Integrating Hybrid-Clustering and Localized Regression for Time Synchronization of a Hierarchical Underwater Acoustic Sensor Array

Abstract

Time synchronization is a critical requirement for the application of underwater acoustic sensor network (UWSN). Although a number of time synchronization protocols have been proposed for UWSN, none of them can be directly applied to stand-alone autonomous acoustic receivers, as they lack hardware platforms permitting communication. In this paper, we propose a machine learning-based time synchronization framework for stand-alone autonomous receiver arrays, using the Juvenile Salmon Acoustic Telemetry System as a case study. The proposed framework consists of array partition and time synchronization. Using detections of receiver-attached beacons as input, this framework synchronizes all receiver clocks to a root receiver clock. The framework has been successfully used in a field study at Trevallyn Dam forebay in Tasmania, Australia.