Blackboard-based software framework and tool for mobile device context awareness: Dissertation

The usage needs of a mobile device vary according to context. Mobile context awareness research aims at providing the device user with a way of usage that suits the situation. Interaction based on context requires acquiring, abstracting and delivering information from multiple sources, such as sensors, to the application or application control. A generic software framework and tool for facilitating the rapid development of mobile device context-aware applications were developed in this work. The blackboard-based framework supports all tasks that are required for context-based application control, where contexts can be any events that are relevant to user interaction with the application, including explicit inputs. The core component of the framework, Context Manager, provides a publish and subscribe mechanism and a database for the applications and application control. The framework provides an application programming interface (API) for developers. As a higher abstraction-level programming interface, a customization tool enables easy end-user development of context-aware features into existing applications without changing them. An extensible ontology is used as a uniform context representation within the framework. The purpose of the ontology, together with the API, is to enable easy access, use and reuse of human-understandable context information. Context information sources, such as sensors, often produce a continuous stream of low abstraction-level data. The framework supports the transformation of a continuous data stream into abstracted context events, described in the ontology. Context information is delivered to applications or application control as abstracted events. The main result of the dissertation is a software framework, ontology and tool, which facilitate the customization of sensor-based human-computer interaction in mobile devices. The practical applicability, scope, and computational efficiency of the implemented framework and customization tool are evaluated with performance measurements and multiple applications implemented in a mobile phone with real sensor sources.