Development of a novel context prediction algorithm and analysis of context prediction schemes

Context-awareness is a broad field that can be further partitioned into sub-disciplines which themselves constitute interesting research topics. A context-aware application might be aware of its present, past or future context. Most work considers present or past context information. In this thesis we investigate issues related to the inference of future context. We discuss the context prediction task and identify chances and challenges that derive from it. This discussion leads to the development of a de? nition of the context prediction task. Dependent on the amount of pre-processing or context abstraction applied to context data, di? erent context prediction schemes are distinguished between. We study context prediction schemes for their prediction accuracy and provide guidelines to application designers as to which prediction scheme to utilise in which situation. Analytical results obtained in these studies are con? rmed in simulations in several application domains. We also develop a novel architecture for context prediction approaches that is applicable to arbitrary context abstraction levels. This architecture allows the distribution of components to various interconnected devices in a mobile ubiquitous computing environment. For context prediction algorithms, we review the state of the art and discuss strengths and weaknesses of these algorithms. Furthermore, a novel context prediction algorithm is developed in the scope of the thesis. On the basis of requirements identi? ed for algorithms in UbiComp environments, we are able to name those approaches that are best suited in these environments. For three exemplary algorithms, simulation results are provided in various application domains.