silvia coradeschi aass mobile robotics laboratory university of Örebro, sweden silvia coradeschi...

Post on 12-Dec-2015

214 Views

Category:

Documents

0 Downloads

Preview:

Click to see full reader

TRANSCRIPT

Silvia Coradeschi

AASS Mobile Robotics LaboratoryUniversity of Örebro, Sweden

Silvia Coradeschi

AASS Mobile Robotics LaboratoryUniversity of Örebro, Sweden

Robot ecology for an ageing society

www.aass.oru.se

Contributors: G. Cortellessa, M. Scopelliti, L. Tiberio (ISTC);A. Saffiotti, F. Pecora, A. Loutfi, J. Rashid, M. Broxvall, L. Karlsson

and the rest of the Mobile Robotic Lab at AASS

An ageing population

• A clear problem in the industrialized world

• Can robotics offer partly a solution?

• Would a traditional robot work in such an application?

• What is an acceptable solution for the user?

• How to test acceptability at an early stage?

Outline

• Single robot vs ecology of intelligent systems

• Expectations of end-users– Video-based evaluation (Sweden vs Italy)

– On-going evaluations with a teleoperated robot

• Wizard of Oz method for evaluation

What do you think when you hear the word Robot?

Robots at home – most popular vision (and also the robocup@home vision ??)

Animation Markus Decker Fraunhofer-Institute for ManufacturingEngineering and Automation, Stuttgart

Robots at home – problems with this vision

• Technological– understand the human– recognize objects– grasp– dexterous manipulation– position tracking– ...– ...

• Socio-economical– will you accept this thing?– how much would you pay?– how will it grow tomorrow?– one size fits all: does it?– ...– ...

Robots at home – an ecological vision

Robots at home – integration of olfaction

Robots at home – tiny devices

Key points in the scenarios

• Many specialized robotic devices– Moving tables, cameras, manipulators, appliances, ...– Highly heterogeneous

• Devices communicate and cooperate– Common communication and cooperation model

• Complex abilities achieved through cooperation– Not by building a single “super-robot”

Research Contributions

• Hardware and software developments– Development of middleware for communication– Integration of new devices (Tiny OS, gas sensors etc.)

• Information Fusion– Maintain coherent information about objects and persons.

• Learning and Artificial Intelligence– Automatic configuration of PEIS ecology to perform different

tasks under different conditions. – Activity monitoring and forecasting, in order to adapt the robots’

activities to the humans’ activities

http://aass.oru.se/~peis

RoboCup@home

• Important to have also distributed solutions beside single robot

• Allow for both variants?

• Technical challenges

• How to adapt rules for both cases?

• How to fix such a system in the short time allowed for the competition?

Robots at home – problems with this vision

• Technological– understand the human– recognize objects– grasp– dexterous manipulation– position tracking– ...– ...

• Socio-economical– will you accept this thing?– how much would you pay?– how will it grow tomorrow?– one size fits all: does it?– ...– ...

• Intelligent homes & robotics have a large potential to assist the aging society.

• Many solutions are either seen from the view of the engineers and technicians or from psychologists and behavioural scientists.

• Absence of common techniques to test and validate our methods, particularly for robotics.

Robots at home – problems with this vision

Our evaluation with elderly grounded on two implemented smart-home prototypes

The ROBOCARE projectDeveloping software and

robotic technology for providing intelligent support to elderly people

A testbed for an a-posteriori evaluation of smart home technology with Italian user groups

http://robocare.istc.cnr.it

Ecologies of PEISAn ecology of physically

embedded intelligent systems (e.g., robots, actuators, sensors, SW agents)

Shares numerous commonalities with ROBOCARE, gives us the opportunity to extend previous results in evaluation

http://aass.oru.se/~peis

• Phase 1 – video sessions– Male vs. Female actors– Robot interface vs. None.– Interface on the actual robot. – 8 different scenarios

• Results– A diverse group of healthy testpersons,– 56-88 years old. – Crosscultural evaluation.

Cooperation with – Pensionärsriksorganization (PRO), BSR ÖU, Teknik ÖU, Institute for Cognitive Science and Technology, Italian National Research Council.

• Phase 2 – Live Sessions: telepresence and Wizard of Oz methodology

HRI – Human Robot Interaction

Video-based evaluation of developed prototypes

• Understanding the real perception of elderly people toward the assistive robot in a real domestic environment

• Identifying meaningful everyday situations for human-robot interaction at home

• Providing a general evaluation of the robotic mediator (physical aspect, capabilities and integration in the domestic environment)

• Understanding the preferences of elder users with respect to the presence of human features in the robot

• Preferences with respect to interaction modalities (Proactive vs. On Demand)

Video-based evaluation of developed prototypes

Eight scenarios

On-demand scenarios• Finding objects• Activity planning• Reminding medication

Proactive scenarios• Environmental safety• Personal safety• Health-related reminders• Suggestions• Reminding events

Two experimental conditions

FACE condition: a robot showing a human speaking face on a notebook monitor

NO-FACE condition: a robot with no reference to human features

NO-ROBOT: just sensors and actuators without a robotic interface

Cross-cultural evaluation methodology

•Similar demographics

•Administration and evaluation methodology kept consistent

•Participants drawn from senior citizen associations and study groups

Demographics

40 Participants(Aged 56-88)

43 Participants(Aged 58-87)

Results: Scenarios

Results: Evaluation of the robotPositive aspects Mean Std. dev.

Obstacle avoidance capability 3.54 .74

People feel safer 3.27 .93

Reduction of age-related impairments 3.15 1.01

Ability to speak 3.13 .85

Results: Emotional reaction to the robot

Results: Human-like features

Results: Family status (Italy)

Results: Family status (Sweden)

Results: Perceived health conditions (Italy)

Results: Gender

ICT in Italy and Sweden

Italy Sweden

2.8% of households of people aged over 65 has access to the Internet;45% of them own a mobile phone

73% of households in 2005 had access to the Internet; 51% of individuals aged 55-74 age use the Internet at least once a week

Different level of confidence in technology• Swedes perceive system as more useful• Italians are more afraid of programming and

maintaining the robot

Data taken from the “Supporting Policy Development for e-Inclusion”

project —seethe the EU country reports section on http://www.ipolicy.eu/

Policy in Italy and Sweden [Sundström and Johansson, 2005]

Italy Sweden35% of elderly people lived with other family members and/or within communities in 2000

98% of elderly people lived either alone or with their spouse in 2002

• Swedes lead more independent lives in old age

• ==> the system threatens their ability to remember things on their own and worry that they will become dependent on the system

South EU Nordic countries

67% of the elderly received care from family members within their own household in 1992

17% of the elderly received care from family members within their own household in 1992

Human-like attributes

• Both groups, when answering “which version of the robot (FACE vs. NO-FACE) do you prefer?” answered “FACE”

• However– both groups indicated a preference for the NO-FACE version in

the other questions

– e.g., better integration in the home, less impersonal, less threat to pets, more reassuring during dialogue, less irritating when it takes decisions autonomously

• This is true across cultures

• The version with the robot is preferred to the version without the robot

Current and next step evaluation in a real environment

• Telepresence evaluated in real environments

• Wizard of Oz in our test apartment

• Test apartment in newly build senior home

Telepresence system

• A “simple to use” telepresence system that can be a reasonable first step to home robotics for elderly

Evaluation of telepresence

• On-going experiments with Giraffe a teleoperated robot

• A first (more acceptable) step towards home robotic

• Several acceptability trials:– Seniors homes, elderly homes

• Communication:– Relatives to elderly

– Elderly to elderly

– Care givers to elderly

Inspired from HCI, WoOz allows us to simulate a system (unknown to a user) and evaluate their response.

Wizard of Oz setup

What the elderly/elderly care people would like (informal survey)?

• Assessment of dangerous situation– During the night

– For elderly living isolated

• Technological help for documentation– Voice recognition

– Events recognition

• Technological help for everyday tasks– Brush teeth

– Eat

– Take a shower

– Walk

Thank you

top related