Download Framework for ambient assistive living PDF

TitleFramework for ambient assistive living
File Size3.1 MB
Total Pages204
Table of Contents
                            Page de garde
Page de garde
Author's Publications
Table of Contents
	High-Level Problematic
		Ageing Population and Dependency
		Cognitive Deficiency
			Normal Ageing Cognitive Decline
			Dementia as a Pathological Cognitive Decline
	Intelligent Environments
		Ambient Intelligence
		Ambient Assisted Living
	Ambient Assistive Living Systems for Cognitive Decline
		Gator Tech Smart House
		SIMBAD System
		CSCC System
		ISISEMD Project
		ROSETTA Project
	Requirements of Ambient Assisted Living Systems
		Context Awareness
		Multiple People Management
		Design for failure
		Dynamism and Adaptability of the system
		Handling Uncertainty
	Thesis Positioning
	Thesis Outline
The State of the Art
	Overview of AAL Dynamism
	Service Model
		Modular Approach vs. Monolithic Approach
		The Service Oriented Approach
			Web Service
			The Open Service Gateway Initiative
	Services Interoperability
		Service Exchange
			Local Service Exchange
			Distributed Service Exchange
		Event Exchange
			Local Event Exchange
			Distributed Event Exchange
		The Device Profile for Web Services Communication Mechanism
	Knowledge Modeling & Reasoning
		The Declarative Approach vs. the Imperative Approach
		Semantic Web Technologies
			Semantic Modeling
			Semantic Reasoning
	Uncertainty Handling
Research Approach & System Design
	Vision of Ambient Assistive Living Systems
	The UbiSMART Framework
		Proposed UbiSMART Architecture
		Architecture Based on Distributed Event Communication
	Handling Dynamism in the UbiSMART framework
		Architecture Based on Device Profile for Web Services
		The Semantic Plug&Play Mechanism
			Assistive Service Semantic Plug&Play
			Sensors/Devices Semantic Plug&Play
		Semantic Context Modeling and Reasoning
	Validation of the Semantic Plug&Play Mechanism
Uncertainty Handling
	Uncertainty Measurement
	Semantic Modeling under Uncertainty
	Reasoning under Uncertainty
		Probabilistic Approach
		Bayesian Inference
		Possibility Theory: Fuzzy Logic
		Evidence Theory: Dempster-Shafer Theory
		Decision-Supporting Mechanism
	Validation of the Uncertainty Handling Approach
	Design & Development
		Components Diagram
		Sequence Diagram for Sensors Discovery and Configuration
		Sequence Diagram for Context Update and Mass Function Calculation
		Sequence Diagram for Decision Making and Service Selection and Provision
		Class Diagram
Deployment and Validation
	Rational of our Deployment Approach
	AMUPADH Project: A Top/Down Deployment Approach
	Research Approach to Real Life Deployment
		Choice of Deployment Environment
		Pre-deployment Observations and Discussions
		Participants' Characteristics and Selection Process
		Deployed System
		Data Gathering & Performance Evaluation
	Outcomes of the Pre-deployment Analysis
	Deployed Services
	System Performance
	Data Analysis
	Qualitative Feedback
	Discussion & Lessons Learned
List of appendices
	GDS for assessment of primary degenerative dementia
	Overview of Ambient Assisted Living
	The DPWS based semantic Plug&Play
	Semantic Modeling and Reasoning
	Uncertainty Handling
List of Figures
List of Tables
Document Text Contents
Page 1

HAL Id: tel-01048706

Submitted on 25 Jul 2014

HAL is a multi-disciplinary open access
archive for the deposit and dissemination of sci-
entific research documents, whether they are pub-
lished or not. The documents may come from
teaching and research institutions in France or
abroad, or from public or private research centers.

L’archive ouverte pluridisciplinaire HAL, est
destinée au dépôt et à la diffusion de documents
scientifiques de niveau recherche, publiés ou non,
émanant des établissements d’enseignement et de
recherche français ou étrangers, des laboratoires
publics ou privés.

Framework for ambient assistive living : handling
dynamism and uncertainty in real time semantic services

Hamdi Aloulou

To cite this version:
Hamdi Aloulou. Framework for ambient assistive living : handling dynamism and uncertainty in real
time semantic services provisioning. Automatic Control Engineering. Institut National des Télécom-
munications, 2013. English. �NNT : 2013TELE0016�. �tel-01048706�

Page 102

4.4. Reasoning under Uncertainty 79

its certainty level needs to be calculated from the certainty levels of the values used

in the antecedent of the rule [97]. This approach has been adopted in our framework,

where the uncertainty is propagated from sensors level to high-level context information

using semantic rules. Next, I provide the transition of a semantic rule from a classical

representation to a new one that transfers uncertainty from sensors level to context level.

This rule is used for localization tracking of the end-user.

Classical representation:

∀ Sensor se; SensorState st; Room r; House h; User u

(se hasCurrentState st) ∧ (se hasLastUpdate true) ∧ (se deployedIn r)

∧ (r partOf h) ∧ (u liveIn h) ⇒ (u detectedIn r)

Uncertainty inclusion:

∀ Sensor se; SensorState st; Room r; House h; User u

(se hasCurrentState [ a Uncertainty; uncertaintyLevel n; relatedObject st])

∧ (se hasLastUpdate true) ∧ (se deployedIn r) ∧ (r partOf h) ∧ (u liveIn h)

⇒ (u detectedIn [ a Uncertainty; uncertaintyLevel n; relatedObject r; accordingTo se])

We believe that the uncertainty aspect will not be tackled only by the engine itself,

but it is rather the way the engine is used and coupled with other techniques that can ever

address this aspect. Especially when aiming at data fusion or higher-ordered reasoning in

AAL systems.

Diverse formalisms exist to deal with uncertainty including probability value assign-

ments, and degrees of set membership for vagueness e.g., probabilistic approach [98],

Bayesian reasoning [92], Dempster-Shafer techniques [99], and fuzzy logic [100]. These

mechanisms can be coupled with semantic web reasoning.

4.4.1 Probabilistic Approach

Probability theory is the first uncertainty management technique to be introduced. This

theory seeks to judge the probability measure for an event Ai given a proposed hypothesis

Hi such that:

0 ≤ Pr(Ai|Hi) ≤ 1

Decision making in probabilistic approach could be realized using different rules. For

example, the likelihood comparison rule suggests accepting the hypothesis Hi if the prob-

ability relationship satisfies the equation:

Page 103

80 Chapter 4. Uncertainty Handling

P (Ak|Hi).P (Hi) > P (Ak|¬Hi).P (¬Hi)

One principal limitation of the traditional probability theory’s characterization of un-

certainty is its incapability of capturing epistemic uncertainty. The application of tra-

ditional probabilistic methods to epistemic or subjective uncertainty is often known as

Bayesian probability [99]. In addition, a probabilistic analysis requires an analyst to have

information on the probability of all events. When this is not available, the uniform distri-

bution function is often used to affect an equal value to all events for which a probability

distribution is not known in a given sample space. This is not totally true. For example, if

we have three sensors in the environment, we assign a probability of failure with 0.4 to one

of them and we are ignoring the probability of the two others. This does not mean that

they have a probability of failure of 0.3 each. Another assumption in classical probability

is that the knowledge of the likelihood of the occurrence of an event can be translated

into the knowledge of the likelihood of that event not occurring. Once again, if we believe

that the system may fail due to the first sensor with a probability of 0.4, this does not

necessarily mean that we believe that the system will not fail due to that sensor with a

probability of 0.6.

Though the assumptions of additivity and the principle of insufficient reason may be

appropriate when modeling the random events associated with aleatory uncertainty, these

constraints are questionable when applied to an issue of knowledge or belief, especially

when information on which to evaluate a probability is limited, ambiguous or conflicting.

It could be reasonable to consider probability measurement as an interval or a set when

a precise probability characterizing the uncertainty is not available. This has three major


• We do not have to give a precise measure of uncertainty if it is not realistic or feasible

to do so.

• Uncertainty could be affected to multiple events together without having to give

assumption about events under ignorance.

• Measure of uncertainty does not have to add up to 1. If it is less than 1, this

means that there is incompatibility between multiple sensors providing conflicting

information. If it is greater than 1, this implies a cooperative effect between multiple

sensors providing the same information.

Page 204


Ageing, 3

Ambient Assisted Living, 14

Ambient Intelligence, 13

Apache Felix, 35

Bayesian Inference, 81

Concierge, 34

Context Awareness, 18

Declaritave Approach, 39

Dementia, 9

Deployment Research Approach, 103

Device Profile for Web Services, 37

Distributed OSGi, 36

Equinox, 35

Event Exchange, 36

Evidence Theory, 82

Gerontechnology, 11

Intelligent Environments, 12

Knopflerfish, 34

Modular Approach, 28

Normal Ageing, 7

Open Service Gateway Initiative, 32

Pathological Ageing, 9

Possibility Theory, 81

Probabilistic Approach, 79

Semantic Modeling, 40

Semantic Modeling & Uncertainty, 76

Semantic Modelling, 61

Semantic Plug&Play, 57

Semantic Reasoning, 41, 61

Semantic Reasoning & Uncertainty, 78

Service Exchange, 36

Service Oriented Approach, 29

UbiSMART Framework, 51

Uncertainty, 42

Uncertainty Measurement, 73

Web Service, 31

Similer Documents