THE UNIVERSITY OF THE WEST INDIES ST. AUGUSTINE, TRINIDAD AND TOBAGO, WEST INDIES FACULTY OF ENGINEERING Department of Electrical and Computer Engineering MASc. Electrical and Computer Engineering
ECNG 6603 Modern Control Strategies Semester II: 2012 - 2013
1
GENERAL INFORMATION
Course level: Course status:
Level M Core Elective Occasional
Semester(s) offered: Delivery mode:
Credits:
3
Estimated total study hours1:
Estimated enrollment:
6-10 students
Course Dependencies2
Pre-Requisites – Math courses. Other – None
Recommended prior knowledge skills3:
I Lecture Online Lab Other 3hrs (weekly) 39hrs (semester)
Upon entering this course, students should be able to: Demonstrate understanding of the fundamental theory of differential equation, system stability, complex numbers, and linear algebra. Perform numerical simulations under Matlab/Simulink environment with basic knowledge of Control System Toolbox of Matlab/Simulink. Apply basic results obtained in linear-control-systems course.
More information on UWI regulations and policies can be found on the web site of the course.
Course Staff
Dr. Salim Ibrir
2 2.1
Position/Role
Senior Lecturer
E-mail
[email protected] [email protected]
Office Phone Ext. 320 83147
Office Hours Time Table – Office Door
COURSE OVERVIEW Course Description
The aim of this course is to introduce graduate students to nonlinear, robust, adaptive, and neural-network control using modern control techniques in state space. The course chapters cover system identification, robust control design, adaptive control, and feedback design using convex optimization. The identification part constitutes one of the key and straightforward solutions to make the dynamics of a system with uncertainties a well-defined one. However, robust control is mainly created to stabilize parameterized systems that are not well identified due to the modeling imprecision, and the presence of unmolded dynamics. The aforementioned topics are discussed both in continuous and discrete time. Throughout this course, a major concern will be paid to the control of uncertain linear systems with a little emphasis on robust estimation and observer design. 2.2
Course Rationale
Due to the nature of existing physical systems, nonlinearities and uncertainties that are present in the system dynamics cannot be handled by static linear feedbacks. The potential need to overcome this inherent dynamics to maintain the system stability and achieve desired performances has created an increasing interest to develop reliable control techniques where uncertainties are handled through different points of views. This course is proposed to let the student aware about existing modern techniques with special interest to some classes of systems. A major concern of this course is to put more light on adaptive and nonlinear control systems that are encountered in industrial engineering and others that can be satisfactorily implemented in the Caribbean region. This course is considered as a core course in the postgraduate programme leading to a Master of Applied Science (MASc) in Electrical & Computer Engineering with majors in Control Systems.
2.3
Course Aims
The basic aims of Modern Control Strategies are: 2.4
System identification with classical numerical procedures; Design of robust feedback laws under system uncertainties and model imprecision; System linearization by state feedback and change of coordinates; Design of adaptive control systems for nonlinear uncertain systems; Neural network control. Course Learning Outcomes
The 5 (five) course-learning outcomes are summarized in the following Table.
Upon successful completion of ECNG 6603, attendees will be able to: 1. Identify the system parameters with static and recursive numerical techniques by the use of Matlab/Simulink software. 2. Construct, demonstrate, and implement robust feedbacks for linear systems with structured-type uncertainties. Implement robust feedbacks using convex optimization. 3. Synthesize of linearized dynamics around equilibrium points. Highlight the need of state transformation to cancel the system nonlinearities. Analyze and synthesize nonlinear feedbacks for a class of linearizable nonlinear systems with numerical simulations using Matlab/Simulink. Demonstrate the effectiveness of nonlinear feedback through real-case examples.
Cognitive Level Knowledge and comprehension Knowledge, synthesis, and application Synthesis and applications
4. Design and discuss nonlinear adaptive systems with stability analysis of the closed-loop systems. Perform numerical simulations with Matlab/Simulink.
Synthesis, evaluation and application
5. Analysis of system approximation with neural nets. Design, assess, simulate, and synthesize neural network-based feedback for system stability and trajectory tracking. Demonstrate the effectiveness of the design by numerical simulations (Matlab/Simulink).
Analysis and comprehension
3 3.1
COURSE ASSESSMENT Breakdown of Assessment Artifacts and Linkage to Course Learning Outcomes
The details of the learning outcomes are given in section 2.4 and they are classified into five independent Los.
Assessment Artifact
Course Work 1
Course LOs Covered LO1 LO2 LO3 LO4 LO5 ⊗
Course Work 2
⊗
⊗
Required to
Weight %
Details (e.g. type - written, oral, practical; duration)
YES
40/3
YES
40/3
Assignment to be done at home. The assignment is mainly based on system identification by learning a model from empirical data with simulation using Matlab/Simulink. Assignment to be done at home. Assignment based on the design of nonlinear feedback design and state transformation. Numerical simulation using Matlab/Simulink. Assignment covering all the learning outcome of the course. Written exam of 3 hours. Closed-book exam. See UWI regulations at the end of this outline.
pass course
Course Work 3
⊗
⊗
⊗
⊗
⊗
YES
40/3
Final Exam
⊗
⊗
⊗
⊗
⊗
YES
60
TOTAL1 Key 1 Distribution of LO percentages indicate approximate contribution of each LO component to final course grade. (Optional) Assessment provides full coverage of the LO Assessment provides partial coverage of the LO
100%
ECNG 6603 Course Plan – (Last Reviewed: 20-Jan-13)
4
4 4.1
COURSE DELIVERY Schedule and Venue
Component Lecture Lab Tutorial Other Activities 4.2
Schedule 3hr/week Not Available
Venue
Particulars
To be defined according to the student need.
Target Delivery Schedule
Week Lecture Topics
Learning Resources Course Notes
1-2
Identification techniques: Modeling and uncertainties; Model selection; Parameter identification; Least-Squares estimation; Recursive LeastSquares estimation; Least Squares with forgetting factor; Parameter identification using numerical techniques.
3-3
Numerical simulation and control design: A Matlab/Simulink tutorial: Course Notes Introduction to Matlab/Simulink; System modeling; Numerical simulation with Simulink; data handling; Different types of simulation. Simulation of nonlinear sysems. Robust control of linear systems: Types of uncertain systems; ; Modeling Course Notes of uncertain systems; Polytopic uncertainties; Norm-bounded uncertainties; Stability of uncertain systems; Stabilizability of uncertain systems. Linear matrix inequalities versus the Algebraic-Ricatti-equation-based designs; Convex functions; Convex optimization problems; Positive definite and negative definite matrices; Minimizing a performance index under the
4-6
LOs Addr essed4 LO1
Other Activities5
Assessment Exercises6 Assigned Due Course Work 1 given on week 2
All Los.
LO2
Course Work 2 given on week 5
Course Work 1 due on week 6.
The course notes are the most important documents to be read carefully.
ECNG 6603 Course Plan – (Last Reviewed: 20-Jan-13)
5
dynamics constraints; Linear optimization and linear matrix inequalities; Solving linear matrix inequalities; The LMI package of Matlab; Examples. 7-9
10-12
13
Nonlinear and adaptive Control: Different notions of stability; System linearization around an equilibrium point; Lyapunov stability and Lypunov control designs; System linearization with state transformation and feedback; Nonlinearity cancelling; Uncertain nonlinear systems; Robust control versus adaptive control; Conception of an adaptive control law; Stability analysis; Adaptive trajectory tracking. Neural network approximation and control: Definition of a neural net; System modeling using neural networks; Differential neural networks; Controller design using neural networks; Lyapunov-based analysis of Neural-Network controllers. Model-predictive control; Matlab implementation of neural networks. Examples.
Course Notes
LO3 and LO4
Course Notes
LO5
Review Session
All LOs.
ECNG 6603 Course Plan – (Last Reviewed: 20-Jan-13)
Course Work 3 assigned on week 8
Preparatio n for the final exam. Handling the students issues
Course Work 3 due on week 13
6
5 5.1
RESOURCES` Required Readings
[1] Hassan K. Khalil, Nonlinear Systems, Prentice Hall, Third edition, 2002. [2] Lennart Ljung, System Identification: theory for the user, second edition, Prentice Hall, 1999. [3] Karl J. Astrom, and Bjorn Wittenmark, Adaptive Control, Second Edition, Dover Publications, INC, Mineola, New-York, ISBN-13: 978-0-486-46278-3, year 2008. [4] The pdf files distributed or put on the web site are considered as required reading documents. [5] Katsuhiko Ogata, Modern control engineering, Prentice Hall, 2002. 5.2
Recommended Readings
[6] Katsuhiko Ogata, Matlab for control engineers, Prentice Hall,2008; [7] Brian D. O. Anderson, and John B. Moore, Optimal control: linear quadratic methods, Prentice Hall International, Inc, 1989. [8] Arthur E. Bryson Jr., and Yu-Chi Ho, Applied optimal control, optimization, estimation, and control, Taylor & Francis, 1975. [9] Jeffrey B. Burl, Linear optimal control, ܪଶ and ܪஶ methods, Addison-Wesley, 1999.
6 6.1
STUDENT CONDUCT Student Attendance
Rule 7 (d) in The Faculty of Engineering: Undergraduate Regulations 2008-2009: “In order for a student to qualify for credit and/or final examination of a course, the student would have had to have a minimum of 75% attendance for that course.” 6.2
Extended Absence from Class
Please note the University’s policy on absence from class as documented in Examination Regulations for First Degrees, Associated Degrees, Diplomas and Certificates 2006/2007: 31. Any candidate who has been absent from the University for a prolonged period during the teaching of a particular course for any reason other than illness or whose attendance at prescribed lectures, classes, practical classes, tutorial, or clinical instructions has been unsatisfactory or who has failed to submit essays or other exercises set by his/her teachers, may be debarred by the relevant Academic, on the recommendation of the relevant Faculty Board, from taking any University examinations. The procedures to be used shall be prescribed in Faculty Regulations. ECNGXXXX– (Last Reviewed: 20-Jan-13)
7
33. (ii) In cases of illness the candidate shall present to the Campus Registrar…a medical certificate, as proof of illness, signed by the University Health Officer or by another medical practitioner approved for this purpose by the University. The candidate shall send the medical certificate within seven days from the date of that part of the examination in which performance of the candidate is affected. A certificate received after this period will be considered only in exceptional circumstances. 6.3
Missed Coursework Exams Policy
Rule 10 in The Faculty of Engineering: Undergraduate Regulations 2008-2009: “A student who is absent from written coursework tests for grave medical reasons, as prescribed in the University Regulations, shall be graded on the tests he/she has taken as if such tests constitute the full test requirement provided that the tests not taken constitute no more than 20% of the total mark for all the tests7. If the tests not taken constitute more than 20% of the total mark for all the tests, the candidate shall have to take make-up tests at a later date.” 6.4
Coursework Late Submission Policy
According to Rule 11 (b) in The Faculty of Engineering: Undergraduate Regulations 2008-2009, Students are required to submit coursework by the prescribed date. 6.5
Policy on Re-Use of Previous Coursework Grade
Rule 11 (a) in The Faculty of Engineering: Undergraduate Regulations 2008-2009: “Students who fail the examination in any course, but pass the coursework may be exempted from redoing only those sections of the coursework comprising laboratory experiments, workshop and/or field exercises.”
6.6
Statement of Academic Honesty
Academic dishonesty has grave consequences which may include receiving “no grade” on the assignment, debarment from class, or even expulsion from the University. Academic dishonesty is a serious offense which should not be taken lightly. Cheating and plagiarism are both forms of academic dishonesty. Rule 32 in The Faculty of Engineering: Undergraduate Regulations 2008-2009:
“ Cheating, Plagiarism and Collusion are serious offences under University Regulations. (a)
7
Cheating is any attempt to benefit one's self or another by deceit or fraud.
20% of entire course weighting. ECNGXXXX– (Last Reviewed: 20-Jan-13)
8
(b)
Plagiarism is the unauthorised and/or unacknowledged use of another person's intellectual efforts and creations howsoever recorded, including whether formally published or in manuscript or in typescript or other printed or electronically presented form and includes taking passages, ideas or structures from another work or author without proper and unequivocal attribution of such source(s), using the conventions for attributions or citing used in this University. Plagiarism is a form of cheating.
(c)
For the purposes of these Regulations, ‘collusion’ shall mean the unauthorised or unlawful collaboration or agreement between two or more students in the preparation, writing or production of a course assignment for examination and assessment, to the extent that they have produced the same or substantially the same paper, project report, as the case may be, as if it were their separate and individual efforts, in circumstances where they knew or had reason to know that the assignment or a part thereof was not intended to be a group project, but was rather to be the product of each student’s individual efforts. Where two or more students have produced the same or substantially the same assignment for examination and assessment in circumstances that the assignment was to be the product of each student’s individual efforts, they shall receive a failing grade in the course. ”
According to the University of the West Indies’ Code of Principles and Responsibilities for Students, a student may appear before a disciplinary committee for the following misconduct: “Item 5. Presentation of the work of any other person as a student's own work. This includes plagiarism from unpublished and/or electronic sources.”
Every student submission made to the Department of Electrical and Computer Engineering is subject to examination through an electronic plagiarism checker.
ECNGXXXX– (Last Reviewed: 20-Jan-13)
9
