Robots in Autism

Robots in Autism

A sampling of

Approaches and Anecdotes


4 Dec 2008

Outline


Who’s who, and their robots


Observed Behaviors


Proto
-
therapies


Key assumptions


Trends

Who’s who


First apparent use:


Using LOGO to catalyse
communication in an autistic child.
1976.
S. Weir and R. Emmanuel,

Technical report
DAI Research Report No. 15, University of Edinburgh,



Common origins:


Rod Brooks’s
lab at MIT,1998
-
99

nexus


Dautenhahn, Scassellati, Kozima

Who’s who:

common roots
---

Rod Brooks


Kerstin Dautenhahn


U. Hertfordshire, UK


Francois Michaud


U. de Sherbrooke, Canada


Brian Scassellati


Yale, USA


Hideki Kozima


Japan


Maja Mataric


USC, USA


Who’s who, and their robots:

Dautenhahn


Kaspar

Who’s who, and their robots:

Dautenhahn, Billard


Robota

Who’s who, and their robots:

Dautenhahn


Non
-
humanoids

Who’s who and their robots:

Dautenhahn & Robins


Humans

Who’s who and their robots:

Michaud


Who’s who and their robots:

Kozima


Who’s who and their robots:

Kozima


Who’s who and their robots:

Kozima


“
My motivation [is] to understand
human communication, and the
developmental process behind
language, so that I can recreate
that model in machines.
…

Autism is a disability in verbal
and nonverbal communication
skills. By understanding autism,
we can gain knowledge of the
normal cognitive processes
behind communication.
”


Who’s who, and their robots:

Mataric


http://cres.usc.edu/pubdb_html/files_upload/549.pdf

What behaviors are seen:

AURORA


Triadic interactions

What behaviors are seen:

AURORA


Turn
-
taking (or lack thereof)


Exclusive dyadic interactions

What behaviors are seen:

AURORA


Imitation

What behaviors are seen:

AURORA


Reaching out

What behaviors are seen:

Kozima


“
Autistic children will generally
be
scared

of Keepon at first, but
they are also very curious. After
several months they will
understand that Keepon will not
harm them, and then we will see
them start to
voluntarily touch

and
make eye
-
contact

with it.
…

almost all autistic children come
to enjoy their interactions with
Keepon. ... Why? I think it
’
s
because Keepon is simple.
”

What behaviors are seen:

Kozima (typical children)


[http://mainline.brynmawr.edu/DevRob05/schedule/papers/kozima.pdf]

What behaviors are seen:

Kozima
(1.5yr study, two 2
-
4 year
-
old children)


“
Keepon's simple appearance
and predictable responses gave
the autistic children a playful and
relaxed mood
”


“
[Spontaneous engagement] in
dyadic

play with Keepon
”


“
[Expansion into] interpersonal
communication where Keepon
worked as the
pivot of triadic
play

with adults or other
children.
”


http://ieeexplore.ieee.org/xpls/abs_all.jsp?arnumber=1513802

What behaviors are seen:

Mataric


The robot in an experimental setting. The
bubble
-

blowing robot

(left) is known to be
less intimidating

than a humanoid

robot (right).

http://cres.usc.edu/pubdb_html/files_upload/589.pdf

What behaviors are seen:

Mataric


A child interacting with a
…

robot that responds to the
child
’
s behavior
…

will
exhibit more social behavior
than when interacting with a
[robot that responds
randomly]

http://cres.usc.edu/pubdb_html/files_upload/589.pdf

What behaviors are seen:

Mataric

Dark: contingent robot

Light: random robot

Total

speech

Robot

speech

Parent

speech

Total

Robot

Inter
-

action

Push

button

Re
-

sponse

to

robot

Di
-

rected

Inter
-

action

http://cres.usc.edu/pubdb_html/files_upload/589.pdf

What (proto) Therapies:

AURORA


Imitation

What (proto) Therapies:

AURORA


Robot as mediator for social
interaction

What (proto) Therapies:

AURORA


Shared attention

What (proto) Therapies:

AURORA


Mutual gaze

What (proto) Therapies

Dautenhahn & Francois


Play


Dyadic: child
-
robot
--

progressively better balanced
interaction


Triadic: child
-
robot
-
experimenter


Therapies:


robot
-
assisted play


automatic recognition of human
-
robot interaction styles in real time


adaptive robot responsive to
different styles

What (proto) therapies:

Kozima


“
If Keepon can become a
trigger

for autistic children to
experience social interaction,
then it might support their
social
development
. If a robot can
become a catalyst between the
child
’
s natural growth and desire
to interact, and the power of [or
a tool for]
people supporting
the child

(such as parents and
therapists), that would be great.
”

Key assumptions


Robots are predictable


Robots are simple … “not toys
nor a human”
--

Kozima
[http://mainline.brynmawr.edu/DevRob05/schedule/papers/kozima.pdf]


Robots engage and hold interest
for children w/ ASD


Robot autonomy and adaptation
is useful


Therapist
-
mediated activity is
useful


Children w/ ASD can play w/
robots

Trends


Use robots that are much
simpler than humans


Support basic contingency
-
based
autonomy OR


Wizard of Oz
-
based contingent
behavior


Children need to acclimate


Dyadic interaction facilitates
triadic interaction