Catálogo de publicaciones - libros

In various application areas for alerting systems, the context and knowledge of several parties affect profile definition and filtering. For example, in healthcare nurses, doctors and patient influence the treatment process. Thus, profiles for alerting systems have to be generated by the explicit collaboration of several parties who may not know each other directly.

We propose the new concept of to capture these different conditions and contexts. These profiles exploit each single party’s expert-knowledge for defining the context under which (health-related) alerting is required. Challenges include the definition and refinement of profiles as well as conflict detection in context definitions.

Providing mobile users with relevant and up-to-date information on the move through wireless communication needs to take the current context of a user into account. In this paper, the context of a user with respect to his movement behaviour as well as device characteristics is under investigation. In outdoor areas, particularly in an urban area, obviously there is often sufficient communication bandwidth available. In some areas though, especially in rural areas, communication bandwidth coverage is often poor. Providing users in such areas with relevant information and making this information available in time is a major challenge. Prefetching tries to overcome these problems by using predefined user context settings. In situations where resource restrictions like limited bandwidth or insufficient memory apply, strategies come into place to optimize the process. Such strategies will be discussed. Evaluating the different types of users supports the approach of getting the relevant information to the user at the right time and the right place.

Peer-to-peer (P2P) networks are receiving increasing attention in a variety of current applications. In this paper, we concentrate on applications where a mobile user queries peers to find either data (e.g., a list of restaurants) or services (e.g., a reservation service). We classify location-based queries in categories depending on parameters such as the user’s velocity, the nature of the desired information, and the anticipated proximity of this information. We then propose query routing strategies to ensure the distributed query evaluation on different peers in the application while optimizing the device and network energy consumption.

In this paper we present a novel approach to design and implement applications that provide location-aware services. We show how a clear separation of design concerns (e.g. applicative, context-specific, etc) helps to improve modularity. We stress that by using dependency mechanism among outstanding components we can get rid of explicit rule-based approach thus simplifying evolution and maintenance. We first discuss some related work in this field. Next, we introduce a simple exemplary scenario and present the big picture of our architectural approach. Then we detail the process of service definition and activation. A discussion on communication and composition mechanisms is next presented and we end presenting some concluding remarks and further work.

Next-generation handhelds are expected to be multi-functional devices capable of executing a broad range of compute-intensive applications. These handheld devices are constraint by their physical size, their computational power and their networking ability. These factors could hinder future performance and versatility of handheld devices. Reconfigurable hardware incorporated within distributed servers can help address portable device constraints. This paper proposes a context-based negotiation and bidding technique enabling a handheld device to intelligently utilise its surrounding reconfigurable resources. The negotiation protocol enables handhelds to optimally offload their computational task. The contextual aspect of the protocol uses the location of a mobile device to identify the urgency of a user request. This helps to optimise the quality of service experienced by a handheld user. An architectural framework currently being deployed and tested as well as overall future objectives are outlined.

The steady rise of mobile computing devices and wireless local-area networks (WLAN) has fostered a growing interest in location-aware systems and services (LBS). Context-awareness in mobile systems give users more convenience with services specific to their preferences, behaviour and physical position. This paper presents an architecture for LBSs in WLANs describing the essential location determination component. An example service using Web Service techniques is described to illustrate efficient service invocation and interaction.

Context-aware mobile systems aim at delivering information and services tailored to the current user’s situation [1], [10]. One major application area of these systems is the tourism domain, assisting tourists especially during their vacation through location-based services (LBS) [4], [7]. Consequently a proliferation of approaches [2], [5], [8], [9], [12], [15], [17], [18] can be observed, whereby an in-depth study of related work has shown that some of these existing mobile tourism information systems exhibit few limitations [3], [19]: First, existing approaches often use proprietary interfaces to other systems (e.g. a Geographic Information System – GIS), and employ their own data repositories, thus falling short in portability and having to deal with time consuming content maintenance. Second, often thick clients are used that may lack out-of-the-box-usage. Third, existing solutions are sometimes inflexible concerning configuration capabilities of the system. To deal with those deficiencies, we present a lightweight framework for LBS that can be used for various application domains. This framework builds on existing GIS standards, incorporates already available Web content, can be employed out-of-the-box, and is configurable by using a Web-based interface. The applicability of the framework is demonstrated by means of a prototype of a mobile tourist guide.

The development of mobile devices and wireless networks made it possible for users to seamlessly use different devices to recognize changes in their computing environment. In this paper, we propose a music information retrieval system (MIRS) that exploits various types of contextual information. Our system is based on JXTA technology which enables to join at any time and direct communication between peers. It supports the personalized retrieval of music information at specific moments and locations. Each peer can run a context interpreter and each edge peer functions as a query requester or a query responser. The query requester’s context interpreter analyzes user’s context and customizes the search result. The query responser’s context interpreter analyzes the pattern for the query melody and selects requested music information.

We propose a formal definition of a pervasive service model targeting the very dynamic environments typical of mobile application scenarios. The model is based on a requirement analysis and evolves from existing service definitions. These are extended with pervasive features that allow for the modeling of context awareness, and pervasive functionality needed for the dynamic (re-) composition of services. In order to demonstrate its value, we present an implementation of a pervasive service platform that makes use of the service model.

This paper targets with applications running on mobile devices and using context informations. Following previous studies from other authors, we extend the notion of context area replacing distance function by cost function. Using this extension, we exhibit three different modes of selection and demonstrate their differences on a mobile applications: the museum visit.