Exploration of Perceptual Computing for Smart-Its

The future success of ubiquitous computing depends to a big part on how well applications can adapt to their environment and act accordingly. This thesis has set itself the goal of exploring perceptual computing for Smart-Its, which is one such ubiquitous computing vision. The thesis addresses this task by first presenting a literature survey of the terms context and context-awareness. Using one of the definitions presented a set of well known context-aware applications from the ubiquitous computing research community are classified. A comparison between these classified applications and new applications envisaged especially for the Smart-Its platform results in the finding that the Smart-Its applications are more reactive in general and use a wider range of context information. This application comparison was conducted to explore the differences between the existing applications and the envisaged Smart-Its applications, and also it was used to derive new perception tasks. Determining how close two Smart-Its are located proved to be one of the most prominent tasks. By describing a wide range of sensors the foundation for exploring perceptual computing is laid. In a first step the perception properties of each single sensor are analyzed. Thereafter different integration techniques for multiple sensors are explored. Concluding, some ideas to solve the proposed perception tasks are presented. The most prominent task of determining the proximity between two Smart-Its is explored in greater detail. A proximity hierarchy is proposed which classifies different physical parameters to determine a semantic proximity measure. On one hand the categorization distinguishes between the dynamics in the environment opposed to the static state, and on the other hand between the dynamics of the object, which the Smart-It is attached to, opposed to the objects static state. In two experiments it is shown how semantic proximity information can be obtained using a variety of sensors.