Abstract
We present a user interface system “RapTapBath” that converts an existing bathtub into various controllers by tapping locations, tones and patterns with their tap sounds on a bathtub edge. This system utilizes embedded piezoelectric sensors in a bathtub edge to analyze acoustic signals of tap sounds, and also a projector installed above the tub to project menus on the edge. Tap locations are detected by differences in signal arrival times of plural sensors. Tap tones are identified by spectrum patterns with neural-network technique. Tap patterns are detected by tapping timings with their probability calculations in a fixed time. This paper describes the system overview of RapTapBath, and each of the tapping user interface events, their specific detection methods by signal processing techniques, and also their performance evaluations. We also give effective applications for spending bathing time using this system. Finally, discussions of the limitations of current tap UI events and the point of view of interaction design with RapTapBath are described.
