Title | Lecture 01:Introduction- Control system design with incomplete information |
---|---|
Course | Adaptive Control Systems |
Institution | Binghamton University |
Pages | 15 |
File Size | 1.5 MB |
File Type | |
Total Downloads | 89 |
Total Views | 133 |
Introduction- Control system design with incomplete information...
ECE 517 LECTURE 01
EE 517 ADAPTIVE CONTROL Spring 2018 Description: Methods, models and applications of model reference and self-tuning adaptive control theory. Mathematical description, computer simulation, analysis, and design of adaptive control systems. Textbooks: (1) Y. D. Landau, "Adaptive Control. The Model Reference Approach", Marcel Dekker, Inc. (2) Astrom, K.J. and Wittenmark, B., "Adaptive Control," Addison-Wesley Publishing Company Coordinator: Victor A. Skormin, Ph.D., Distinguished Service Professor of Electrical Engineering. Office: ES 2316, Telephone: 607-777-4013, E-mail: [email protected] Prerequisites: Methods of analog and digital state-variable control, Scientific Computing/ Major Software Tools (MATLAB/ SIMULINK/ VISSIM/ CC) Topics: 1. Introduction. Control system design with incomplete information 2. The Model Reference (MR) approach in parameter estimation and control 3. Local parametric optimization in MR control 4. Application of Liapunov function in MR control 5. Application of positivity and hyperstability principle in MR control 6. Adaptive identifiers and state estimators utilizing the MR approach 7. Adaptive control systems utilizing the MR approach 8. Methods of Self-Tuning (ST) control 9. Recursive least squares method in ST control systems 10. Pole-zero placement in ST controllers (Diophantine equation) 11. Explicit self-tuning and its applications in ST control 12. Implicit self-tuning and its applications in ST control 13. On-line parameter optimization in control systems 14. Methods of gain scheduling 15. Case studies Assignments and Grading Policy: Test (topics 1-7) 10 Homework Assignments Final Exam/Project (topics 1-15)
- 30% - 30% - 40%
NOTE: All students must adhere to the Student Academic Honesty Code of the University and the Watson School, see: http://www.binghamton.edu/ece/grad/academic-honesty.html This policy is consistent with the principles of professional ethics of a practicing American engineer. 1
ECE 517 LECTURE 01
2
ECE 517 LECTURE 01
3
ECE 517 LECTURE 01
Parameters of the adaptive controller are adjustable. They may depend on time, current system behavior, and the most interesting, particular signals within the system. This ultimately leads to the system nonlinearity 4
ECE 517 LECTURE 01
5
ECE 517 LECTURE 01
MODEL REFERENCE (MR) CONTROL
SELF-TUNING (ST) CONTROL Reference signal _
Physical process with poorly known parameters
Parameter estimation procedure Estimated process parameters (Re)Design of the controller Parameters of the controller Adjustable CONTROLLER
6
ECE 517 LECTURE 01
GAIN SCHEDULING
7
ECE 517 LECTURE 01
MODEL REFERENCE ADAPTIVE CONTROL
8
ECE 517 LECTURE 01
9
ECE 517 LECTURE 01
MR FOR PARAMETER ESTIMATION (IDENTIFICATION): Parameters of a software simulator are adjusted to minimize the discrepancy between the output of the real process (MR) and the output of the software simulator (Adjustable System). Then parameters of the AS may (or may not) represent parameters of the MR
MR FOR STATE ESTIMATION (STATE OBSERVER): Adaptation signal “forces” the output of the software simulator (Adjustable System) to be close to the output of the real process (MR). Then the state vector of the simulator (AS) will the unknown state vector of the MR. Parametric adaptation version may be also available Physical process with poorly known parameters or states
Input signal
+
Software simulator with adjustable parameters
+ _
Error
Adaptation mechanism generating signals for parametric adaptation or signal adaptation
parametric adaptation signal adaptation
10
ECE 517 LECTURE 01
MR FOR CONTROL SYSTEM: Parameters of the controller are adjusted to minimize the discrepancy between the output of the control system (AS) and the output of the software simulator representing the desired system performance (MR) – in parametric adaptation. Adaptation signal “forces” the output of the control system (AS) to follow the output of the software simulator representing the desired system performance (MR) – in signal adaptation. Software simulator representing the desired properties of the closed-loop system
Input of the CL system (reference signal)
+
_
Physical controlled process with poorly known properties (poorly known and/or time-varying parameters)
+ _
Error
Adaptation mechanism generating signals for parametric adaptation or signal adaptation
Controller designed for a process with “best guess” properties
parametric adaptation signal adaptation
11
ECE 517 LECTURE 01
Control problem, state-variable definition, signal adaptation:
12
ECE 517 LECTURE 01
13
ECE 517 LECTURE 01
THE MR APPROACH IMPLEMENTS THE NEGATIVE FEEDBACK PRINCIPLE: Advantage of FB:
14
ECE 517 LECTURE 01
EE 517 Homework Assignment #1 Given the “best guess” transfer function of a controlled plant: G (s)
3s 2 8s 16 s3 10s2 2s 2
The design specifications call for the settling time of 2 sec and the overshoot of the system step response of approximately 3 % Problem 1. a) Design an output feedback controller. Verify your design by computer simulations for the “best guess” controlled plant. b) Assume the “true” plant to be GTRUE ( s)
2s 2 8s 5 Investigate the system operation with s3 2 s2 s 1
the “true” controlled plant. c) Develop a schematic of a model reference control system with parametric adaptation that could be applicable in this situation without the explicit definition of the adaptation mechanism (i.e. just represent it by a box). Describe the principle of operation of this system d) Develop a schematic of a model reference control system with signal adaptation that could be applicable in this situation without the explicit definition of the adaptation mechanism. Describe the principle of operation of this system Problem 2. a) Design a state-variable controller and a state observer. Verify your design by computer simulations for the “best guess” controlled plant. b) Assume the “true plant” to be G
TRUE
2 s2 8s 5 . Investigate the system operation with (s) 3 s 4s 2 s 1
the “true” controlled plant c) Develop the schematic of a model reference control system with parametric adaptation that could be applicable in this situation without the explicit definition of the adaptation mechanism. Describe the principle of operation of this system. d) Develop the schematic of a model reference control system with signal adaptation that could be applicable in this situation without the explicit definition of the adaptation mechanism. Describe the principle of operation of this system
15...