Estimation theory for signal processing and communications

Aalborg University

Description:

An estimator is a way to “guess” the value of an entity from data stochastically related to it.  Estimation theory, which may be considered as an element of “machine learning”, is highly relevant for researchers who are often facing problems of inferring parameters from data gathered from measurements or simulations.  This is in particular true for researchers within communications and signal processing.

The aim of this course is to give an overview of the field of estimation theory and the tools to derive and evaluate new estimators for specific engineering applications.  The course is centered around three main topics:

1. Introduction to point estimators and performance bounds
2. Connection between estimation theory and information theory
3. Bayesian inference on graphical models

The teachers of the course are all actively developing estimators as a part of their research. Therefore, we will during the course give examples of how we use the theory in our own research work.

Prerequisites:

Background in probability theory and statistics at the level taught at the MSc at AAU. Some background in stochastic processes is useful, but not required.

Learning objectives:

The intended learning outcome is that the participants will be able to apply methods from the course to problems within their own fields of research.

Teaching methods:

The course will consist of lectures combined with exercises and work on the students own estimation problem. As a part of the exercises, the students will formulate an estimation/inference problem related to their field of research and to apply some of the methods to that problem.

Criteria for assessment:

Active participation is expected. At the end of the course, the students will present their results of the work during the course and give/receive feedback from lecturers and peers.

Key literature:

The course will be based on lecture notes writing by the teachers supplemented by selected chapters from text books by SM Kay, Scharf, van den Boos, Bishop and Cover&Thomas (TBC).

Back