Balance of Control between Users and Context-Aware Pervasive Systems

Bob Hardian (2008). Balance of Control between Users and Context-Aware Pervasive Systems PhD Thesis, School of Information Technol and Elec Engineering, The University of Queensland.

Attached Files (Some files may be inaccessible until you login with your UQ eSpace credentials)
Name Description MIMEType Size Downloads
n40640916_PhD_abstract.pdf 40640916_PhD_abstract.pdf application/pdf 21.10KB 3
n40640916_PhD_totalthesis.pdf Final Thesis Lodgement application/pdf 1.64MB 9
Author Bob Hardian
Thesis Title Balance of Control between Users and Context-Aware Pervasive Systems
School, Centre or Institute School of Information Technol and Elec Engineering
Institution The University of Queensland
Publication date 2008-08
Thesis type PhD Thesis
Supervisor Prof. Jadwiga Indulska
Dr. Karen Henricksen
Total pages 178
Total colour pages 9
Total black and white pages 169
Subjects 290000 Engineering and Technology
Abstract/Summary Context-awareness in pervasive computing environments can reduce user interactions with computing devices by making applications adaptive and autonomous. Context-aware applications rely on information about user context and user preferences to guide their own behaviour. However, context-aware applications do not always behave as users expect due to imperfection of context information, incorrect user preferences or incorrect adaptation rules. This may cause users to feel loss of control over their applications. To mitigate these problems, context-aware systems must provide mechanisms to strike a suitable balance between user control and software autonomy. Allowing users to scrutinise the system and allowing the system to sometimes include users in the adaptation decision making, can provide a balance of user control. This thesis addresses the shortcoming in development of context-aware pervasive systems with regard to providing balance between user control and software autonomy. The thesis shows that rather than making a context-aware application a complete black box, it is possible to allow user control of application adaptations. The system can reveal to the user what context information the system uses and how it arrives at adaptation decisions if the user requests such information. The user may decide to alter the adaptive behaviour of the system to achieve desired outcomes. Hence, a context-aware application becomes a closed loop system where the user is put into the loop if requested. The proposed approach is developed under an assumption that users differ in the level of their technology expertise and therefore the system has to provide explanations that are suitable for a particular level of user expertise. The thesis makes two important research contributions: design of the architectural framework and development of the platform exposing autonomic behaviour of context-aware applications. The architectural framework supports developers of context aware-applications in providing balance of control between users and software autonomy. The framework describes a set of models that allow revealing the adaptation behaviour of context-aware applications in a way suitable for users with various levels of expertise. The framework consists of: (i) a model for exposing elements that influence the context-aware behaviour, (ii) a generic architecture for providing balance of control, (iii) a user model, and (iv) a context graph based overview of context-aware adaptations. The platform exposing autonomic behaviour of context-aware applications is a proof of concept prototype of a software infrastructure (middleware) providing balance of control. The software infrastructure includes: (i) a Semantic Manager, developed to serve the description of elements required for explanations of the application behaviour; (ii) an extension of the PACE Middleware, to enable the middleware to expose the context information, preferences, adaptation rules and their evaluation traces, respectively. (iii) supporting tools for the application designer to prepare the overview of context-aware adaptations and review the evaluation traces. Finally, this thesis presents a case study that illustrates and evaluates the system supporting balance of control. This evaluation involves the existing application which is developed using the previous version of the PACE middleware. The case study validates the architectural framework and illustrates the process and issues involved in developing context-aware application that are able to expose elements that influence context-aware behaviour.
Keyword balance of control
context awareness
autonomic computing
pervasive systems
user interaction
Additional Notes 113, 127, 128, 137, 154, 155, 156, 157, 158

Citation counts: Google Scholar Search Google Scholar
Created: Fri, 15 May 2009, 23:09:05 EST by Mr Bob Hardian on behalf of Library - Information Access Service