About the project
This project will develop machine learning (ML) techniques for processing acoustic data from sensor arrays, allowing the enhancement of signals and localization of sound sources. While ML has been used for single sensors, this project will create new methodologies for processing data from multiple sensors, improving underwater acoustic analysis.
We have been successfully developing methods for the analysis of acoustic data collected from a single hydrophone using advanced machine learning techniques. In many practical underwater scenarios, acoustic data is collected via an array of multiple hydrophones. This approach provides the ability to localize sounds and enhance them relative to background noise. However, the integration of ML into the processing chain for such arrays has largely been overlooked.
This project aims to bridge that gap by exploiting data from an array sited off the Sussex coast. The array collects acoustic data, transmits it ashore in real-time, and record it for further analysis. You will explore various approaches to applying ML data from acoustic arrays, beginning with classical processing techniques at the front end and progressively incorporating ML solutions throughout the processing chain.
This project will address key challenges in multichannel data fusion, aiming to develop innovative ML algorithms that can effectively process and analyse data from multiple hydrophones. By doing so, we hope to enhance the accuracy and robustness of underwater acoustic signal analysis, with potential applications in underwater communication, marine life monitoring, and environmental noise assessment.
You will have the opportunity to work with leading experts in the field, access advanced facilities, and contribute to innovative research that advances the field of underwater acoustics.
This is a collaborative project between the Institute of Sound and Vibration Research and the School of Ocean and Earth Science, in partnership with our industrial partner.
