The welfare of the marine environment is of increasing global concern. Underwater acoustics is a promising approach to assess and mitigate human impact on marine mammals, as well as to determine population sizes, monitor behavior and migration patterns, identify critical habitat, and study bioacoustics. The research funded by this award is expected to lead to radically improved methods for passive acoustic monitoring of marine animals by using modern signal processing methods. Passive acoustic monitoring methods are non-invasive, unobtrusive, and can be used in poor weather conditions and lack of daylight. They enable continuous and remote sensing in a cost and time efficient manner. For these reasons, passive acoustic techniques are valuable on their own, and as a complement to existing visual and tagging techniques for studying marine mammals.
Many marine mammals are vocal so by processing data collected by hydrophones (underwater microphones), they can be detected, monitored and studied. Due to the complex nature of the marine environment and the vocalizations, advanced signal processing methods are needed. The project will consider the following specific problems 1) Channel estimation and use of this for localization of mammals. 2) Separation of multiple mammals. Signal processing methods for blind source separation can be used to separate the vocalizations of multiple mammals within the range of the hydrophones. 3) Detection of vocalizations. Most of the time sound recordings are pure noise. The project will develop methods for separating noise and signal segments. 4) Analysis of vocalizations. The project will then look into further analyzing the sounds, for example for classification of species, establishing information content in vocalizations, and extracting bioacoustic properties.
