Signal Processing Algorithms (Master, M.Sc.)

Typ: Vorlesung + Übung/Tutorium
SWS: 5
Credit Points: 6

Kursbeschreibung / -kommentar

The course shall provide knowledge in the field of linear and nonlinear digital signal processing algorithms and their hardware implementations. Especially nonlinear concepts in digital signal processing are of actual interest in a wide area of signal, bio-systems, image and multimedia processing applications. After the course the student is able to analyze and map algorithms onto different hardware platforms, such as DSPs and ASICs (FPGAs).
Contents:One keypoint is the implementation of algorithms in DSPs and FPGAs or ASICs with the help of linear systolic arrays (LSAs). The theory and practical aspects of systolic designs, optimal array scheduling, Cut-Set-Retiming procedures and the design of processing elements (PEs) as well as hardware-software co-design will be highlighted. Different LSA-implementations for hardware accelerators will be discussed:

* Correlation and Convolution
* FIR filters
* Wavelets
* Spectral transforms
* 1D- and 2D- position invariant transforms (PIT) and Fuzzy-Pattern-Classification (FPC)