from smart home to smart care : pervasive assistance for cognitively impaired people sylvain giroux

From

Smart

Home to

Smart

Care :

Pervasive

Assistance for

Cognitively I

mpaired People

Sylvain

Giroux

Plan

Context

Objectives and Approach

Pervasive and mobile computing, Tangible User interface

From homes…

Hardware level: networks, sensors, effectors

… to smart homes …

Middleware level: pervasive infrastructure

… to smart care !

Application level: cognitive assistance & tele-monitoring

Validation

Clinical studies

Conclusion

ContextP

eople suffering of cognitive impairments in Quebec Alzheimer disease : 5.1% of people over 65 years old Head trauma : 3000 new cases each yearSchizophrenia : 1% of the population

In many cases, they would be able to stay at home

if light assistance was provided.

But healthcare resources are scarce. So relatives have to take responsibility for care.

It then turns to an exhausting burden.Hence relatives and caregivers urge for help.

Objectives

Provide adapted and personalized environmental cues to

Foster the autonomy of cognitively impaired people

Reduce risks and hazards

Pervasive computing & Tangible user interfaces

Keep ensuring continuous cognitive assistance outside people’s home

Mobile computing & Location-based services

Help relatives and caregivers to stay in touch at distance with cognitively impaired people

From homes…S

mart homes are augmented environments

Heterogeneous networks

Sensors networks

Embedded processors in devices, clothes, jewels…

Information appliances

Networked communicating objects

DOMUS: an augmented apartment

… smart homes …S

mart homes are augmented environments

Heterogeneous networks

Wireless : WiFi, Bluetooth, RFID, UWB...Wired : Ethernet, Electrical wires, X10, power line…

ServersFull control over sounds and video streams

… smart homes …S

mart homes are augmented environments

Sensors networks

Identification and localization of objects and people

Ubisense tags, UWBSmart tags (RFID)

… smart homes …S

mart homes are augmented environments

Embedded processors in devices and clothesNot yet investigated

… smart homes …S

mart homes are augmented environments

Information appliancesFixed: smart boards, Icebox… Mobile: laptop, wireless screen, PDAs…

… smart homes …S

mart homes are augmented environments

Networked communicating objectsSightHearing

Smell (not yet investigated)Touch (not yet investigated)

Taste (is it possible ?)

From

homes to smart

homes

Towards a pervasive infrastructure

Some issues investigated at DOMUS

Spontaneous networkingHeterogeneous networksAutonomic computingMobile code and agentsLocation and context awareness

Security and privacy

Some prototypes

A pervasive reminder systemMulti-channel delivery of services

A

Pervasive

Reminder

System

for

Smart

Homes

How to localize a user from simple sensors information ?

How to achieve pervasiveness?

Follow-me : transparent user friendly migration of sessions

How to use spontaneous networking and service discovery to build zero-configuration system ?

How to cope with heterogeneity of devices, networks, and OS ?

How to keep the system in a clean state ?

Multi-channel

delivery of services

On-the-fly generation of user interfaces from raw code

Means to control complex interactions with a user

Ready to use service delivery infrastructure

… to smart careS

mart homes can assist cognitively impaired people foster their autonomy.

The whole home becomes a true cognitive prosthesis.

As well, smart homes can

help caregivers to grant better care give a sense of security to residents and their relatives

Cognitive assistance

What is the available information ?

Identification and location of people and objectsObjects involved in an activityPrimitive actions

Who the user is ?

personalization

What is the user doing ?

Activity recognitionHierarchical modelsLattice-theory

How to assist the user?

Highligth objectsTangible user interfaces

Activity recognition

Based on plan recognition

Hierarchical Markov modelsLattice-based models

Based on involved objects

Perkowitz et al., Mining Models of Human Activities from the Web, WWW 2004, May 17-22, 2004, New York, NY USA.

Personalization

Every person is different

Cognitive modelling exploiting episodic memory

To know from life habits, how one usually performs an activity

method usually used to achieve a taskestimations on time (averge time of completion…)most likely location

Cognitive deficits

Deficits addressed

Initiation

Memory

Planning

Attention

Initiation deficitsInitiation deficits leads to inactive periods whereas the person is supposed to perform actions

Wandering for a long time could be attributed to an initiation deficit

>> Prompt the resident

Memory deficits

Difficulties to remember

the activity to perform the steps of the activity the locations of the tools and materials involved in that activity.

>> Show-me objects + « Follow-me » applied to objects

The lamp turns off when the object is too far away.

The lamp turns on to highlight the searched red book

Planning deficitsD

ifficulties to perform an appropriate sequence of actions in the rightorder to achieve a goal.

>> Prepare_coffee = {take_milk, take_cup…}

>> Show to the user where to perform the next action

Attention deficitsS

hifts of attention from the activity under progress to a

stimulus causing interference

The current activity may be forgotten and never completed

>> Remind the activity under progress to the resident

Validation

Clinical studies using prototypes are in preparation and will soon be on-going at

Fernand-Séguin research center, L-H Lafontaine Hospital, Montreal

Dr Emmanuel Stip, psychiatristSchizophrenia

Centre de Réadaptation Estrie, SherbrookeHead trauma

Conclusion

Pervasive computing and tangible user interfaces can help to transform home into smart homes adapted to cognitively impaired people

Going beyond the usual view of computing as “PC-based”

Pervasive computing enables a seamless integration of assistance in residents’ everyday life

Going beyond traditional human computer interfaces

TUI helps to turn the whole house into a cognitive prosthesis

Such smart homes can

Foster people’s autonomy

Lead to smarter care

Our team

Researchers from the Faculties of science, engineering, and administration

Sylvain Giroux, Ph. D. in Computer scienceHélène Pigot, Ph. D. in Computer science and B. in occupational therapy André Mayers, Ph. D. in Computer science and M. inpsychologyPhilippe Mabilleau, Ph.D. in engineeringClaude Caron (geo-business)

Analyst

Francis Bouchard

Students

6 Ph. D. students12 M. Sc. students6 B.Sc. students, 2 international trainees (M. Sc. level)

Some collaborations

CRE, Centre de réadaptation Estrie Centre de recherches Fernand Séguin, Computer scienceUniversité Joseph Fourier, Grenoble

France TelecomAriane Controls, Canada

Thank you for your attention

http://domus.usherbrooke.ca/