| Dozent: | Prof. Dr.-Ing. Walter Kellermann | |
| Übungsleiter: | M.Sc. Roland Maas | |
| Termin | Vorlesung: | Block course |
| Leistungspunkte: | 2.5 ECTS | |
| Voraussetzungen: | Basics of digital signal processing and probability and random processes | |
| Sprache der Veranstaltung: | English or German | |
Contents
 |
 |
|---|
Parameter estimation is of paramount importance in signal processing and communications. In almost any product related to these areas, algorithms for the estimation of unknown parameters are employed. This seminar starts with the general estimation problem - extracting parameter values from an observed data set in a statistically optimum sense - and with a specification of the differences between estimation theory and related fields such as hypothesis testing, decision theory and pattern recognition. The main concepts of estimation theory are developed along with practical applications of parameter estimation algorithms in signal processing and communications, such as
- signal estimation for statistically optimum filtering (e.g., for signal prediction, speech enhancement, communication channel equalization),
- carrier frequency estimation for digital transmission,
- parameter estimation for signal models (e.g., Gaussian Mixture Models, Hidden Markov Models),
- source localization (time difference of arrival estimation).
As the basis of these applications we will first study classical estimation approaches including general minimum variance unbiased estimation, best linear unbiased estimation, maximum-likelihood estimation, MMSE signal estimation and the method of least squares. The second part of the seminar focuses on Bayesian estimation and its differences to classical approaches. General and linear Bayesian estimators are discussed including MMSE and MAP estimators. The Cramer-Rao Bound is derived as a lower bound for the variance of any unbiased estimator.
Dates
block course, first meeting: October 21, 10:00 a.m., room N 6.17
Registration
via email to maas@LNT.de
Audience
Bachelor and Master students
Organization
Prof. Dr.-Ing. Walter Kellermann
Prof. Dr.-Ing. Wolfgang Gerstacker
Roland Maas, M.Sc.
Dipl.-Ing. Michael Ruder
Nach oben