ECE5530: Multivariable Control Systems II
Taught: Spr. 2004, Spr. 2002, Spr. 2001, Spr. 1999.

 Summary

To date, I have taught ECE5530 three times. The lecture notes for this course are largely taken from notes I took at a similar course at Stanford University. Some features of the current notes include:
A review of continuous-time and discrete-time control systems. State-space analysis is stressed, and Matlab control-system design procedures are reviewed (sections 1–3).
“Optimal control” (LQR) of systems when complete, perfect, measurements are available (section 4).
A review of random processes (section 5).
A development of least-squares (optimal) estimation theory for estimating a static state, given noisy measurements (section 6).
An application of least-squares estimation in system identification (section 7).
A development of least-squares estimation theory for estimating the state of a dynamic system, leading to development of the Kalman filter (sections 8 and 9).
Application of the Kalman filter to control theory, developing optimal control (LQR and LQG) of systems with incomplete, noisy measurements (section 10).
An introduction to robust control theory, explaining why LQG is not the ultimate control-system design, and why the control designer needs to take care when using it (section 11).
I typeset the lecture notes.

 Files available for viewing

 > View the ECE5530 Course Reader (311 pages; Requires Adobe Acrobat Reader–only available on CD-ROM dossier)

 < Go to ECE4899 page
 ^ Return to the “Documentation of Teaching” index
 ^ Return to the main Table of Contents