Since an ever-growing demand for 'smart' computer systems capable of monitoring or adapting to their environments arose throughout the preceding decade, sensing technologies have gradually found their way from embedded systems and cybernetic research into various fields of computer science (e.g., ubiquitous computing, medical computer science). This thesis is concerned with the design and implementation of a mobile context-sensing framework based on a wireless sensor network. The particular contributions of this work are a generic software architecture for context-aware applications and an approach to overcome the calibration problem in wireless sensor networks. The calibration approach proposed involves automatic as well as user-controlled calibration of sensors. To demonstrate the feasibility of the context-framework, it has been prototypically implemented for the use in mobile scenarios. For this purpose, wireless network nodes equipped with sensors capable of measuring environmental conditions, such as light, temperature, humidity, and sound, and sensing motion have been used to provide information on the context of a moving person. In this thesis, the design and implementation of the context-sensing framework is detailed and the experiments carried out to describe the calibration challenges are summarised. Finally, the thesis outlines the design and results of a mobile HCI study where the prototype of the framework has been successfully used. In this mobile HCI study, the influence of motion and environmental conditions on the users of mobile applications has been investigated by means of the context information captured by the context-sensing framework.

Top