AG Kommunikationstheorie

Thema:

Applications of robust optimization in signal processing

Abstract:

In standard (deterministic) optimization the problem related data is assumed to be fixed and known. However, there exist many problems with some uncertainty in the data, e.g., measurement errors, imperfect knowledge or wrong model assumptions. In contrast to standard optimization, robust optimization methods incorporate uncertainty about the data.

This talk will start with a short introduction on robust optimization. After that, we will present three different applications. Firstly, we will analyze the design of a robust FIR Chebyshev equalizer. Secondly, we will introduce a new robust design for allpass transformed filter banks. And finally, we present the detection problem of 16-QAM signaling in MIMO channels for a simple case of a uncertain channel matrix. The last section concludes the presented material and provides a small survey on current research activities related to robust optimization.



zurück zur Terminübersicht