Eventually,

she

didn’t

speak

at

all

until

the

end

of

trial
.

In

a

trial

with

robot

instructor,

in

a

few

moments

she

met

ifbot
,

she

tried

to

touching

ifbot’s

face

and

was

looking

at

upside

of

ifbot’s

head

for

a

while
.

However,

when

ifbot

started

to

conversation,

she

almost

didn’t

respond

and

turn

her

posture

to

other

direction
.

At

the

moment

ifbot

raise

its

arms,

she

was

back

to

the

facial

expression

game

and

answered

to

Yes/No

question

at

least
.


Results of Study 2


Case 2.

O is a boy, 11 years old, more active than other
participants. His trial started to interact with
ifbot

first.
When O was introduced
ifbot
, he was very excited and
interested in
ifbot
. He was good at making facial
expression and talking with
ifbot
, but he didn’t respond
when
ifbot

asked a question was exactly same meaning
in a different way. After that happened, he was getting
distracted. A trial with the experiment was ended very
quickly and wasn’t completed. He seemed to know
that the trial will be going in the same way just like
before he was done with
ifbot

and kept saying that the
experiment is not a robot.

Results of Study 2


Case

3
.

M

is

a

boy,

8

years

old,

active

and

usually

talking

to

himself
.

At

the

beginning

of

trial,

Y

was

being

at

distance,

but

when

ifbot

moved,

he

was

exploring

very

closely

ifbot

face

features

in

a

few

moments

and

also

touched

ifbot’s

arm

several

times
.

He

didn’t

make

eye
-
contact

with

ifbot
,

even

talking

to

it
.

As

time

went

on,

it

seemed

he

became

lost

his

interest

to

ifbot

and

started

talking

to

himself

again
.

In

a

trial

with

the

experiment,

Y

was

sitting

very

close

to

experimenter

but

he

didn’t

make

eye
-
contact

and

was

talking

himself
.

He

continued

talking

to

himself,

even

the

facial

expression

game

started
.

Results of Study 2


Trends of all participants:
The
attention level which
included the number of making eye
-
contact, the number of response in
verbal communication and facial expression game score of each child.
(Each item response is scored: 1=Very poor, 2 =Poor, 3 =Good, 4 = Very
good), Facial expression game score is calculated by summing up the each
scores of four tasks. (Each score is coded as 0=no response, 1=response)


Results of Study 2

Eye Contact

Response in Verbal
Communication

Facial Expression

Game
Score

Robot

Human

Robot

Human

Robot

Human

CASE1

4

1

3

1

4

0

CASE2

3

2

2

1

0

0

CASE3

3

3

3

2

4

0

CASE4

3

2

3

3

0

0

CASE5

4

3

2

3

0

0

CASE6

1

4

1

1

0

0

Mean

3.00

2.50

2.33

1.83

1.33

0.00

SD

1.10

1.05

0.82

0.98

2.07

0.00

Summary


Autistic children with
ifbot


We found a robot for autism therapy needs to contain
a face and movement of limbs based on Study 1.



The results of Study 1 showed sound function doesn’t
have the effect on communication skills.



Verbal function of robot is required with more
systematic approach for adapting as an element of
the robot from Study 2

.



Summary


Autistic children with
ifbot


We found a robot for autism therapy needs to contain
a face and movement of limbs based on Study 1.



The results of Study 1 showed sound function doesn’t
have the effect on communication skills.



Verbal function of robot is required with more
systematic approach for adapting as an element of
the robot from Study 2

.



Summary


Our research in mental problems with
ifbot


ifbot

stimulates more user’s feelings in comparison with
computers when the users conduct search tasks. Note
that the subjects were familiar with web search but were
not with
ifbot
.



Elderly people who live alone tend to understand the
conversation with
ifbot

is real . The proof appears in their
wrong answers to the question of recognition rate of
words.



Ifbot

has a big potential for curing autistic children. We
regard that the face and movement of limb are the keys of
imitation and the imitation might lead to improve autistic
children’s communication skills.




Discussion

“Uncanny valley” proposed by Prof. M. Mori in 1976

Affinity

Degree of similarity

to human

a toy doll

H
uman

For an example: moving
corpse

Uncanny valley

No one knows the exact axes and the shape of function.

Discussion

What is the adequate feature of robot for children with
autism in robot
-
assisted therapy or education?

Effect

on

Improvement of
communication
skill

Imitation level or
function of
communication

a toy doll

H
uman

Valley ?

a good robot for

improving communication
skill of autistic children

Discussion



Communication

Robots

can

be

of

service

for

people

with

appropriate

therapists,

or

care
-
givers,

or

teachers

if

there

exists

the

necessity

.

We

can

find

the

cost

benefit

in

such

cases
.




We

cannot

identify

the

origin

of

such

effects

or

differences

with

communication

robots
.

At

this

stgae,

there

exist

a

lot

of

discussions

on

the

requirements

for

the

robots

and

the

design

specifications

which

meet

those

requirements
.

Further

research

will

be

required
.



Thank you. Questions?


These experiments described in
this presentation
were partly supported by Japanese Ministry of
Education,


and were conducted with my colleagues Dr. K.
Matsumoto, Dr. H. Takehashi, T. Suzuki, N. Takeda and
J. Lee. I would like to acknowledge them.