A Context-aware and workflow-based framework for pervasive environments

Avenoğlu, Bilgin
In this thesis study, a framework offering people help and guidance for organizing their daily activities is proposed. This framework allows users to model their daily activities in the form of workflows. Due to the dynamic nature of pervasive environments, workflows are enabled to be adaptable at runtime according to context information collected in pervasive environments. A workflow engine is used for modeling and management of workflows and a separate rule engine with Complex Event Processing (CEP) capability is incorporated into the framework for both enhancing workflow adaptation and execution and for rule-based context reasoning. The adaptation model in the framework allows modeling activities in a hierarchical manner, from high level abstract activities to more detailed ones. Event-driven Architecture (EDA) is utilized for loosely coupled interaction between the workflow engine and the rule engine, allowing these engines, other context sources and user-level applications to exchange data among each other. Since workflow engines are sufficiently complex, EDA allows incorporation of context information into the workflow models without modifying the workflow language. A higher automation level than the level supported by workflows is proposed by processing events in pervasive environments using CEP. A prototype implementation is developed and the framework is evaluated with some real life examples and experimental tests demonstrating its applicability.