ELEC4410 Control System Design and Management

Purpose and Objectives

ELEC4410 Control System Design and Management builds on ELEC4400 Automatic Control offering a more advanced discussion of control systems, and introducing modern control techniques and practical implementation issues. In particular, students who successfully complete this course should have:

An exposure to modern control tools (e.g., internal model control, state feedback control, observers, etc.).
An in-depth introduction to the fundamental concepts of linear system theory (e.g., internal stability, realisations, controllability and observability, etc.).
A basic understanding of various factors which limit the achievable control system performance (e.g. time delays, non minimum phase zeros, etc.).
Experience in several lab implementations of control systems.
An initial exposure to various control implementation issues (e.g., sampled data systems, actuator saturation, anti-windup schemes, etc.).
An initial exposure to more advanced control system topics (e.g. multivariable systems, Kalman Filters, etc.).
Knowledge of case studies of successful modern control implementations.

Topics

Internal model control (16 hours), fundamental limitations in control design (5 hours), system identification (2 hours), state space system theory (21 hours), state space control design (14 hours), optimal estimation (5 hours), industrial case study (2 hours).

Assumed Knowledge

It is assumed an active knowledge of linear algebra and Laplace transforms, and an elemental knowledge of complex variables and linear ordinary differential equations. It is also assumed an active knowledge of transfer functions, frequency response, feedback principles, closed-loop stability, and basic control design issues for SISO systems, as given in ELEC4400.

Syllabus

Introduction and motivation to ELEC4410
Mathematical description of systems
1. A taxonomy of systems
2. Linear time invariant systems
3. Discrete-time systems
Control design via affine parameterisations (Internal Model Control)
1. Affine parameterisation for stable systems
2. PID synthesis via the affine parameterisation
3. Affine parameterisation for systems with time delays
4. Undesirable closed-loop poles
5. Saturation and anti-windup
6. Affine parameterisation for unstable systems
7. Affine parameterisation for MIMO systems
Fundamental limitations in control design
1. Sensors, actuators, perturbations and model errors
2. Structural limitations
  1. Delays
  2. Unstable poles
  3. Non-minimum phase zeros
3. Design tradeoffs in the step response
  1. Interpolations constraints
  2. Design specifications
  3. Design limitations
Elements of system identification

Introduction
Least squares model fitting

Introduction to state space system theory
1. Solution of LTI state equations
  1. Discretisation
  2. Discrete-time state equations
2. Realisations
3. Equivalent state space equations
4. Stability
  1. External and internal stability
  2. Lyapunov Theorem
5. Controllability
6. Observability
7. Canonical forms
8. Discrete-time state equations
  1. Controllability after sampling
Control design via state space methods
1. State feedback
2. Regulation and tracking
  1. Robust tracking: integral action
3. State estimation
4. Feedback from estimated states
  1. The Separation Principle
  2. Design considerations
5. MIMO state feedback
6. MIMO state estimation
7. MIMO feedback from estimated states
Introduction to optimal control
1. The basic optimal control problem
2. The Kalman Filter

Teaching Modes

The material is presented in a number of ways based around a lecture format, including case studies, and supplemented with Tutorials, Computer Simulations and Laboratories to reinforce student learning.

Lecture Times and Places

Activity	Day	Time	Room
Semester 2 - 2003
Lecture	Monday	11.00 - 12.00	[EAG01]
and Tutorial	Wednesday	08.00 - 10.00	[EF14]
and Lecture	Thursday	13.00 - 15.00	[EAG01]

First lecture 21 of July. Last lecture 6 of November.

Grading Policies

The course will be assessed as follows:

Assignments (15%)
Labs (10%)
Quiz (15%)
Exam (60%)

The Quiz will be on Monday 15th September 2003, 11am to 12pm, ES204. It will be open book - you are welcome to bring lecture notes and/or reference books. It will cover the material of the first eight weeks (up to and including Wednesday 10th September 2003), with emphasis on lecture and tutorial material.

Workload

Similar workload guidelines to those indicated by Prof. Brett Ninness for ELEC2400 apply for ELEC4410, that is:

As a ten credit point subject, ELEC4410 is one quarter of a full-time load of 40 credit points per semester. If a full time study load is equated to 40 hours per five day week (an arguably modest figure, particularly for engineering students), then ELEC4410 should be consuming (at least) 10 hours and more than one full day per week - every week.

Given that there are one one-hour and one two-hour lectures, and one two-hour tutorial session each week, students should dedicate a further five hours per week of non face-to-face time (that is, private study) to ELEC4410.

Acknowledgement

Much of the course structure and material builds on the design of ELEC4410 2002 by Dr. Will Heath's, whom the current instructors gratefully acknowledge.

ELEC4410 ­ Control System Design and Management