Actively exploiting propagation delay for acoustic systems

Abstract

Propagation delay refers to the length of time it takes for a signal to travel from point A to point B. Many existing systems, including Global Positioning System (GPS) localization, vehicular imaging, and microphone array beamforming, have taken advantage of propagation delay. This dissertation revisits different properties of propagation delay to enable new acoustic techniques and applications. For instance: (1) We leverage the propagation delay difference between two very different frequencies -- radio frequency (RF), and acoustics -- to improve active noise cancellation. By "piggybacking" sound over RF, our proposed system is able to compute anti-noise signals more precisely, and ultimately attain better cancellation performance. (2) We develop solutions that exploit the propagation delays of multipath echoes to localize an indoor human speaker. By aligning the arrivals of the voice signal at different times, we compute user location within an optimization framework, serving as a valuable context for smart voice assistants like Amazon Echo and Google Home. (3) We design 3D directional sound by actively synthesizing different propagation delays at two ears using earphones. We develop algorithms that accurately track the 3D orientation of the head, a key enabler for designing 3D acoustics. In general, this dissertation shows that while propagation delay has been studied for a long time and for many applications, there is still opportunity for new techniques and systems, by carefully looking at different properties of the propagation delay, across frequencies, time, and space.