ECE 2695: Adaptive Control (3 Credits, Fall 2008)

 

Description: Adaptation and learning play an essential role in biological systems, and these characteristics have been widely incorporated in modern control systems. This course introduces the general principles of adaptive control and learning. Topics to be covered include: real-time parameter estimation, self-tuning regulators, model-reference adaptive systems, adaptive control of nonlinear systems, practical aspects and implementation of adaptive control systems, introduction to computational learning theory and learning in neural systems, and an example of adaptive control by the cerebellum.

 

Prerequisite: Knowledge of feedback control systems (ECE 1673, notes available at http://www.engr.pitt.edu/electrical/faculty-staff/mao/1673/) and linear system theory (ECE 2646, notes available at http://www.engr.pitt.edu/electrical/faculty-staff/mao/2646/).

 

Time and Place: Monday 6:00 pm−8:30 pm, Benedum 426.

 

Instructor: Dr. Zhi-Hong Mao, (office) 434 Benedum Hall, (phone) 412-624-9674, (email) maozh@engr.pitt.edu, (office hours) Tuesday 4:00 pm−7:00 pm.

 

Text: S. Sastry and M. Bodson, Adaptive Control: Stability, Convergence, and Robustness, Prentice-Hall, 1989-1994, Sastry & Bodson, 1994, available for free download at http://www.ece.utah.edu/~bodson/acscr/.

 

Course Evaluation: Homework and class participation 30%, midterm exam 30%, and final exam 40%.

 

Tentative schedule for lectures (notes will be available at http://www.engr.pitt.edu/electrical/faculty-staff/mao/2695/ ):

 

Date

Topic

Reading

 

August 25

Lecture 1: Course organization and introduction to adaptive control

 

Chapter 0 of the textbook

September 1

Labor Day

 

 

September 8

Lecture 2: Mathematical description of systems and Lyapunov stability theory

 

Section 1.4

September 15

Lecture 3: System identification (I)

 

Section 2.0

September 22

Lecture 4: System identification (II)

 

Sections 0.3, 1.3, 2.0-2.3, and 2.5

September 29

Lecture 5: Model reference adaptive control (I)

 

Material about the MIT rule

October 6

Lecture 6: Model reference adaptive control (II)

 

Section 3.0

October 14

 

Midterm

(Tuesday, not Monday—October 13 is Fall Break)

 

 

October 20

 

Lecture 7: Model reference adaptive control (III)

 

TBA

October 27

 

Lecture 8: Self-tuning regulators (I)

 

TBA

November 3

 

Lecture 9: Self-tuning regulators (II)

 

TBA

November 10

Lecture 10: Robustness, practical aspects, and alternatives to adaptive control

 

TBA

November 17

Lecture 11: Introduction to computational learning theory and learning in neural systems

 

TBA

November 24

Lecture 13: Adaptive control by the cerebellum

 

TBA

December 1

Review of classes

 

 

December 8

Final exam