Trends in Robot Assisted Therapy and Robot Enhanced Therapy · Robotics and Embedded Systems Department of Computer Science Technical University of Munich Prof. Dr.-Ing. habil. Alois

Trends in Robot Assisted Therapy and Robot Enhanced Therapy

Presentation by Cristina Gil Ávila and Yousef Yegane

Machine Learning in Robot Assisted Therapy Summer Semester 2018

M. Sc. Sina Shafei

Robotics and Embedded Systems ▪ Department of Computer Science ▪ Technical University of Munich

Prof. Dr.-Ing. habil. Alois Knoll

Overview

Robot Assisted Therapy vs. Robot Enhanced TherapyWhy Robot Assisted Therapy in Autism?What are the current needs of Autistic people?

SS2018 2Machine Learning in Robot Assisted Therapy / Trends in Robot Assisted Therapy and Robot Enhanced Therapy

Introduction


Our vision

Conclusions

Targeted fields in Autism by RAT and RETMapping robots to ASD objectivesReview of robots uses in RAT and RATDistribution of projects by countryCommercial availability and budget

Methodology of research

Towards more autonomous robots (‘supervised autonomy’)Orientation towards self-care and independent livingLow-cost robots that can easily be used at home

Technological approach



Introduction

Initial distinction made by the DREAM project1

SAR: Socially Assistive Robotics

RAT: Robot Assisted Therapy

Controlled by a human agent

RET: Robot Enhanced Therapy (Next Generation of RAT)

Semi/Fully automated robots

Able to infer psychological disposition and asses

behaviour

Tailored to individual needs

Diagnostic tool


Robot Assisted Therapy and Robot Enhanced Therapy in Autism

Source: DREAM project1

1https://www.dream2020.eu/



Introduction

• Robots can provide human cues while maintaining object-like simplicity

• Robots can be controlled so that only relevant information is presented

• Lighten the burden of human therapists

• Reducing the costs of standard cognitive therapies

• More rational beliefs and lower intensity level of emotional responses have been reported in RAT


Why RAT and RET in Autism?

(Boucenna et al., 2014)



Introduction

Mc Call, 2017: UK falling to meet the needs of people with Autism (Report from the National Autism Project)

• Lack of strategic research: £32 billions invested each year with little impact

• High unemployment rates: ‘only 16% of adults diagnosed with autism in England are in full-

time employment, and two-thirds are not working at all’. Main reason is attributed to lack of

support from society and social and communication difficulties

• Lack of consideration towards health needs. Examples are controlling weight and

maintaining healthy diets.

• Address the high variability inside the autistic spectrum. ‘When you’ve met one autistic

person, you’ve met one autistic person’.


What are the current needs of Autistic people?

(Mc Call, 2017)



Methodology of Research

We used statistical approach to review the current state of research.

Types of therapies based on 2791 documents with the keyword of robot assisted/enhanced therapy

These are general main categories that are performed by the assistance of a robot:

• BCI (Brain Computer Interface) Assisted Robotic Therapy• Robot Assisted Rehabilitation Therapy• Post Stroke Robotic Assisted Therapy• Social Robot Assisted Therapy

6




Robot Assisted/Enhanced Therapy in Autism (Based on the research papers*)

• Categorized under social assisted therapy• 143 papers out of 2791 papers in RAT/RET therapies (about 5 percent)• Relatively new trend in the field of RET/RAT therapies• Most of the research in this area was done in the field of CS

* June 2018-Source Scopus with the keywords of “Robot AND (Assisted OR Enhanced) And Therapy” with the refined keywords of “ASD OR Autism OR Autistic” in Titles, Keywords and Abstract

7




Understanding research patterns:

General steps that are performed in RET/RAT projects

● Physical Robot Design○ Size○ Anthropomorphism○ Range of Movement○ Body Parts

● Human-Robot Interaction Design○ Joint Attention○ Turn Taking○ Imitation

● Evaluations Of Robots In Therapy-like Settings○ Currently not performed in sense of medical standards in fields of CS/Mathematics

8




DiagnosisCollecting data for interaction and automate behaviour evaluation systems (eg. eye tracking).

Social interactions and reactions

Self-initiated interactions: encourage the child to engage the therapy proactively

Turn-taking activities: get children used to wait for a response after a performed action

Imitation: recognize ‘social others’, help children realize their actions are related and have an impact on

their environment

Joint attention: ability to consciously focus on the same object with other person. Robots improve that

automatically focusing their gaze towards children.

Emotional wellbeingEmotion recognition: robots simplify the complexity of a human face

.


Targeted fields in Autism by RAT and RET

(Dickstein-Fischer & Fischer, 2014; Ricks & Colton, 2010)





Mapping robots to ASD Objectives

(Huijnen, Lexis, Jansens, & de Witte, 2016)





Review of robots used in RAT and RET

Nao Robota Probo Keepon Cat robot I-sobot Tito GIPY

HOAP 3 KASPAR Robot arm Pleo Labo 1 ifbot CHARLIE PABI

(Huijnen, Lexis, Jansens, & de Witte, 2016)

Image

not

available





Distribution of projects by country





Commercial availability and budget

Commercially available Price Commercially unavailable (research)

NAO N.A.

Keepon 170$

I-sobot 600$

Hoap 3 N.A

KASPAR 2000$

Pleo 469$

Robota

Cat robot

Probo

Tito

GIPY

Robot arm

Ifbot

Labo 1



Our vision

14

1. Needs of joint research in Medical and Computer Science research groups

2. Towards more autonomous robots (‘supervised autonomy’)

3. Orientation towards self-care and independent living

4. Low-cost robots that can easily be used at home



Our visionNeeds of joint research in Medical and Computer Science

● Computer Science is the most active field of science to investigate RAT/RET in ASD● Question: Why in the field of medicine has not performed numerous investigations about ?● Our interpretation:

○ This field is still under development and we have a distance until we get to a reliable product.○ CS research approach: Careful quantifications and calculation wit limited group of samples○ Medical research approach: Large group of samples based on observation

15



Our vision

16

At short term, we strongly suggest following the guidelines of the DREAM project in any future attempt to

construct a social robot for autism.

Wizard of Oz technique is not a good option in the long run.

Supervised autonomy: the psychologist, therapist or teacher gives the robot a particular task and the robot

autonomously works in that task under the supervision of the user.

Esteban et al., 2017 proposes an architecture with three subsystems:

1. Sensing and interpretation

2. Child behaviour classification

3. Social cognitive controller

At long term, we propose a shared supervision between parents and therapists. It will reduce the costs and

could be easily done at home after a required training of the parents.

Towards more autonomous robots



Our vision

17

Mc Call, 2017 stated that one of the main problems of adults with autism in England is carrying an independent

living. So far none of the robots used in RAT and RET has addressed this problem.

We believe that increasing the level of autonomy during the childhood will positively impact in the adulthood.

For that reason we encourage the targeting of objectives such as personal hygiene, potty training, healthy

eating and drinking, (un)dressing, etc. This objectives could be reached through imitation and rewarding (eg.

imitation of daily life activities such as brushing their teeths).

Nevertheless, this doesn’t mean decreasing the focus in social interactions and reactions, which has been

reported to be one of the major needs in ASD.

Orientation towards self-care and independent living



Our vision

18

Little commercial development has been developed so far, with prices that are sometimes not affordable for

average household incomes.

Low-cost robots that can easily be used at home

We propose future development towards low-cost portable

devices for home-based autism therapies. As an example we

show Probolino (Cao et al., 2015) which faces the challenges of

dimension, cost, and autonomy.

Source of the photos: Cao et al., 2015



Conclusions

19

Current approaches in Robot Assisted Therapy and Robot Enhanced Therapy focus mainly in the development

of social interactions and reactions.

According to the current needs of autistic people, social interactions and reactions should be reinforced and

self-care and independent living should start be targeted by RAT and RET.

Future approaches should be framed into Robot Enhanced Therapy, with focus in a ‘supervised autonomy’.

Current commercial approaches are often expensive and cumbersome to transport. There is a need of low-

cost, portable, prototypes that can be easily used at home.



Bibliography

20

Boucenna, S., Narzisi, A., Tilmont, E., Muratori, F., Pioggia, G., Cohen, D., & Chetouani, M. (2014). Interactive Technologies for Autistic

Children: A Review. Cognitive Computation, 6(4), 722–740. https://doi.org/10.1007/s12559-014-9276-x

Cao, H.-L., Pop, C., Simut, R., Furnemon, R., De Beir, A., Van de Perre, G., … Vanderborght, B. (2017). Probolino: A Portable Low-Cost Social

Device for Home-Based Autism Therapy.

Dickstein-Fischer, L., & Fischer, G. S. (2014). Combining psychological and engineering approaches to utilizing social robots with children with

Autism. 2014 36th Annual International Conference of the IEEE Engineering in Medicine and Biology Society, EMBC 2014, 792–795.

https://doi.org/10.1109/EMBC.2014.6943710

Esteban, P. G., Baxter, P., Belpaeme, T., Billing, E., Cai, H., Cao, H. L., … Ziemke, T. (2017). How to build a supervised autonomous system

for robot-enhanced therapy for children with autism spectrum disorder. Paladyn, 8(1), 18–38. https://doi.org/10.1515/pjbr-2017-0002

Huijnen, C. A. G. J., Lexis, M. A. S., Jansens, R., & de Witte, L. P. (2016). Mapping Robots to Therapy and Educational Objectives for Children

with Autism Spectrum Disorder. Journal of Autism and Developmental Disorders, 46(6), 2100–2114. https://doi.org/10.1007/s10803-016-

2740-6

McCall, B. (2017). UK failing to meet the needs of people with autism. The Lancet, 389(10067), 351. https://doi.org/10.1016/S0140-

6736(17)30197-6

Ricks, D. J., & Colton, M. B. (2010). Trends and considerations in robot-assisted autism therapy. Proceedings - IEEE International Conference

on Robotics and Automation, 4354–4359. https://doi.org/10.1109/ROBOT.2010.5509327

Brian Scassellati, Henny Admoni, and Maja Mataric (2012). Robots for Use in Autism Research. The Annual Review of Biomedical Engineering

https://www.dream2020.eu/

https://www.dream2020.eu/

Documents

Trends in Robot Assisted Therapy and Robot Enhanced Therapy · Robotics and Embedded Systems Department of Computer Science Technical University of Munich Prof. Dr.-Ing. habil. Alois