1,721,063 research outputs found
Context-aware pervasive interfaces
No full textThe proliferation of pervasive services requires advanced methods to adapt the service provision to the user's context. The author presents a hybrid statistical and semantic framework for interface selection and adaptation. The approach is to find the best compromise between urgency and privacy requirements, avoiding interference with the user's activities
Opportunistic pervasive computing: adaptive context recognition and interfaces
No full textPervasive computing is promising to radically change people’s life in several dimensions, including the way we work, travel, have leisure, and take care of ourselves. In order to realize the goal, pervasive technologies must be aware of people’s context and react accordingly by tailoring services and interfaces to the current situation. Unfortunately, today’s pervasive applications have a restricted scope: their visibility is limited to their specific application domain. Hence, they are very well suited to address their specific objective, but lack the overall landscape of the user’s context and goals. In this paper, we put forward the vision of opportunistic pervasive computing: pervasive technologies that can dynamically exploit the available data sources and heterogeneous reasoning services to opportunistically reconstruct the whole landscape of the users’ context and seamlessly adapt to their needs and expectations. We point out the research challenges involved in this vision, and we present a conceptual model and architecture to address these issues, that include solutions for interoperability, reasoning integration, and artificial intelligence tools. Moreover, we introduce technical solutions for opportunistic context data discovery, reasoning and integration, as well as a hybrid method for opportunistic context-aware adaptation. We have also developed a prototype implementation considering an adaptive healthcare application for self-administration of therapies, and a system for opportunistic recognition of different activities based on sensor data. Experimental results indicate the viability and effectiveness of our solution
Special Issue on Context Awareness
Full text linkContext-awareness is a fundamental ingredient of pervasive computing.[...
Towards the combination of statistical and symbolic techniques for activity recognition
Full text linkContinuous media adaptation for mobile computing using coarse-grained asynchronous notifications
Full text linkThe recent spreading of public wireless infrastructures allowing for higher data rates makes mobile communications networks a very attractive platform for distribution of multimedia content. At the same time, limited resources in public wireless networks pose serious questions on how to bring services and multimedia to terminals to be used anywhere. Content adaptation is required in order to bring the best perceptual experience to the end-user while optimizing resources usage. Unfortunately, content adaptation is very difficult to achieve and is usually related to band-width availability only. In this paper we propose to extend existing service provisioning architectures with an asynchronous notification system to keep up-to-date the whole set of user profile data during service provisioning. We argue that the average multimedia application behavior, still adhering to a model based on a very limited number of choices, is not affected by increased reaction time and coarse-grained parameters responsivity. Furthermore, introduction of asynchronous notifications will enable service providers to adapt content considering any parameter characterizing the user profile, not just available bandwidth. © 2005 IEEE
Unsupervised Recognition of Multi-Resident Activities in Smart-Homes
Full text linkSeveral methods have been proposed in the last two decades to recognize human activities based on sensor data acquired in smart-homes. While most existing methods assume the presence of a single inhabitant, a few techniques tackle the challenging issue of multi-resident activity recognition. To the best of our knowledge, all existing methods for multi-inhabitant activity recognition require the acquisition of a labeled training set of activities and sensor events. Unfortunately, activity labeling is costly and may disrupt the users' privacy. In this article, we introduce a novel technique to recognize multi-inhabitant activities without the need of labeled datasets. Our technique relies on an unlabeled sensor data stream acquired from a single resident, and on ontological reasoning to extract probabilistic associations among sensor events and activities. Extensive experiments with a large dataset of multi-inhabitant activities show that our technique achieves an average accuracy very close to the one of state-of-the-art supervised methods, without requiring the acquisition of labeled data
Reasoning with smart objects’ affordance for personalized behavior monitoring in pervasive information systems
No full textThe miniaturization of sensors and their integration in everyday appliances have opened the way for ecologically monitoring people's behavior based on their interaction with smart objects. Thanks to behavior monitoring, mobile, and ubiquitous information systems in the areas of e-health, home automation, and smart cities are becoming more and more "smart," being able to dynamically adapt themselves to the current users' context and situation. However, human behavior is characterized by large variability due to individual habits, physical disabilities or cognitive impairment. This aspect makes behavior monitoring a challenging task. On the one side, execution variability makes it hard to acquire sufficiently large activity datasets needed by supervised learning methods. On the other side, being based on a strict definition of activities in terms of constituting simpler actions, existing knowledge-based frameworks fall short in adapting to the specific characteristics of the subject. Hence, the variability of activity execution by different subjects calls for personalized methods to capture human activities and interaction in smart spaces at a fine-grained level. In this paper, we address this challenge by proposing a novel hybrid reasoning framework to capture fine-grained interaction with smart objects considering the specific features of individuals. Our model has its roots in the well-founded psychological theory of affordances, i.e., those features of an object that naturally explain its possible uses and how it should be used. The core of the framework is the ontological model of smart objects affordance, expressed through the OWL 2 Web Ontology Language. Through a use case in pervasive healthcare, we show how our framework can be applied to personalized recognition of abnormal behaviors. In particular, we tackle a particularly challenging issue: how to recognize early behavioral symptoms of mild cognitive impairment in subjects with physical disabilities. Moreover, an extensive experimental evaluation with real-world datasets acquired from 24 subjects shows the effectiveness of our framework in recognizing human activities and fine-grained manipulative gestures in different pervasive computing environments
Sensor-based activity recognition: One picture is worth a thousand words
Full text linkIn several domains, including healthcare and home automation, it is important to unobtrusively monitor the activities of daily living (ADLs) carried out by people at home. A popular approach consists in the use of sensors attached to everyday objects to capture user interaction, and ADL models to recognize the current activity based on the temporal sequence of used objects. Often, ADL models are automatically extracted from labeled datasets of activities and sensor events, using supervised learning techniques. Unfortunately, acquiring such datasets in smart homes is expensive and violates users’ privacy. Hence, an alternative solution consists in manually defining ADL models based on common sense, exploiting logic languages such as description logics. However, manual specification of ADL ontologies is cumbersome, and rigid ontological definitions fail to capture the variability of activity execution. In this paper, we introduce a radically new approach enabled by the recent proliferation of tagged visual contents available on the Web. Indeed, thanks to the popularity of social network applications, people increasingly share pictures and videos taken during the execution of every kind of activity. Often, shared contents are tagged with metadata, manually specified by their owners, that concisely describe the depicted activity. Those metadata represent an implicit activity label of the picture or video. Moreover, today’s computer vision tools support accurate extraction of tags describing the situation and the objects that appear in the visual content. By reasoning with those tags and their corresponding activity labels, we can reconstruct accurate models of a comprehensive set of human activities executed in the most disparate situations. This approach overcomes the main shortcomings of existing techniques. Compared to supervised learning methods, it does not require the acquisition of training sets of sensor events and activities. Compared to knowledge-based methods, it does not involve any manual modeling effort, and it captures a comprehensive array of execution modalities. Through extensive experiments with large datasets of real-world ADLs, we show that this approach is practical and effective
Demonstration of a sensor-based app for self-monitoring of medicine intake
No full textAccurate adherence to prescribed medications is essential for the effectiveness of therapies, but several studies show that when patients are responsible for treatment administration, poor adherence is prevalent. Existing apps to support self-administration of drugs may interfere with the normal routine of patients by providing unnecessary reminders. More sophisticated solutions, including the use of smart packaging and ingestible sensors, are currently restricted to patients involved in a few clinical studies. In this paper, we demonstrate a novel app to support self-administration of drugs without interfering with the patient's routines. The system relies on cheap wireless sensors attached to medicine boxes to detect medicine intake. The app uses machine learning to detect intake events, and active learning to improve recognition based on the user's feedback. In the demonstration, we show a working prototype of the system, which includes a Web dashboard for physicians to monitor the rate of intakes. Copyright is held by the author/owner(s)
- …
