ME190Y
Practical Control System Design: A Systematic Optimization
Approach [1 unit]
ONLINE RESOURCES:
CATALOG DESCRIPTION
The Youla-parametrization of all stabilizing controllers allows certain time-domain and frequency-domain closed-loop design objectives to be cast as convex optimizations, and solved reliably using off-the-shelf numerical optimization codes. This course covers the Youla parametrization, basic elements of convex optimizaation, and finally control design using these techniques.
COURSE PREREQUISITES
ME 132 or EECS 128 (EECS 20 may suffice) or similar introductory experience regarding feedback control systems. The student should understand basic properties of feedback systems, be comfortable with transfer function and differential equation descriptions of systems, and be familiar with typical feedback objectives such as disturbance rejection, command following, noise insensitivity and closed loop stability.
TEXTBOOK(S) AND/OR OTHER
REQUIRED MATERIAL
Notes and slides in class, both based on Linear Controller Design: Limits of
performance, by Stephen Boyd and Craig Barratt, available in pdf format at
Notes and slides in class, both based on Linear Controller Design: Limits of
performance, by Stephen Boyd and Craig Barratt, available in pdf format at
www.stanford.edu/~boyd/lcdbook
COURSE OBJECTIVES
DESIRED COURSE OUTCOMES
TOPICS COVERED
- (1 lecture) Review of design objectives in a feedback system, role of the sensitivity and complementary sensitivity functions;
- (3 lectures) Conservation laws in feedback systems: Bode integral theorem, maximum modulus theorem, closed-loop consequences of open-loop right-half-plane poles and zeros;
- (3 lectures) Youla parametrization of all stabilizing controllers for a plant;
- (2 lectures) Convex sets, convex functions, convex optimization;
- (2 lectures) Formulating control design problems as optimizations: pitfalls and best practices
- (3 lectures) Design examples, interpreting the results in classical control context;
- Last lecture: 1-hour final exam.
CLASS/LABORATORY SCHEDULE
1 hour of lecture.
CONTRIBUTION OF THE COURSE TO MEETING THE
PROFESSIONAL COMPONENT
The focus of the course is optimization-based control system design, using sound theory
to ensure that the optimizations are convex. The use of numerical tools to enable (and
automate) advanced design is prevalent in industry, and this class reinforces this notion.
RELATIONSHIP OF THE COURSE TO
ABET PROGRAM OUTCOMES
· an ability to apply knowledge of
mathematics, science, and engineering
· an ability to identify, formulate,
and solve engineering problems
· an ability to communicate effectively
· an ability to use the techniques,
skills, and modern engineering tools necessary for
engineering practice.
ASSESSMENT OF STUDENT PROGRESS TOWARD COURSE OBJECTIVES
PERSON(S) WHO PREPARED THIS DESCRIPTION:
Professor
Andrew Packard
