Spatial Awareness of Autonomous Embedded Systems

The invisible integration of technology into everyday objects like home appliances, cars and mobile phones, which is the declared vision and fundament of research in the field of pervasive computing, leads to huge quantities of smart objects which are situated in the surrounding physical space. Equipped with embedded systems technology, they become increasingly heterogeneous and interconnected, raising the challenge of a semantically meaningful interplay with each other. In this thesis, such physical objects with embedded computing and communication technology are referred to as digital artifacts. One approach to meet this challenge is to design and implement systems that are able to operate autonomously in the background, namely with as little human intervention as possible, and interact with humans in a more unobtrusive way. In order to achieve autonomy, two aspects are particularly important: (i) context-awareness, which refers to the ability of a digital artifact to acquire environmental information in order to become aware of its situation and adapt to changing situations at runtime, and (ii) context sharing, which relates to an artifact's ability of exchanging context information with other artifacts in communication range. An essential part of the context of spatially distributed objects is their position, direction and spatial extension with respect to an external reference system or with respect to other objects. The focus of this thesis is to make digital artifacts aware of such context information in order to enable their autonomous adaptation to spatial changes in the environment, whereas our main interest is on qualitatively abstracted spatial relations and their use at the application level. Concepts for recognizing, representing and reasoning about qualitative spatial relations among autonomous artifacts and their changes over time are presented, as well as an according architecture which has been implemented in a flexible software framework that builds upon qualitative relationship abstractions and the rule-based inference of conclusions from them. Special attention is paid to the adaptivity to different application domains, which influence the semantics of spatial abstractions and thus the behavior of digital artifacts. Evaluation results show the feasibility of the proposed concepts for developing spatially aware applications which involve one or more spontaneously interacting digital artifacts, and in particular that qualitatively abstracted relations can constitute an adequate basis for it.