Le paradigme d'émergence spontanée de services (Spontaneous Emergence Paradigm, SEP) pour la mise en œuvre de l’informatique omniprésente ou diffuse (pervasive computing), consiste à fournir des services aux utilisateurs en fonction de leur environnement ad hoc. Ce paradigme a prouvé son efficacité et son adéquation aux applications distribuées sensibles au contexte. Ce travail se concentre sur deux aspects liés à la mise en œuvre des services SEP. Le premier se distingue de ce qui est proposé dans la littérature, par une approche physiologique de « Dead Reckoning » (Physiological Pedestrian Dead Reckoning Approach, PDR), qui permet de corriger les erreurs de positionnement à partir de données physiologiques comme la fréquence cardiaque et la vitesse de déplacement. Cette approche utilise des modèles stochastiques, qui reposent sur la distribution des probabilités des fréquences cardiaques précédemment collectées par rapport à la vitesse de déplacement de l’utilisateur (par exemple, par apprentissage via une marche sur un tapis roulant). Le modèle proposé permet ainsi d’ajuster les vitesses lues dans n’importe quel futur système PDR. Le deuxième aspect de cette recherche concerne la détection de réseaux d’affinités (Social Relationship Detection framework, SRD). Cette approche propose d’explorer le niveau d’affinité entre personnes en fonction de leur connexions et interactions selon plusieurs conditions. Ces connexions à des bornes WAP (Wireless Access Point), par rapport à chaque condition, sont consolidées pour former des distributions de probabilités. Par la suite, ces distributions de probabilités sont introduites dans un réseau de neurones (back propagation neural network) pour détecter et déterminer les affinités entre utilisateurs. Cette approche permet une prise de décision adaptative à chaque connexion au WAP et selon le comportement de l’utilisateur, contrairement aux approches classiques reposant uniquement sur une connaissance globale au préalable de la convergence des attributs des utilisateurs concernés. / Service Emergence Paradigm (SEP), based on a pervasive computing approach, consists of providing services to users within intended applications, transparently from their intentions, relatively to their needs, and regardless of the time and devices used. This considered paradigm has been proved sufficiently in applications that have context awareness schemes and various working threads. This research focuses on two aspects for providing SEP services. This first aspect is to implement a Physiological Pedestrian Dead Reckoning Approach (PPDR) that corrects position errors in pedestrian dead reckoning systems based on the heart rate parameter. This approach uses specific stochastic models that rely on probability distributions of previously collected heart rate values with respect to their velocities of a user during a treadmill walk. The reason behind this collection is to form a pattern that contributes in adjusting currently read velocities in any future PDR system. The second aspect of this research is the Social Relationship Detection framework (SRD). This approach proposes a mean for exploring the level of relationship between people by observing the network connection pattern of each person over several conditions. These connections, with respect to each condition, are consolidated to form a set that holds different probability distributions. These distributions are formed based on the users’ Wireless Access Point (WAP) connection establishment routine. Afterwards, the distributions of the probabilities are fed into a trained back propagation neural network to detect the level of relationship between candidate users for forthcoming friendship recommendations. This approach is elaborated in a technique that achieves intelligent decision-making whenever the connection pattern to the WAP and the user behavior are changing with time. It is also distinguished from other classical approaches that rely solely on prior knowledge of the convergence of the users’ attributes.

The Internet is a vastly under-utilised resource, only used for half of the IT story. Describe the Internet in two words and many might say ‘sharing knowledge’. But sharing information is more accurate. It’s just that all the principle ways we use the Internet — the Web, email and media streaming — happen to be examples where information is in the form of knowledge. But IT — Information Technology — has another side: the realm of software programming where information means the dynamic control of how things work. The Internet is the driving force in the IT industry, so why isn’t it also known for sharing control? True, there are examples of specialised, one-off software applications interfacing with each other via the Internet, but there has yet to be any systematic and universal attempt to exploit the potential of the Internet for control-IT in the way we have seen it for knowledge-IT. Taking the strengths of the Web model as a starting point, this thesis proposes a parallel, dynamic world to the Web called The Mesh. In the same way that the Web seamlessly connects databases of the world to provide a global font of knowledge, the Mesh would connect software of the world to provide a global means of control. The Mesh would embody all the successful, empowering features of the Web. Everyone would have a say in how things work, mirroring Web 2.0’s user-generated content but for software instead of media. In being a universally integrated design approach for device control, the Mesh would encompass a number of research areas working on the control issue at the big picture level. It would address the problems of universal usability and ubiquitous computing. It would also provide solutions in agent-based systems and grid computing. But many features of the Mesh would simply be unique. They would change the way we go about software design, leading to new opportunities for users, programmers and manufacturers alike. The key to everything is design simplicity. A concept demonstrator has been developed as an integral part of this research project. It shows that the Mesh is both feasible and practical. Examples of programs run in the concept demonstrator are discussed, showing exactly how the Mesh would be built and how it would work. / Thesis (Ph.D.) -- University of Adelaide, School of Electrical and Electronic Engineering, 2007

Researchers have approached the problem of designing context-aware systems, computer systems that can react intelligently to the context in which they are used, from both software engineering and human-centred perspectives. While engineered context-aware systems have found successful application in optimising technical settings without requiring mundane and explicit interaction from humans, challenges remain in furthering their application to more complex environments. In particular, when technical decisions and human actions are intertwined in sophisticated work environments (rather than, for example, the simple act of transferring a mobile phone call to receive a signal from the nearest tower), the problem of designing context-aware systems demands further examination and calls for a new approach. A natural extension of the software engineering approach in more challenging environments is to try to implement increasingly sophisticated algorithms for managing context in the hope that this will lead to autonomous application behaviour. In contrast, research from a human-centred perspective, grounded in the social sciences and philosophy, demonstrates that context is not entirely objective and measurable but rather is dynamic and created through human action. It may not be easily detected through technical sensing systems, and therefore simply employing more sophisticated algorithms within the technical sphere may not be effective. While engineering approaches have continued to evolve, the problem of how to design interfaces to context-aware applications still remains. The research reported in this dissertation investigated the problem of designing context-aware systems for the complex environment of a dental surgery. I undertook, though a participatory design approach (a) to better characterise the problem of designing context-aware systems, and (b) to understand how design methods could be employed to bridge the human and software engineering approaches. The gap in existing research on context-aware systems is evident in the way that the methods applied to designing systems don't provide an insight into how people actively create the context in which they work (in a practical rather than theoretical sense) they don‘t closely examine the behaviours of people, the role and arrangement of artefacts, and the dynamic relationships between people and artefacts. I found that an understanding of how these features of work and human behaviour are realised in practice in a given environment is fundamental to being able to design an effective context-aware system for that environment. The challenge is to design at the boundary between the technical and the social. The contribution of this thesis is an approach that explores context-aware design through synthesis. The synthetic approach leads to design opportunities and guidelines based on an understanding of the processes through which people actively co-create the context in which they work. I have applied and built upon a number of existing user-centred design and participatory design methods, in addition to creating some new methods in order to develop an understanding of how designers can examine the human aspects surrounding the co-creation of context and apply these in a way that progressively informs the design process. The methods collectively represent a novel approach to designing context-aware applications and differ from the more traditional technical approaches of developing software frameworks and infrastructures, and formal models of context, tasks, users, and systems. The techniques presented have focused primarily on developing an understanding of how humans find meaning in their actions along with their interaction with other people and technology. Participatory design methods help participants to reveal potential implicit technical resources that can be presented explicitly in technologies in order to assist humans in managing their interactions with and amidst technical systems gracefully. The methods introduced and the design approach proposed complement existing research on context-awareness from both a human-centred and software engineering perspective. This research builds on the notion of providing resources which allow users to manage their own context and also manage shifts in control while interacting with other people and with a variety of technical artefacts. It does this by examining a complex work environment, in particular looking at the kinds of resources people use and expect to use (and the constraints around these), the form(s) it is appropriate for them to take, and the patterns of interaction they will ultimately be used within.

Malgré des progrès récents, développer une application d'informatique ubiquitaire reste un défi à cause d'un manque de canevas conceptuels et d'outils aidant au développement. Ce défi implique de prendre en charge des objets communicants hétérogènes, de surmonter la complexité des technologies de systèmes distribués, de définir l'architecture d'une application, et d'encoder cela dans un programme. De plus, tester des applications d'informatique ubiquitaire est problématique car cela implique d'acquérir, de tester et d'interfacer une variété d'entités logicielles et matérielles. Ce procédé peut rapidement devenir coûteux en argent et en temps lorsque l'environnement ciblé implique de nombreuses entités. Cette thèse propose une méthodologie outillée pour dévelop- per et tester des applications d'informatique ubiquitaire. Notre méthodologie fournit tout d'abord le langage de conception DiaSpec. Ce langage permet de définir une taxonomie d'entités spécifiques à un domaine applicatif, s'abstrayant ainsi de leur hétérogénéité. Ce langage inclut également une couche permettant de définir l'architecture d'une application. Notre suite outillée fournit un compilateur qui, à partir de descriptions DiaSpec, génère un canevas de programmation guidant les phases d'implémentation et de test. Afin d'aider à la phase de test, nous proposons une approche de simulation et un outil intégré dans notre méthodologie outillée : l'outil DiaSim. Notre approche utilise le support de test généré par DiaSpec pour tester les applications de manière transparente dans un environnement physique simulé. La simulation d'une application est rendue graphiquement dans un outil de visualisation 2D. Nous avons combiné DiaSim avec un langage dédié permet- tant de décrire les phénomènes physiques en tant qu'équations différentielles, permettant des simulations réalistes. DiaSim a été utilisé pour simuler des applications dans des domaines applicatifs variés. Notre approche de simulation a également été appliquée à un système avionique, démontrant la généralité de notre approche de simulation.

Ambient intelligence refers to the vision of computationally augmented everyday environments that are sensitive, adaptive and responsive to humans and intelligently support their daily lives. Ambient ecologies are the infrastructures of ambient intelligence. To enable system developers to frame and manage the dynamic and complex interaction of humans with ambient ecologies consisting of a mixture of physical (real) and virtual (digital) objects, novel interaction paradigms are needed. Traditional interaction paradigms like the WIMP (windows, icon, menus, and pointing devices) paradigm for desktop computing operate in a closed world, unaware of the physical, social and cultural context. They restrict human perception and action to screen, mouse and keyboard with the assumption that human attention will be fully devoted to interaction with the computer. Emerging interaction paradigms for ambient intelligence are typically centered on specific devices, specific computing environments or specific human capabilities. Also, many of them are driven by technological advancements rather than viewing the human agent as their starting point. A principled, theoretical approach centered in the individual human agent, their situation and activities that are comprehensive and integrated while at the same time instrumental in the design of ambient ecologies has been lacking. This thesis introduces egocentric interaction as an approach towards the modeling of ambient ecologies with the distinguishing feature of taking the human agent’s body, situation and activities as center of reference, as opposed to the more common device-centric approaches in facilitating human-environment interaction. Egocentric interaction is encapsulated in a number of assumptions and principles such as situatedness, the proximity principle, the physical-virtual equity principle, perception and action instead of “input” and “output,” and activity-centeredness. A situative space model is proposed based on some of these principles. It is intended to capture what a specific human agent can perceive and not perceive, reach and not reach at any given moment in time. The situative space model is for the egocentric interaction paradigm what the virtual desktop is for the WIMP interaction paradigm: more or less everything of interest to a specific human agent is assumed and supposed to happen here. In addition, the conception and implementation of the easy ADL ecology based on egocentric interaction, comprising of smart objects, a personal activity-centric middleware, ambient intelligence applications aimed at everyday activity support, and a human agent literally in the middle of it all is described. The middleware was developed to address important challenges in ambient intelligence: (1) tracking and managing smart objects; (2) tracking a human agent’s situative spaces; (3) recognizing human activities and actions; (4) managing and facilitating human-environment interaction; and (5) to ease up the development of ambient intelligence applications. The easy ADL ecology was first simulated in immersive virtual reality, and then set up physically as a living laboratory to evaluate: (1) the technological and technical performance of individual middleware components, (2) to perform a user experience evaluation assessing various aspects of user satisfaction in relation to the support offered by the easy ADL ecology, and (3) to use it as a research test bed for addressing challenges in ambient intelligence. While it is problematic to directly compare the “proof-of-concept” easy ADL ecology with related research efforts, it is clear from the user experience evaluation that the subjects were positive with the services it offered. / easy ADL project

This thesis investigates the design of sociable technologies and is divided into three main parts described below. In the first part, we introduce sociable technologies. We review our the definition of technology and propose categories of technologies according to the motivation underlying their design: improvement of control, improvement of communication or improvement of cooperation. Sociable technologies are then presented as an extension of techniques to improve cooperation. The design of sociable technologies are then discussed leading to the observation that acquisition of social common sense is a key challenge for designing sociable technologies. Finally, polite technologies are presented as an approach for acquiring social common sense. In the second part, we focus on the premises for the design of sociable technologies. A key aspect of social common sense is the ability to act appropriately in social situations. Associating appropriate behaviour with social situations is presented as a key method for implementing polite technologies. Reinforcement learning is proposed as a method for learning such associations and variation of this algorithm are experimentally evaluated. Learning the association between situation and behaviour relies on the strong assumption that mutual understanding of social situations can be achieved between technologies and people during interaction. We argue that in order to design sociable technologies, we must change the model of communication used by our technologies. We propose to replace the well-known code model of communication, with the ostensive-inferential model proposed by Sperber and Wilson. Hypotheses raised by this approach are evaluated in an experiment conducted in a smart environment, where, subjects by group of two or three are asked to collaborate with a smart environment in order to teach it how to behave in an automated meeting. A novel experimental methodology is presented: The Sorceress of Oz. The results collected from this experiment validate our hypothesis and provide insightful information for the design. We conclude by presenting, what we believe are, the premises for the design of sociable technologies. The final part of the thesis concerns an infrastructure for the design of sociable technologies. This infrastructure provides the support for three fundamental components. First, it provides the support for an inferential model of context. This inferential model of context is presented; a software architecture is proposed and evaluated in an experiment conducted in a smart-environment. Second, it provides the support for reasoning by analogy and introduces the concept of eigensituations. The advantage of this representation are discussed and evaluated in an experiment. Finally, it provides the support for ostensive-inferential communication and introduces the concept of ostensive interface.

En France, comme dans l'ensemble des pays développés, le même constat est dressé : la population vieillit. Cette évolution démographique sans précédent dans l'histoire de l'humanité pose un certain nombre d'enjeux économiques et sociaux liés à la prise en charge des personnes âgées dépendantes et/ou en situation de handicap, dont le nombre augmente considérablement au fil des ans. Le nombre d'aidants potentiels ne pouvant augmenter en conséquence, il faut penser dès aujourd'hui à des solutions alternatives pour permettre à ces personnes de garder un niveau d'autonomie suffisant leur permettant de rester dans leur environnement domestique préféré le plus longtemps possible. Les systèmes domotiques, qui regroupent l'ensemble des technologies de l'informatique, de l'électronique et des télécommunications au sein d'un habitat, permet d'y apporter des solutions technologiques. Dans ces travaux, nous proposons un flot de conception dirigé par les modèles pour la commande et la supervision de systèmes domotiques d'assistance. Dans un premier temps, nous cherchons à intégrer une personne non-experte en domotique (e.g. membre de la famille / aidant, ergothérapeute) dans certaines étapes de la conception pour une meilleure prise en compte des besoins de personnes en situation de handicap mais aussi pour limiter l'intervention relativement coûteuse de l'expert. Pour cela, nous proposons un langage de modélisation (DSML : Domain Specific Modeling Language) permettant de décrire un système domotique en manipulant des concepts à un haut niveau d'abstraction. Ce DSML est composé de deux vues : une vue physique pour représenter l'environnement domotique et une vue fonctionnelle pour définir les interactions entre l'utilisateur et son système domotique. Dans cette dernière vue, deux types d'interactions sont définis pour prendre en compte un nombre de besoins utilisateur le plus grand possible. Ainsi, il est possible de définir des requêtes de différents types (i.e. services) comme des interdictions temporaires (i.e. modes) sur des ressources domotiques. À partir des modèles définis à l'aide de ce langage, notre flot de conception permet de générer la partie commande à l'aide de transformations de modèles successives. Une évaluation sur le cas d'étude des appartements tremplins du centre de Kerpape avec des étudiants puis des ergothérapeutes permet d'avoir des retours intéressants sur l'utilisabilité du DSML et du flot de conception proposé. Dans un deuxième temps, nous cherchons à offrir une solution alternative aux approches de supervision basées sur l'utilisation massive de capteurs, qui peut parfois être problématique lorsque l'on s'intéresse à l'acceptabilité de la solution par l'utilisateur (coût, intrusivité). Nous proposons pour cela d'intégrer dans le flot de conception un modèle de supervision non-intrusif basé sur une ontologie pour représenter les interactions utilisateur-système. Contrairement aux solutions existantes, ce modèle n'est pas basé sur des données de capteurs, mais sur l'analyse sémantique des commandes domotiques envoyées par l'utilisateur. L'ontologie et les règles qui lui sont associées permettent à un moteur d'inférences d'extraire des informations habituellement captées : suivant la manière dont l'utilisateur interagit avec son système domotique, sa localisation et son intention peuvent être déduites par inférences logiques, permettant d'aboutir à la proposition de services contextuels. Une évaluation sur les jeux de données de l'appartement Domus permet de montrer l'intérêt d'une telle approche tant pour limiter l'utilisation de capteurs que pour augmenter la sûreté de fonctionnement par redondance d'informations. Une partie de ces travaux est reprise dans le cadre du projet ASIM porté par l'entreprise Vity Technology. La syntaxe concrète (i.e. représentation graphique) du langage de modélisation sera notamment retravaillé dans un but de commercialisation.

Highly distributed environments (HDEs) are deployment environments that include powerful and robust machines in addition to resource-constrained and mobile devices such as laptops, personal digital assistants (or PDAs), smart-phones, GPS devices, sensors, etc. Developing software for HDEs is fundamentally different from the software development for central systems and stable distributed systems. This argument is discussed deeply and in-details throughout this dissertation. HDE applications are challenged by two problems: unreliable networks, and heterogeneity of hardware and software. Both challenges need careful handling, where the system must continue functioning and delivering the expected QoS. This dissertation is a direct response to the mentioned challenges of HDEs. The contribution of this dissertation is the cloud component model and its related formal language and tools. This is the general title. However, and to make this contribution clear, we prefer to present it in the following detailed form: (1) We propose a paradigm shift from distribution transparency to localization acknowledgment being the first class concern. (2) To achieve the above mentioned objective, we propose a novel component model called cloud component (CC). (3) In this dissertation we propose a new approach to assemble CCs using systematic methodology that maintains the properties of CC model. (4) Cloud component development process and cloud component based systems development process. (5) Location modeling and advanced localization for HDEs are the pivotal key in our contribution. (6) Formal language to model single CC, CC assembly, CC development process, and CC based systems. (7) We finally present our fully-developed supporting tools: the cloud component management system CCMS, and the Registry utility. To respond to the challenges posed by HDEs, and to maintain expected software quality at the user endpoint, we think we need to pass a ¿paradigm shift¿ from the way software is designed and implemented currently to our new vision that this dissertation is devoted to. This is a paradigm shift from distribution transparency to localization acknowledgment being the first class concern. The contribution in this thesis has several faces as explained above, but still, these faces are cohesive. Each of these faces form a partial contribution, however, this partial contribution does not mean anything if isolated from the overall proposal. Moreover, the merit of the overall proposal can not be grasped by reading one partial contribution. The merit of the proposal is evident only if all parts of this work are cohesively organized. Finally, we claim that our proposal spans the entire software development process for HDEs, from requirements to deployment and runtime management.

The dream of weaving technology into our everyday fabric of life is recently being made possible by advances in ubiquitous computing and sensor technologies. Countless sensors of various sizes have made their way into everyday commercial applications. Many projects aim to explore new ways to utilize these new technologies to aid and interact with the general population. Context-aware systems use available context information to assist users automatically, without explicit user input. By inferring user intent and configuring the system proactively for each user, context-aware systems are an integral part of achieving user-friendly ubiquitous computing environments. <br /><br /> A common issue with building a distributed context-aware system is the need to develop a supporting infrastructure providing features such as storage, distributed messaging, and security, before the real work on processing context information can be done. This thesis proposes a coordination framework that provides an effective common foundation for context-aware systems. The separation between the context-processing logic component and the underlying supporting foundation allows researchers to focus their energy at the context-processing part of the system, instead of spending their time re-inventing the supporting infrastructure. <br /><br /> As part of an ongoing project, Impress, the framework uses the open standard, Jabber, as its communication protocol. The Publish-Subscribe (pubsub) extension to Jabber provides interesting features that match those needed by a context-aware system. The main contribution of this thesis is the design and implementation of a coordination framework, called the Impress Context Store, that provides an effective common foundation for context-aware systems. The separation between the context-processing logic and the underlying supporting foundation allows researchers to focus their energy at the context-processing part of the system, instead of spending their time re-inventing the supporting infrastructure.

The dream of weaving technology into our everyday fabric of life is recently being made possible by advances in ubiquitous computing and sensor technologies. Countless sensors of various sizes have made their way into everyday commercial applications. Many projects aim to explore new ways to utilize these new technologies to aid and interact with the general population. Context-aware systems use available context information to assist users automatically, without explicit user input. By inferring user intent and configuring the system proactively for each user, context-aware systems are an integral part of achieving user-friendly ubiquitous computing environments. <br /><br /> A common issue with building a distributed context-aware system is the need to develop a supporting infrastructure providing features such as storage, distributed messaging, and security, before the real work on processing context information can be done. This thesis proposes a coordination framework that provides an effective common foundation for context-aware systems. The separation between the context-processing logic component and the underlying supporting foundation allows researchers to focus their energy at the context-processing part of the system, instead of spending their time re-inventing the supporting infrastructure. <br /><br /> As part of an ongoing project, Impress, the framework uses the open standard, Jabber, as its communication protocol. The Publish-Subscribe (pubsub) extension to Jabber provides interesting features that match those needed by a context-aware system. The main contribution of this thesis is the design and implementation of a coordination framework, called the Impress Context Store, that provides an effective common foundation for context-aware systems. The separation between the context-processing logic and the underlying supporting foundation allows researchers to focus their energy at the context-processing part of the system, instead of spending their time re-inventing the supporting infrastructure.