Sampling Rate Offset Estimation in Acoustic Sensor Networks

Abstract

Acoustic sensor networks can achieve both a wide area coverage and coherent gain in a low-cost, scalable fashion. Their applications have been explored both in-air and underwater for robotic navigation, teleconferencing and target detection. While flexible and scalable, these networks bring new synchronization challenges to array processing. In this paper, we present methods for estimating the sampling rate offsets, or small deviations in sampling frequencies, of an entire collection of unsynchronized acoustic sensors in a single estimation procedure. We develop methods which improve upon existing techniques by leveraging the scale of the network to make the estimation more robust to noise and random source motion.