On Friday, June 21, Prof. Tülay Adali from University of Maryland, Baltimore County will give a lecture at Tallinn University of Technology (Ehitajate tee 5) in room VI-229 at 13:00. The lecture is a part of IEEE Signal Processing Society’s Distinguished Lecturer Program and is titled as “Complex-valued Adaptive Signal Processing: When and how to take non-circularity into account?” Access is free for everybody.
Complex-valued signals arise frequently in applications as diverse as communications, radar, geophysics, optics, and biomedicine, as most practical modulation formats are of complex type and applications such as radar and magnetic resonance imaging lead to data that are inherently complex valued. There are two key issues in the statistical signal processing of complex-valued data: (1) making use of the complete statistical information by taking the potential noncircularity of the signal into account; and (2) optimization of real-valued cost functions with respect to complex parameters.
This talk first reviews recent advances related to both issues, which are coupled. It is shown that noncircularity is an intrinsic characteristic of many signals of practical interest, and when taken into account, the methods developed for their processing provide significant performance gains. On the other hand, by using Wirtinger calculus, optimization can be carried out in a way similar to the real-valued case without the need to evaluate derivatives separately with respect to real and imaginary parts.
Hence, expressions are kept simple and one does not need to invoke unrealistic assumptions such as circularity of the signals, an unrealistic assumption that helps with tractability but prevents one from making use of the information in the phase of the signal.
After a review of the basic tools for statistical characterization and optimization, the talk presents examples in adaptive filtering, model order selection, and blind source separation, and emphasizes when and how one should account for the potential noncircularity of the signal.
(more…)