Numerics for Filtering and Identification

August 20-24, 2012

General Information

The course is given August 20-24 by Michel Verhaegen, and it will be held at the department of Automatic Control

Contents

The course covers the use of linear algebra in solving filtering and identification problems. The main ingredient is to study a variety of signal processing problems, such as Kalman filtering and subspace identification in the context of the weighted linear least squares problem. Attention will also be given to numerical aspects such as reliability and efficiency. The lectures consists of two parts: Lectures in the morning and exercise sessions in the afternoon.

Organization and Examination

• Lectures: 0815-1200 Monday to Friday
• Location: Visionen, Building B, Ground Floor (Level 2), Campus Valla, Linköping. Map (on Monday 10-12 the lecture will be in Signalen)
• Exercises: Participants will receive exercises to be discussed and solved each afternoon Monday to Thursday
• Course credits: 3hp
• Examination: Completion of the afternoon exercise problems

Lectures

Slides for lectures 1 and 2 for Monday 20 can be found here and here.
Slides for lectures 3 and 4 for Tuesday 21 can be found here and here.
Slides for lectures 5 and 6 for Wednesday 22 can be found here and here.
Slides for lectures 7 and 8 for Thursday 23 can be found here and here.
Slides for lectures 9 and 10 for Friday 24 can be found here and here.

Exercises

Exercise 1 can be found here.
Exercise 2 can be found here.
Exercise 3 can be found here with m-file.
Exercise 4 can be found here with m-file Example2.m and m-file Example3.m

Course Literature

The following book is recommended

Matlab code

Code used for demos during the lectures.
Acker_test.m
RLS_demo.m
RangeDemo.m
SensNeq.m.
Spiegelman.m
rrqrdemo.m
linCost.m
nlCost.m
Comp_Obs_SRCF.m
SubidPlot.m
SIDemo.m
subid_basic.m

Page responsible: Torkel Glad
Last updated: 2024-08-20