OERA Webinar Series: Real-time detection of marine mammals in high flow environments

October 22, 2020 1:00pm – 2:00pm ADT (4:00pm – 5:00pm UTC)

Dr. Jean-François Bousquet, Dalhousie University

This talk will focus on the development of a remote sensing node to detect the presence of marine mammals in tidal environments. The work has two main contributions: 1) the development of tracking algorithms to localize different species in real-time, and 2) the development of a wireless communication system to forward the information from the remote instrument to a gateway at or above the sea surface. A sea trial in the summer 2019 with a 5-element horizontal acoustic sensor array has allowed us to gather different animal signatures as well as confirm the challenges of communicating wirelessly in high flow environment. As will be demonstrated, in this work, animal localization relies on the use of different signal processing algorithms, and their performance will be assessed for signatures of animals that occupy extreme frequencies in the acoustic spectrum: our case study focuses on the humpback whale and the harbour porpoise, two species that are prominent in the Bay of Fundy. For the high frequency echolocations produced by harbour porpoises, a MUSIC algorithm is applied to obtain the direction of arrival. For low-frequency vocalization, processing in the frequency domain is implemented using the Multi-Channel Cross-Correlation Coefficient Algorithm as well as the Minimum Variance Distortionless Response (MVDR) beamformer. The exact location of the sources is obtained using matched field processing (MFP), and for this purpose a model of the shallow water environment is included. The objective is to implement these algorithms on a real-time processing engine, and to forward events above the surface using a short-range magneto-inductive link.