Trial lecture
“Deep Learning for Direction of Arrival Estimation”
Time and place: June 25, 2025 10:15. Kristen Nygaards sal (5370), Ole-Johan Dahls hus.
Main research findings
In the current thesis, I provide novel approaches to estimate the performance bounds for both adaptive and conventional beamformers. Beamforming combines the signal of multiple transducers to form a narrow beam pointing towards the target. Such an array can consist of antenna elements as used in mobile communication, or ultrasonic sensors as used in sonars. Adaptive beamformers have, in general, a better potential to suppress unwanted signals then conventional beamformers, but they also tend to be more sensitive to calibration errors. My results contribute to a better understanding of beamforming performance. I foresee these results to be applicable in the development of novel and efficient beamformers for specific applications, such as bathymetric sonar, medical ultrasound, imaging radar and mobile communication. The current thesis provides both novel theoretical results derived from first principles, and open-source tools for evaluating beamformers. The results provide a balanced emphasis on the accuracy and robustness of beamformers, which encourages engineers to pay equal consideration to these aspects when choosing the appropriate beamformer for an application.
Adjudication committee
- Professor Kathleen E. Waage, George Mason University, USA
- Senior Researcher Carl-Inge Colombo Nilsen, Norwegian Computer Center, Norway
- Associate Professor Anis Yazidi, University of Oslo, Norway
Supervisors
-
Professor Andreas Austeng, University of Oslo, Norway
-
Associate Professor Roy Edgar Hansen, University of Oslo, Norway
-
Associate Professor Tor Inge Birkenes Lønmo, Kongsberg Maritime, Norway
Chair of defence
- Professor Emeritus Dag Langmyhr, Department of Informatics, University of Oslo
Contact information at Department: Mozhdeh Sheibani Harat