Human-machine-interfaces in automotive environments using capacitive proximity sensors

The need for novel interaction paradigms in automotive human-machine interface (HMI) applications has become apparent within the last decade. The number of functions to be controlled in modern cars rises constantly. In parallel, increasing traffic density demands more and more attention and concentration from the driver. Natural interaction paradigms, like gesture-based interaction, provide prospects for more intuitive and less distracting function control. Prominent research projects use vision-based approaches like cameras to replace buttons successively. Capacitive proximity sensing constitutes a promising alternative technology to realize contactless gesture-based interaction. The required sensor electrode surfaces are installed at locations at which interaction should be possible. This thesis describes an approach for free-space contactless gesture recognition with transparent electrode surfaces, which are attached to a common monitor display. An adaption on the Condensation algorithm is proposed for object tracking. Recognition and discrimination of gestures is realized with an approach employing hidden Markov models. The developed method with its required hardware and electrode layouts is prototypically realized as a demonstrator. Based upon the finalized prototype, a user survey is undertaken to evaluate the user experience and relevance of contactless gesture-based interaction paradigms for in-vehicle applications.