Energy Efficient Collaborative Sensing-Based Design: Soft Keyboard Case Study

With our approach, for the synthesis and energy-efficient operation of a pressure sensor-based soft keyboard, energy is optimized at five levels of abstraction: 1) architecture of sensor systems, 2) sensing schedule, 3) sensor data processing, 4) use of semantic information, and 5) customization. We have been able to achieve up to a factor of 63 energy reduction in a keyboard case study over the standard using various techniques. For example, when each sensor senses several keyboard events and each event is sensed by three sensors, sensing and data processing results in a 12- to 42-fold energy savings depending on typing speeds. Customization techniques accounting for user typing speed and duration of typing pulses support another 1.5-fold savings. We demonstrate using a proof of concept prototype evaluation with five participants.