Using Inaudible Audio to Improve Indoor-Localization- and Proximity-Aware Intelligent Applications (2002.00091v1)

Abstract: While it is often critical for indoor-location- and proximity-aware applications to know whether a user is in a space or not (e.g., a specific room or office), a key challenge is that the difference between standing on one side or another of a doorway or wall is well within the error range of most RF-based approaches. In this work, we address this challenge by augmenting RF-based localization and proximity detection with active ultrasonic sensing, taking advantage of the limited propagation of sound waves. This simple and cost-effective approach can allow, for example, a Bluetooth smart-lock to discern whether a user is inside or outside their home in order to lock or unlock doors automatically. We describe a configurable architecture for our solution and present experiments that validate this approach but also demonstrate that different user behavior and application needs can impact system configuration decisions. Finally, we describe applications that could benefit from our solution and address privacy concerns.