ScudCORE: A Context-driven Reasoning Engine

Context-aware Computing is an important research domain in Ubiquitous Computing, which adapts computing systems to ever-changing contexts so as to provide users services and information for their tasks. It is critical to develop an effective inference mechanism. Context-aware Computing shows new characteristics such as: to be context-driven, to deal with frequently change of contexts, and to work in a real-time manner, which are different from traditional methods. The new characteristics put forward a number of challenges on reasoning engine.To deal with the rising challenges, we present a novel context-driven reasoning engine for Context-aware Computing. The reasoning engine can be used to develop reasoning modules in context-aware systems, with a great reduction of developers' burden.The main contributions of this thesis includes:1) We have analyzed the features of the Context-aware Computing inference engine distinct from traditional inference engine. On this basis, we have presented a context-driven rule-based inference engine, ScudCORE, along with its architecture and mechanisms. We also have realized a prototype of ScudCORE, taking the open source engine CLIPS as the foundation;2) We have designed and developed a wizard-style rule editor, ScudRutor, which supports context ontology. It can simplify the difficulty of rule-editing;s3) We applied the ScudCORE to the smart car space prototype developed by our research group. The rule sets for experimental scenarios of the car prototype were built by our ScudRutor. The prototype demonstrates the effectiveness of the ScudCORE and the ScudRutor.