mae 280a linear dynamic systems
DESCRIPTION
MAE 280A Linear Dynamic Systems. Robert E Skelton, [email protected] , 858 822 1054 office hours (help session): 4:00-5:00 TU, 1804 EBU-1 text 1 : skelton, dynamic systems control , Wiley 1988 ISBN 0-471-83779-2 - PowerPoint PPT PresentationTRANSCRIPT
MAE 280A Linear Dynamic SystemsMAE 280A Linear Dynamic Systems
Robert E Skelton, [email protected], 858 822 1054 office hours (help session): 4:00-5:00 TU, 1804 EBU-1
text 1: skelton, dynamic systems control, Wiley 1988 ISBN 0-471-83779-2
text 2: skelton, iwasaki and grigoriadis, a unified algebraic approach to control design, Taylor and Francis 1998 ISBN 0-7484-0592-5
prerequisites: linear algebra, differential eqs
homework turned in on every Monday, late homework cannot be accepted (solutions will be posted on web)
hwnotebook = corrected homework solutions, bound in a notebook (due last week of class)
grade=.25exam + .25exam + .3final + .2hwnotebook
read assignment before lecture (come to class with questions in your head)
Homework 1: read chapter 3, text 1. Do exercises 3.2, 3.7, and 3.8
MAE280A Syllabus
1. How to get Models of Dynamics
2. How to get linear Models of Dynamics
3. How to get solution of linear models
4. How to measure performance of dynamic system
5. How to compute performance without solving the ODEs
6. How to modify performance with control
MAE280A Syllabus
1. How to get Models of Dynamics– Dynamics– State space models– Linearization
Modeling, the Most Difficult partModeling, the Most Difficult part
• How should we model a pendulum?• Should we model:
– Flexibility of rod?
– Bearing dynamics?
– Friction?
– Aerodynamic disturbances?
• Depends on control accuracy required of y• Control accuracy will depend on model, hence,• Modeling and Control Problem not independent• How do we get a model suitable for control design?• An ongoing research topic!
• Particle dynamics
• Put model in state form
How to get Dynamic Models
mg
x
y
x
y
m x u
um y mg
u x
y
x Ax Bu
Rigid body dynamics
Linearize about = 90, ux = 0
Put model in state form
How to get Dynamic Models
mg
x
y
( )x y
x
y
u u
u
J Cos Sin
m x
y
r
m mg u
u x
y
x Ax Bu
r
What is a Linear System?What is a Linear System?• A linear algebraic system
• A linear dynamic system
yAx
y
y
x
x
x
aaa
aaa
yxaxaxa
yxaxaxa
2
1
3
2
1
232221
131211
2323222121
1313212111
2
1
2
2
1
1
1
2221
1211
2222121
1212111
000
000
u
u
x
x
x
x
x
aa
aa
uxaxa
uxaxa
State Space form of Dynamic Models
( ) ( ) ( ), ( ) (( ) ( ) ( ) ( )) ( )x t x t u t y tA t B t C t Dx t u tt
( ) ( ) ( ), ( ) ( ) ( )x t x t u t y t x tA B C uD t
1( ) , (( ) ( ) ( ))G s G s C sIy s u s A B D
Nonlinear Models
LTI (Linear Time-Invariant) Models
LaPlace Transform of LTI Model
( ) ( ( ), ( )), ( ) ( ( ), ( ))x t f x t u t y t g x t u t
LTV (Linear Time-Varying) Models
State form of Dynamic Models, Discrete
1( ) ( ) ( )( ) ( ) , ( ) ( ) ( )( ) ( )k k k k k kk kk kx t x t u t y tA t B t x t tt D t uC
1( ) ( ) ( ), ( ) ( ) ( )k k k k k kx t x t u t y tA B C Dx t u t
1( ) , (( ) ( ) ( ))G z G z C zIy z u z A B D
Nonlinear Models
LTI (Linear Time-Invariant) Models, Discrete
z Transform of LTI Model, Discrete
( ) ( ( ), ( )), ( ) ( ( ), ( ))x t f x t u t y t g x t u t
LTV (Linear Time-Varying) Models, Discrete
What is a Linear System?What is a Linear System?
• The math model is an abstraction (always erroneous) of the Real System
• Are there any Real Systems that are linear?
Yes. Annually compounded interest at the bank.
0)1(1
System Time-cretelinear Dis A1
02
1112
0001
!
: )1(
)1()1(
)1(
year of beginningat principal
7% of rateinterest for 07.
k
k
P
Pr
kk
k
PrP
PrPrrPPP
PrrPPP
kP
r
Taylor’s seriesTaylor’s series
)()(2
1)()(
)()(
)()(
2
1)()()(
)()(
2
1))((
)()(
)(
2
1)(
)()(
)()()(
:,
,::...))(...(,...)!
1(...
)(!
1...)(
)(
2
1)(
)()()(:,
_
2
2___
_
22
_
11
22
2
12
221
2
21
2
_
22
_
11_
22
_
11
21
_
2222
2
2
221121
22
_
1121
2_
222
_
111
_
21
_
_
0
2_
2
2__1
xxx
fxxxx
x
fxf
xx
xx
x
xf
xx
xfxx
xf
x
xf
xxxxxx
xxx
xf
x
xfxf
xxx
xfxxxx
xx
xfxx
x
xfxx
x
xfxx
x
xfxfxf
RxRf
AfindAzzHomeworkxxzx
f
ia
zaxxx
f
ixx
x
xfxx
x
xfxfxfRxf
T
iTi
iT
Ti
ii
i
Nonlinear Systems/Taylor’s Series
)()(2
1)()()(
_
2
2___
xxx
fxxxx
x
fxfxfx T
MAE280A Syllabus
1. How to get Models of Dynamics2. How to get linear Models of Dynamics3. How to get solution of linear models
1. Coordinate Transformations2. The Liapunov Transformation3. The State Transition Matrix
HW2: chapter 4, exercises 4.11, 4.13, 4.14, 4.23, 4.25, 4.28
Coordinate Transformations
0
0
0
1 1
10
1 10 0
0 0 0
, ,
, [( ) ]
( ) ( ) ( ) ( ) ( ) .
( ) ( ) ( ) ( ) ( ) ( ) ( ) ( )
( , ) ( ) ( , ) ( ) ( ) , ( , )
integrating Hence,t
t
t
t
t
t
x Ax Bu x Tz T AT
Tz Tz ATz Bu z T AT T z Bu T Bu
z t z t T B u d
x t T t T t x t T t T B u d
t t x t t B u d t t
0
0 0
( ) ( , ),
( , )
A t t t
t t I
LTI Systems
0
0 0 0 0
0
0
( ) 2 20 0 0 0
0
( ) ( ) ( )
( ) ( )0
1( , ) [ ( )] ( ( ) ( ) ......
2
:
, ,
( ) ( ) ( )
-1factorial)
Hence
A t t i
i
A t t A t t A t t
tA t t A t
t
t t e A t t i I A t t A t t
proof
de Ae e I
dt
x t e x t e Bu
State: enough IC required to SOLVE the ODE (together with u(t))
LTI Solutions0
0
( )0
( )
1 2
0
( )
( )
, ( ) ( )
( , ) [ ...... ]
( , ) ( , ) ,r
i=1
ZIR:
ZSR:
Impulse Response:
OAT Impulse Response:
Deterministic Covariance:
u
A t t
t A t
t
At
Ati i n
Tu
At Ti i i i
x e x t
x e Bu d
x e Bu u t u t
x t i e bu B b b b
X x t i x t i dt r n
e bu u b
0
21
220
23
( ) ( )
0
( ) (
0
,
0 0
, 0 0
0 0
4.12 ( )
( ) (0) , lim (0) 0,
( ) lim
r
i=1
has solution
Hence, if
T
T
T T T
T
A tu
At T A t
T
tAt A t A t A t At A t
t
t A t A
t
e dt r n
u
e BUB e dt U u
u
Theorem text
Z AZ ZA W
Z t e Z e e We d e Z e
Z Z e We
)
0, lim 0
0 .
Hence, ifTt A A
td e We d Z
AZ ZA W
MAE 280A Outline• Modeling, introduction to state space models• linearization• vectors, inner products, linear independence• Linear algebra problems, matrices, matrix calculus• least squares• Spectral decomposition of matrices: Eigenvalues/eingenvectors• coordinate transformations• solutions of linear ode’s• controllability• pole assignment• observability• state estimation• stability• trackability• optimality• model reduction