To subscribe to the ORE seminar mailing list, click here.
For a (working draft) list of upcoming seminars, click here.
If you are interested in giving an ORE seminar, please contact us at nosal [at] hawaii [dot] edu.
- This event has passed.
PhD Defense. Of RATs and Men: Underwater Passive Acoustic Localization Investigations using Relative Arrival Times and Blind Channel Estimation
2 December 2022 @ 9:30 am - 10:30 am
Brendan P. Rideout
Department of Ocean and Resources Engineering
University of Hawai’i at Mānoa
**This defense is hybrid**
In person in HIG 110
Zoom meeting ID: 935 1677 2350, Passcode: ORE
Please join us afterwards (~12:30 pm) in the HIG courtyard to congratulate Brendan
Understanding the ecology of any organism requires an understanding of all its life stages. Underwater acoustics provides the ability to observe the submerged lives of marine mammals in ways not possible through visual means. The complexities of underwater acoustic propagation yield both challenges and opportunities to extract information from recorded data, of which the estimation of underwater location is one example. This dissertation presents two signal processing approaches related to passive underwater acoustic localization.
Blind channel estimation is a computational approach to estimating a set of impulse responses based on simultaneous recordings of the same unknown source by different receivers. The estimated impulse responses from this approach may simplify passive acoustic localization for some types of sound sources by making arrival times easier to identify. Differences among the received waveforms are interpreted as evidence of differences in the underlying channel impulse responses. Using a sparse assumption on these impulse responses, several different optimization approaches (OMP, CoSaMP, and NESTA) are applied to simulated ocean acoustic data. Estimated channels are a good fit to the true channel when the channels are static, but the introduction of a time-varying characteristic to the true channels negatively impacts channel recovery performance.
Single hydrophone passive acoustic localization is the practice of estimating location characteristics of an underwater sound source using acoustic recordings from a single receiver. This work develops an approach for estimating the horizontal range between a submerged source and receiver in a deep ocean environment without relying on modal dispersion. We develop a cost function and optimization approach that are robust to some significant sources of noise and environmental uncertainty. Results from simulations and ground-truthed measured data demonstrate the accuracy of this localization approach.
Underwater acoustic data recorded by the ALOHA Cabled Observatory (ACO) are processed using this single hydrophone localization approach. Acoustic recordings at ACO made between February 2007 – September 2017 yield 41,481,171 fin whale 20 Hz call detections, of which 3,445,568 detections remain after pruning away suspected sei whale and duplicate call detections. Fin whale 20 Hz calls are concentrated in the November-April time period near ACO. Estimated fin whale call parameters, including inter-call interval, are more clearly estimated from the recorded data following pruning to reduce the false positive rate.