www.uhasselt.be
DSpace

Document Server@UHasselt >
Research >
Research publications >

Please use this identifier to cite or link to this item: http://hdl.handle.net/1942/13768

Title: The Variability of Location Context in Pervasive Environments: Modelling, Representation and Visualisation
Authors: Aksenov, Petr
Advisors: Coninx, Karin
Luyten, Kris
Issue Date: 2012
Abstract: Context and context-awareness are crucial concepts in pervasive computing. Among the many and diverse types of contextual information, information about location is generally seen as one of the most important. There is a great deal of situations the users happen to be in throughout a pervasive environment. This fact imposes restrictions on the use of the available resources and the surroundings as well as requires taking into account the users‟ personal preferences and requested customisations. As a result, location context undergoes permanent and often considerable changes over time and space: e.g., loses existing location data, faces changes in quality of these data, or receives unknown data. In other words, location context in a pervasive environment is exposed to high variability, which puts limitations on the applicability of a pervasive application that is expected to use this context. This dissertation explores how some aspects of the variability of location context in a pervasive environment could be approached in a way suitable to such environment and utilised therein. We introduce a user-centric model of spatial arrangements around as well as between resources in the environment with a special focus on the model‟s ease of reference and manipulation at all stages of an application‟s development and use, and with support of uncertain and incomplete knowledge. We show how the proposed model can be integrated into a framework for dealing with pervasive environments, thereby making the framework capable of managing location-based tasks. Further on, we present an approach to coping with a multitude of location determination technologies by introducing a unified view on the existing diversity of location sensing data and formats in terms of their representation and processing within the application. We provide the approach with a number of software tools and instruments that aim at easing the inclusion of an arbitrary location provider. We then use the proposed unified approach as the underlying echanism to facilitate our investigation of user needs in awareness of location context variability. Based on two real-life user studies throughout a medium-to-large-scale environment, we identify users‟ preferences regarding knowing about their location tracking conditions and come up with a number of design guidelines and implications on visualising location context variability, which can be taken into account when developing location-aware applications. Finally, we demonstrate the benefits of combining location context with other types of context. By combining information about run-time location of members of a vehicular ad-hoc network, i.e. vehicles, with personal preferences of their drivers, we achieve a more efficient data dissemination scheme within that netAbstract iv work and manage to deliver more relevant information as compared to the state-of-the-art algorithms.
URI: http://hdl.handle.net/1942/13768
Category: T1
Type: Theses and Dissertations
Appears in Collections: PhD theses
Research publications

Files in This Item:

Description SizeFormat
PhD Petr Aksenov3.96 MBAdobe PDF

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.