The role of intelligent habitats in upholding elders in residence - PowerPoint PPT Presentation

The role of intelligent habitats in upholding elders in residence Hélène Pigot 1 Bernard Lefebvre 2 Jean-Guy Meunier 2 Brigitte Kerhervé 2 André Mayers 1 Sylvain Giroux 1 1) Département de mathématiques et d'informatique Université de Sherbrooke Canada. 2) Département d'informatique Université du Québec à Montréal Canada. 1 de 23

Introduction The need for intelligent habitats in upholding elders in residence Their number increases They legitimately wish to remain at home as long as possible For economic reasons, governments also want to maintain them in their residence Elders are suffering from several chronic diseases The safety problems Immediate risks Long-term risks This talk introduces the theoretical and computational models needed to assist elders in their Activities of Daily Living (ADL) to inform relatives and caregivers as soon as necessary. 2 de 23

Outline Introduction (done) A layered computer infrastructure hardware layer code middleware layer code Interpretation and decision making code Computational and theoretical model and metamodel of the person the task the environment Implementation consideration Conclusion 3 de 23

A layered computer infrastructure cognitive assistance telemonitoring module module Interpretation and decision making integrator layer code module middle layer code hardware computer micro-wave bathroom door kitchen table glycometer layer code 4 de 23

The Grenoble Experimental Smart Home Movement c c Door contact b) Entrance hall � � � e e Scale c) Kitchen f f Tensiometer d) Living room d d � e) Bedroom Oxymeter ? g g f) Shower Fall detector g) WC Sound antenna b b h) Technical room Microphone h h Control panel 5 de 23

hardware layer code Four categories of sensors activity actimetry physiology environment 6 de 23

middleware layer code Purpose Characteristics wireless spontaneous networks autonomous components distributed systems distributed algorithms Framework (abstract classes and tools) 7 de 23

Interpretation and decision making code Activity meta-model Environment meta-model Person meta-model Activities Instantiated Instantiated in process environment person gr Inte ator Assistance HIT Kitchen Telemonitoring sensor ware midd le Code for the integrator module, cognitive assistance module and telemonitoring module 8 de 23

Outline Introduction (done) A layered computer infrastructure hardware layer code middleware layer code Interpretation and decision making code Computational and theoretical model and metamodel of the person the task the environment Implementation consideration Conclusion 9 de 23

The computational and theoretical model behind the architecture Environment meta-model Person meta-model Activity meta-model Sensors Cognitive Activities of daily living appliances physiological (ADL) visual, accoustic interfaces behavioral scripts Instantiated Activities Instantiated person in process environment Intelligent habitat Competence model with a real person (Rousseau) middle layer code cognitive distance integrator module assistance monitoring module module Decision and action: competence or handicap 10 de 23

Person meta-model The cognitive model Act-r (Anderson) Miace (Mayers) CS/SAS (Norman & Shallice) The physiological model blood pressure osteoporosis alimentation hygiene 11 de 23

Ethics We don’t pretend and don’t want to replace human communication We want to alleviate distress and fatigue for relatives. We want security, autonomy and human dignity for elders. 12 de 23

Adapted interventions Characterized domain: hygiene, alimentation … severity: “doing nothing”, advice, drastic frequency: low … Personnalized cognitive capacities: decreasing rate of memory trace activation cognitive abilities: “cooks spaghetti sauce wihout any hesitation” habits: nap after lunchtime … preferences: acoustic, visual signal … history: “eat chicken for lunch today” 13 de 23

Activity meta-model basic movements: infered from intelligent sensors (opening of the bathroom door, opening of hot water tap, …) actions: meaningful sequence of movement (brushing teeth) Activities of Daily Living (ADL): meaningful sequence of actions (morning hygiene) 14 de 23

Activity meta-model Contains Successful and failed scripts of ADL Failed scripts pre-selected erroneous situations due to cognitive impairments. Interventions of the environment linked to failed scripts in order to help the person. Tools under analysis Cogent, Epitalk 15 de 23

The environment model Description of generic physical and human environment Physical environment sensors appliances transmitters Human environment family neighbour medical staff 16 de 23

Outline Introduction (done) A layered computer infrastructure hardware layer code middleware layer code Interpretation and decision making code Computational and theoretical model and metamodel of the person the task the environment Implementation consideration Conclusion 17 de 23

Implementation consideration Analysis of the behavioural patterns to detect abnormal behaviour of the monitored person. This mechanism must be as accurate as possible noise resistant able to adapt to failures absence of sensors. 18 de 23

Various mathematical methods and logics multi-factorial analysis neural networks hidden Markov models bayesian networks knowledge representation first order logic temporal logic description logic 19 de 23

Tools SATIM system Epitalk 20 de 23

You don’t see the computer for the interpretation and decision making layer as well as the middle layer cognitive assistance telemonitoring Interpretation module module and decision making supervisor module layer middle layer hardware computer micro-wave bathroom door kitchen table glycometer layer because the code is distributed over all the processors that are included in the harware layer components 21 de 23

Outline Introduction (done) A layered computer infrastructure hardware layer code middleware layer code Interpretation and decision making code Computational and theoretical model and metamodel of the person the task the environment Implementation consideration Conclusion 22 de 23

Conclusion A pervasive computational infrastructure and applications A combination of cognitive assistance telemonitoring Models of person, activities and environment The implementation pervasive computing spontaneous networking distributed systems. A middleware and many frameworks 23 de 23