Array signal processing aims to infer information about a signal from spatio-temporal measurements of the associated wavefield on an array of sensors. Combining the outputs of the sensors on an array enhances the signal over noise and attributes directionality to the system, allowing to characterize not only the spectral content of the recorded wavefield but also the number and locations of the sources that produce it. Therefore, array signal processing is an active research area in diverse fields, e.g., in radar, seismic, and acoustic imaging. In acoustics, array signal processing is used for visualization of sound fields, identification and localization of sound sources, or acoustic imaging of objects that scatter sound. Machine learning methods allow parameter inference of increasingly complex systems by capitalizing on large amount of data. Hence, data-driven inference with machine learning methods has gained popularity over traditional model-based methods as it requires no particular assumptions or decisions about the underlying physics. This presentation will cover some model-based and data-driven methods for sonar signal processing. Synthetic aperture sonar processing, sparse reconstruction and representation learning for high-resolution acoustic imaging will be discussed in more detail, highlighting the experimental results and the scientific contributions.
Angeliki Xenaki received the Diploma degree in electrical engineering and computer science from the National Technical University of Athens, Greece, in 2007, and the M.Sc. and Ph.D. degrees in acoustics from the Technical University of Denmark (DTU), in 2010 and 2015, respectively. From 2012 to 2014 she was a visiting researcher at the Scripps Institution of Oceanography, University of California San Diego and from 2015 to 2016 she was a postdoctoral researcher at DTU. From 2016 to 2018 she was a research scientist at GN Hearing A/S, Denmark, specializing in array signal processing for hearing devices. In 2018, she joined the Centre for Maritime Research and Experimentation, Italy, as a Scientist working in the field of synthetic aperture sonar. Her research interests include signal processing, statistical modeling and machine learning.