Many pervasive computing applications would benefit from room-level indoor localization. One example might be an E911 application. By providing emergency responders with your building/floor/room number they could arrive at your exact location sooner. Indoor localization is a popular topic and existing solutions do exist for environments when GPS is unavailable. A common approach is to utilize radio signal strength (WiFi or cellular) information, from possibly multiple sources, to estimate location using a triangulation algorithm. While this provides some useful information, it is only reliable to approximately 10 meters (in 3D space). This means the reported location may be in error by 1 or 2 rooms or floors. This error arises from the fact that interiors of most buildings exhibit complex RF radio propagation which causes nonlinearities in RF signal strength throughout the building.

An alternative approach to RF localization is to use acoustic information to determine location. The microphone on most smartphones is sampled at 48 kHz allowing signals up to the Nyquist-rate of 24 kHz to convey information. Because most humans can only hear up to 20 kHz there is approximately 4 kHz of bandwidth available. Ultrasonic signals are ideal for room-level localization because common building materials (e.g., drywall) strongly attenuate these signals therefore containing the localization signal to the room it was broadcast.

This project seeks to develop and deploy low cost ultrasonic room level localizers for pervasive computing applications (our goal is to build and deploy 100 prototypes). The localizer will be battery powered or energy harvesting and contain a small ultrasonic transducer. It will be programmed to broadcast the building name and room number where it is located. Then it would be placed at that location and periodically broadcast this information in the ultrasonic frequency range. A smartphone application (android/iOS) would be developed to sample the microphone and perform the necessary DSP to detect and decode the location.