Cette thèse s’intéresse à la problématique de l’adaptation automatique de services dans ledomaine de l’intelligence ambiante. L’étude de la littérature montre que la sensibilité aucontexte est devenue un élément central pour la conception et la mise en place de servicesadaptatifs. Cependant, sa prise en compte se limite généralement à des descriptionsélémentaires de situations ou à des modèles prédéfinis. Afin de permettre une adaptation auxchangements d’habitudes des utilisateurs, à la dynamique de l’environnement et àl’hétérogénéité des sources de perception, nous proposons des mécanismes de découverte decontexte et de situations déclencheurs d’adaptation. Ces mécanismes s’appuient sur destechniques de fouille de données et sont intégrés au sein d’une architecture d’adaptationautomatique de services. Ces travaux ont été réalisés et appliqués à des projets d’intelligenceambiante pour de l’assistance à des personnes et plus particulièrement dans le cadre du projetITEA- MIDAS. / This thesis addresses the problem of dynamic adaptation of services in the context of ambientintelligence applications. Literature study shows how context-awareness plays a central rolein the design and implementation of adaptive services. However, its use is still limited toelementary descriptions and predefined situational models. Dynamic adaptation should becapable of following user habits to yield dynamic answers to environmental change, and tosupport heterogeneous sources of context. To this end, we propose mechanisms to discovercontexts and situations that trigger adaptation. These mechanisms rely on data miningtechniques, and are integrated within an architecture for dynamic adaptation of services. Thiswork was carried out and applied to ambient intelligence projects for the elderly, providingsupport and assistance in their daily lives, particularly in the context of the ITEA-MIDASproject.

L'instrumentation permet à notre environnement, maison ou bâtiment, de devenir intelligent en s'adaptant à nos modes de vie et en nous assistant au quotidien. Un environnement intelligent est sensible et réactif à nos activités, afin d'améliorer notre qualité de vie. La fiabilité d'identification des activités est ainsi essentielle pour cette intelligence ambiante : elle est directement dépendante du positionnement des capteurs au sein de l'environnement. Cette question essentielle du placement des capteurs est très peu considérée par les systèmes ambiants commercialisés ou même dans la littérature. Pourtant, elle est la source principale de leurs dysfonctionnements où une mauvaise reconnaissance des activités entraîne une mauvaise assistance fournie. Le placement de capteurs consiste à choisir et à positionner des capteurs pertinents pour une identification fiable des activités. Dans cette thèse, nous développons et détaillons une méthodologie de placement de capteurs axée sur l'identifiabilité des activités d'intérêt. Nous la qualifions en nous intéressant à deux évaluations différentes : la couverture des intérêts et l'incertitude de mesures. Dans un premier temps, nous proposons un modèle de l'activité où nous décomposons l'activité en actions caractérisées afin d'être indépendant de toute technologie ambiante (axée connaissances ou données). Nous représentons actions et capteurs par un modèle ensembliste unifiant, permettant de fusionner des informations homogènes de capteurs hétérogènes. Nous en évaluons l'identifiabilité des actions d'intérêt au regard des capteurs placés, par des notions de précision (performance d'identification) et de sensibilité (couverture des actions). Notre algorithme de placement des capteurs utilise la Pareto-optimalité pour proposer une large palette de placements-solutions pertinents et variés, pour ces multiples identifiabilités à maximiser. Nous illustrons notre méthodologie et notre évaluation en utilisant des capteurs de présence, et en choisissant optimalement la caractéristique à couvrir pour chaque action. Dans un deuxième temps, nous nous intéressons à la planification optimale des expériences où l'analyse de la matrice d'information permet de quantifier l'influence des sources d'incertitudes sur l'identification d'une caractéristique d'action. Nous représentons les capteurs continus et l'action caractérisée par un modèle analytique, et montrons que certaines incertitudes doivent être prises en compte et intégrées dans une nouvelle matrice d'information. Nous y appliquons les indices d'observabilité directement pour évaluer l'identifiabilité d'une action caractérisée (incertitude d'identification). Nous illustrons cette évaluation alternative en utilisant des capteurs d'angle, et nous la comparons à la matrice d'information classique. Nous discutons des deux évaluations abordées et de leur complémentarité pour la conception d’environnement instrumenté pour la veille à la personne. / Instrumentation enables our environment, house or building, to get smart through self-adjustment to our lifestyles and through assistance of our daily-life. A smart environment is sensitive and responsive to our activities, in order to improve our quality of life. Reliability of activities' identification is absolutely necessary to such ambient intelligence: it depends directly on sensors' positioning within the environment. This fundamental issue of sensor placement is hardly considered by marketed ambient systems or even into the literature. Yet, it is the main source of ambient systems' malfunctions and failures, because a bad activity recognition leads to a bad delivered assistance. Sensor placement is about choosing and positioning relevant sensors for a reliable identification of activities. In this thesis, we develop and detail a sensor placement methodology driven by identifiability of activities of interest. We quantify it by looking at two different evaluations: coverage of interests and uncertainty of measures. First, we present an activity model that decomposes each activity into characterised actions to be technology-free (either knowledge or data driven one). We depict actions and sensors by a set theoretic model, enabling to fuse homogeneous informations of heterogeneous sensors. We then evaluate each action of interest's identifiability regarding placed sensors, through notions of precision (identification's performance) and sensitivity (action's coverage). Our sensor placement algorithm use Pareto-optimality to offer a wide range of relevant solution-placements, for these multiple identifiabilities to maximise. We showcase our methodology and our evaluation through solving a problem featuring motion and binary sensors, by optimally choosing for each action the characteristic to cover. Finally, we look into optimal design of experiments by analysing the information matrix to quantify how sources of uncertainties influence the identification of an action's characteristic. We depict continuous sensors and the characterised action by an analytical model, and we show that some uncertainties should be considered and included in a new information matrix. We then apply directly observability indexes to evaluate identifiability of a characterised action (uncertainty of identification), and compare our new information matrix to the classical one. We showcase our alternate evaluation through solving a sensor placement problem featuring angular sensors. We discuss both covered evaluations and their complementarity towards the design of instrumented environment for human monitoring.

Tirant parti de notre expérience avec une approche traditionnelle des environnements d'assistance ambiante (AAL) qui repose sur l'utilisation de nombreuses technologies hétérogènes dans les déploiements, cette thèse étudie la possibilité d'une approche simplifiée et complémentaire, ou seul un sous-ensemble hardware réduit est déployé, initiant un transfert de complexité vers le côté logiciel. Axé sur les aspects de raisonnement dans les systèmes AAL, ce travail a permis à la proposition d'un moteur d'inférence sémantique adapté à l'utilisation particulière à ces systèmes, répondant ainsi à un besoin de la communauté scientifique. Prenant en compte la grossière granularité des données situationnelles disponible avec une telle approche, un ensemble de règles dédiées avec des stratégies d'inférence adaptées est proposé, implémenté et validé en utilisant ce moteur. Un mécanisme de raisonnement sémantique novateur est proposé sur la base d'une architecture de raisonnement inspiré du système cognitif. Enfin, le système de raisonnement est intégré dans un framework de provision de services sensible au contexte, se chargeant de l'intelligence vis-à-vis des données contextuelles en effectuant un traitement des événements en direct par des manipulations ontologiques complexes. L’ensemble du système est validé par des déploiements in-situ dans une maison de retraite ainsi que dans des maisons privées, ce qui en soi est remarquable dans un domaine de recherche principalement cantonné aux laboratoires / Leveraging our experience with the traditional approach to ambient assisted living (AAL) which relies on a large spread of heterogeneous technologies in deployments, this thesis studies the possibility of a more “stripped down” and complementary approach, where only a reduced hardware subset is deployed, probing a transfer of complexity towards the software side, and enhancing the large scale deployability of the solution. Focused on the reasoning aspects in AAL systems, this work has allowed the finding of a suitable semantic inference engine for the peculiar use in these systems, responding to a need in this scientific community. Considering the coarse granularity of situational data available, dedicated rule-sets with adapted inference strategies are proposed, implemented, and validated using this engine. A novel semantic reasoning mechanism is proposed based on a cognitively inspired reasoning architecture. Finally, the whole reasoning system is integrated in a fully featured context-aware service framework, powering its context awareness by performing live event processing through complex ontological manipulation. the overall system is validated through in-situ deployments in a nursing home as well as private homes over a few months period, which itself is noticeable in a mainly laboratory-bound research domain

Social contact is known to be vital for older adults' mental and physical health, but few studies of social interactions have taken place in long-term care settings. The current study investigated whether the psychological well-being of assisted living residents was influenced by factors associated with residents' social interactions involving humor. Specific aims of the present study were to develop and test a measure related to humor-related social exchanges, to examine how humor-related social exchanges affect residents' mental health, and to explore whether humor-related social exchanges mediated the effects of resident and facility characteristics on indices of mental health. One hundred and forty older adults residing in 14 assisted living facilities in the Portland, Oregon metropolitan area were interviewed about the frequency and types of social interactions they experienced with members of their facility-based social networks, as well as depression, mood, loneliness, self-esteem, and self-rated health. A 12-item, two-factor model of humor-related social exchanges was identified through confirmatory factor analysis, including both positive and negative humor-related social exchange factors. The newly developed scale displayed evidence of adequate reliability and validly in the current sample. Results indicated that both positive and negative humor-related exchanges were associated with various aspects of mental health, although negative humor-related exchanges appeared to be a stronger predictor of mental health than positive humor-related exchanges. Both positive and negative humor-related exchanges also served as mediators between resident and facility characteristics and indicators of mental health. Cultivating a better understanding of the relationships between humor-related social exchanges and mental health may be beneficial for researchers interested in the way humor impacts older adults' ability to cope with stress. This research may also be of value to long-term care providers who create interventions designed at improving residents' mental health and overall quality of life.

No description available.

No / OBJECTIVES: The Enriched Opportunities Programme (EOP) is a multi-level intervention focussing on improved quality of life for people with dementia. This study compared the experience of people living with dementia and other mental health problems in extra care housing schemes that utilised EOP with schemes that employed an active control intervention. METHOD: Ten extra care housing schemes were cluster randomised to receive either the EOP intervention or an active control intervention for an 18-month period. Residents with dementia or other significant mental health problems (20-30 per scheme) were assessed on a number of outcome measures at baseline, six months, one year and 18 months. The primary outcome measure was quality of life. Self-reported depression was an important secondary outcome. RESULTS: The EOP-participating residents rated their quality of life more positively over time (4.0 (SE 0.6) units; 14% p < 0.001) than the active control (1.3 (SE 0.6) units; 4% p = 0.003). There was also a significant group-time interaction for depressive symptoms (p = 0.003). The EOP-participating residents reported a reduction of 25% at both six and 12 months and a 37% reduction at 18 months (all p's < 0.001). EOP residents were less likely than residents in the active control sites to move to a care home or to be admitted to a hospital inpatient bed. They were more likely to be seen by a range of community health professionals. CONCLUSION: The EOP had a positive impact on the quality of life of people with dementia in well-staffed extra care housing schemes.

Recent years have seen an influx of medical technologies capable of remotely monitoring the health and behaviours of individuals to detect, manage and prevent health problems. Known collectively as personal health monitoring (PHM), these systems are intended to supplement medical care with health monitoring outside traditional care environments such as hospitals, ranging in complexity from mobile devices to complex networks of sensors measuring physiological parameters and behaviours. This research project assesses the potential ethical implications of PHM as an emerging medical technology, amenable to anticipatory action intended to prevent or mitigate problematic ethical issues in the future. PHM fundamentally changes how medical care can be delivered: patients can be monitored and consulted at a distance, eliminating opportunities for face-to-face actions and potentially undermining the importance of social, emotional and psychological aspects of medical care. The norms evident in this movement may clash with existing standards of 'good' medical practice from the perspective of patients, clinicians and institutions. By relating utilitarianism, virtue ethics and theories of surveillance to Habermas' concept of colonisation of the lifeworld, a conceptual framework is created which can explain how PHM may be allowed to change medicine as a practice in an ethically problematic way. The framework relates the inhibition of virtuous behaviour among practitioners of medicine, understood as a moral practice, to the movement in medicine towards remote monitoring. To assess the explanatory power of the conceptual framework and expand its borders, a qualitative interview empirical study with potential users of PHM in England is carried out. Recognising that the inherent uncertainty of the future undermines the validity of empirical research, a novel epistemological framework based in Habermas' discourse ethics is created to justify the empirical study. By developing Habermas' concept of translation into a procedure for assessing the credibility of uncertain normative claims about the future, a novel methodology for empirical ethical assessment of emerging technologies is created and tested. Various methods of analysis are employed, including review of academic discourses, empirical and theoretical analyses of the moral potential of PHM. Recommendations are made concerning ethical issues in the deployment and design of PHM systems, analysis and application of PHM data, and the shortcomings of existing research and protection mechanisms in responding to potential ethical implications of the technology.

I mitt examensarbete har jag undersökt min prekära situation som konsthantverkare och hemtjänstpersonal. Arbetet innehåller reflektioner från den prekära situation det innebär att arbeta i hemtjänsten och att vara brukare. Jag belyser mina medvetna och omedvetna översättningsprocesser från hemtjänsten till mitt konsthantverkskap. Arbetsprocessen i verkstaden binds ihop med prekariatets mindre synliga delar. Identitet, tillhörighet och gemenskap.

Die zunehmende Überalterung der Bevölkerung in Europa und speziell in Deutschland stellt die Gesellschaft vor große personelle und finanzielle Herausforderungen. Diese Problematik wird im Forschungsfeld Ambient Assisted Living (AAL) adressiert. AAL hat die Zielsetzung, ältere Menschen bei Aktivitäten des täglichen Lebens zu unterstützen. Ein wesentliches Element hierbei ist die Vernetzung verschiedener Systeme im Haushalt, sodass das Verhalten einer Person erfasst werden kann. Angehörige und medizinisches Personal können somit im Notfall automatisiert verständigt werden. Hierfür ist das Monitoring von Vitalparametern ein wichtiger Baustein, da damit der momentane Gesundheitszustand einer Person erfasst und analysiert werden kann. Konventionelle Systeme zur Bestimmung von Vitalparametern erfordern eine direkte Applizierung am Körper und werden häufig als störend empfunden. Verfahren, welche auf einer kontaktlosen, optischen Erkennung von Vitalparametern mittels Bildverarbeitung basieren, können diese Problematik lösen. Allerdings sind bestehende Verfahren eher ungeeignet, da sie die notwendigen Anforderungen an Genauigkeit und Robustheit nicht erfüllen. Diese Dissertation leistet einen Beitrag, um diese Lücke zu schließen und genaue, robuste Verfahren zur Bestimmung von Vitalparametern mit intelligenten Verfahren der Bildverarbeitung zu untersuchen. Hierfür werden die Vitalparameter der Herzfrequenz, der Respirationsfrequenz und der Sauerstoffsättigung analysiert. Auf Grundlage einer Quantifizierung der beiden stärksten Störquellen, der Intensitäts- und Bewegungsartefakte, wird ein neues Verfahren, die Ratio-Methode, zur Herzfrequenzbestimmung vorgestellt. Zur Beseitigung dieser beiden Störquellen wird ein weiteres Verfahren präsentiert, welches auf einem individuellen, situationsabhängigen Hautfarbenmodell, einem erweiterten KLT-Tracking, einer ICA, einer Kanalextraktion sowie einer adaptiven Filterung beruht. Aus der Evaluierung geht hervor, dass der gemittelte RMSE für alle Szenarien stets unter einer Schwelle von 3 BPM liegt. Die Ermittlung der Respirationsfrequenz im sichtbaren Wellenlängenbereich wird mit Methoden des optischen Flusses, einer PCA sowie einer FFT analysiert. Es wird gezeigt, dass sich der gemittelte RMSE für alle Testfälle unterhalb von 2,5 Atemzyklen pro Minute befindet. Als weiterer Vitalparameter wird die Sauerstoffsättigung untersucht. Dabei werden sowohl Systeme bestehend aus einer Kamera als auch aus zwei Kameras betrachtet. Neben dem sichtbaren Wellenlängenbereich wird die Vitalparameterbestimmung auch im mittleren Infrarotbereich analysiert. Die Untersuchung fokussiert sich dabei auf die Vitalparameter der Respirations- und der Herzfrequenz. Um die Einsatzfähigkeit von Vitalparametern im Bereich AAL realistisch zu evaluieren, werden Messungen in der AAL-Testwohnung der TU Chemnitz durchgeführt und analysiert. Dabei hat sich herausgestellt, dass sowohl die Herz- als auch die Respirationsfrequenz zuverlässig erfasst werden können. Zudem wird gezeigt, dass neben perspektivischen Kameras auch omnirektionale Kameras zur Bestimmung von Vitalparametern geeignet sind. In einer weiteren Untersuchung erfolgt die Analyse für die Vitalparameterbestimmung bei Dunkelheit im nahen Infrarotbereich. Neben dem Bereich AAL eröffnen sich weitere Anwendungsszenarien auf den folgenden Feldern: Fahrerüberwachung im Fahrzeug, Detektion des plötzlichen Kindstods, E-Rehabilitation und Triage im Krankenhaus. / Europe, in particular Germany, is facing the problem of a steadily ageing society. This development goes hand in hand with a higher demand for technical assistance systems, which can assist elderly people with their self-determined living. A key element is the interconnection of different systems in the home environment in order to recognise the behaviour of a person. In case of an emergency, relatives and medical personnel can be notified automatically. In order to ensure that functionality, the monitoring of vital parameters is a crucial element to determine and to analyse the health status of a person. Conventional systems that determine vital parameters require body contact and are therefore uncomfortable for the person who wars such as system. In contrast to that, contact-less optical methods based on image processing do not share these problems. However, existing methods in this field do not fulfil the requirements regarding accuracy and robustness. This thesis contributes to close this gap by investigating accurate and robust methods to determine vital parameters by means of highly sophisticated image processing techniques. To that aim, vital parameters such as the heart rate, the respiration rate and the oxygen saturation are considered. On the basis of the quantisation of intensity and motion artefacts, a new method, the so-called ratio method, is introduced to determine the heart. In order to eliminate these artefacts another method based on an individual, scene depending skin colour model, an extended KLT tracking, an ICA, a channel selection and an adaptive filtering is presented. The evaluation shows that the mean RMSE is always below a threshold of 3 BPM. The determination of the respiration rate in the visual spectrum is realised by using optical flow, a PCA and an FFT. It can be shown that the mean RMSE is below a threshold of 2.5 breath cycles per minute for all test cases. For the oxygen saturation two setups consisting of one and two cameras respectively are investigated. Determining vital parameters is not limited to the visual spectrum. It is as well feasible to analyse long-wavelength thermal images in order to determine the heart rate and the respiration rate, which was also studied in this thesis. For a realistic evaluation the developed algorithms in the field of AAL, measurements in the AAL living lab of TU Chemnitz are conducted and analysed. It can be demonstrated that the heart rate and the respiration rate can be reliably detected. Moreover, it can be shown that omni-directional cameras can be used in the home environment for vital parameter extraction. Furthermore, this thesis provides the evidence that the determination of the heart rate and the respiration rate is also possible in case of short-wavelength infrared light. Besides the field of AAL, there exist more application areas such as driver monitoring, detection of sudden infant death syndrome, e-rehabilitation or triage in hospital.

The ability of human beings to perceive the environment around them with their visual system is called motion perception. This means that the attention of our visual system is primarily focused on those objects that are moving. The property of human motion perception is used in this dissertation to infer human activity from data using artificial neural networks. One of the main aims of this thesis is to discover which modalities, namely RGB images, optical flow and human keypoints, are best suited for HAR in omnidirectional data. Since these modalities are not yet available for omnidirectional cameras, they are synthetically generated and captured with an omnidirectional camera. During data generation, a distinction is made between synthetically generated omnidirectional data and a real omnidirectional dataset that was recorded in a Living Lab at Chemnitz University of Technology and subsequently annotated by hand. The synthetically generated dataset, called OmniFlow, consists of RGB images, optical flow in forward and backward directions, segmentation masks, bounding boxes for the class people, as well as human keypoints. The real-world dataset, OmniLab, contains RGB images from two top-view scenes as well as manually annotated human keypoints and estimated forward optical flow. In this thesis, the generation of the synthetic and real-world datasets is explained. The OmniFlow dataset is generated using the 3D rendering engine Blender, in which a fully configurable 3D indoor environment is created with artificially textured rooms, human activities, objects and different lighting scenarios. A randomly placed virtual camera following the omnidirectional camera model renders the RGB images, all other modalities and 15 predefined activities. The result of modelling the 3D indoor environment is the OmniFlow dataset. Due to the lack of omnidirectional optical flow data, the OmniFlow dataset is validated using Test-Time Augmentation (TTA). Compared to the baseline, which contains Recurrent All-Pairs Field Transforms (RAFT) trained on the FlyingChairs and FlyingThings3D datasets, it was found that only about 1000 images need to be used for fine-tuning to obtain a very low End-point Error (EE). Furthermore, it was shown that the influence of TTA on the test dataset of OmniFlow affects EE by about a factor of three. As a basis for generating artificial keypoints on OmniFlow with action labels, the Carnegie Mellon University motion capture database is used with a large number of sports and household activities as skeletal data defined in the BVH format. From the BVH-skeletal data, the skeletal points of the people performing the activities can be directly derived or extrapolated by projecting these points from the 3D world into an omnidirectional 2D image. The real-world dataset, OmniLab, was recorded in two rooms of the Living Lab with five different people mimicking the 15 actions of OmniFlow. Human keypoint annotations were added manually in two iterations to reduce the error rate of incorrect annotations. The activity-level evaluation was investigated using a TSN and a PoseC3D network. The TSN consists of two CNNs, a spatial component trained on RGB images and a temporal component trained on the dense optical flow fields of OmniFlow. The PoseC3D network, an approach to skeleton-based activity recognition, uses a heatmap stack of keypoints in combination with 3D convolution, making the network more effective at learning spatio-temporal features than methods based on 2D convolution. In the first step, the networks were trained and validated on the synthetically generated dataset OmniFlow. In the second step, the training was performed on OmniFlow and the validation on the real-world dataset OmniLab. For both networks, TSN and PoseC3D, three hyperparameters were varied and the top-1, top-5 and mean accuracy given. First, the learning rate of the stochastic gradient descent (Stochastic Gradient Descent (SGD)) was varied. Secondly, the clip length, which indicates the number of consecutive frames for learning the network, was varied, and thirdly, the spatial resolution of the input data was varied. For the spatial resolution variation, five different image sizes were generated from the original dataset by cropping from the original dataset of OmniFlow and OmniLab. It was found that keypoint-based HAR with PoseC3D performed best compared to human activity classification based on optical flow and RGB images. This means that the top-1 accuracy was 0.3636, the top-5 accuracy was 0.7273 and the mean accuracy was 0.3750, showing that the most appropriate output resolution is 128px × 128px and the clip length is at least 24 consecutive frames. The best results could be achieved with a learning rate of PoseC3D of 10-3. In addition, confusion matrices indicating the class-wise accuracy of the 15 activity classes have been given for the modalities RGB images, optical flow and human keypoints. The confusion matrix for the modality RGB images shows the best classification result of the TSN for the action walk with an accuracy of 1.00, but almost all other actions are also classified as walking in real-world data. The classification of human actions based on optical flow works best on the action sit in chair and stand up with an accuracy of 1.00 and walk with 0.50. Furthermore, it is noticeable that almost all actions are classified as sit in chair and stand up, which indicates that the intra-class variance is low, so that the TSN is not able to distinguish between the selected action classes. Validated on real-world data for the modality keypoint the actions rugpull (1.00) and cleaning windows (0.75) performs best. Therefore, the PoseC3D network on a time-series of human keypoints is less sensitive to variations in the image angle between the synthetic and real-world data than for the modalities RGB images and optical flow. The pipeline for the generation of synthetic data with regard to a more uniform distribution of the motion magnitudes needs to be investigated in future work. Random placement of the person and other objects is not sufficient for a complete coverage of all movement magnitudes. An additional improvement of the synthetic data could be the rotation of the person around their own axis, so that the person moves in a different direction while performing the activity and thus the movement magnitudes contain more variance. Furthermore, the domain transition between synthetic and real-world data should be considered further in terms of viewpoint invariance and augmentation methods. It may be necessary to generate a new synthetic dataset with only top-view data and re-train the TSN and PoseC3D. As an augmentation method, for example, the Fourier Domain Adaption (FDA) could reduce the domain gap between the synthetically generated and the real-world dataset.:1 Introduction 2 Theoretical Background 3 Related Work 4 Omnidirectional Synthetic Human Optical Flow 5 Human Keypoints for Pose in Omnidirectional Images 6 Human Activity Recognition in Indoor Scenarios 7 Conclusion and Future Work A Chapter 4: Flow Dataset Statistics B Chapter 5: 3D Rotation Matrices C Chapter 6: Network Training Parameters