Topic #13
16.31 Feedback Control
State-Space Systems
? Full-state Feedback Control
? How do we change the poles of the state-space system?
? Or, even if we can change the pole locations.
? Where do we change the pole locations to?
? How well does this approach work?
Copy right 2001 by Jon at h an H ow.
1
Fall 2001 16.31 13–1
Full-state Feedback Controller
? Assumethatthesingle-inputsystemdynamicsaregivenby
˙x = Ax+Bu
y = Cx
sothat D =0.
– The multi-actuator case is quite a bit more complicated as we
wouldhavemanyextradegreesoffreedom.
? Recallthatthesystempolesaregivenbytheeigenvaluesof A.
– Want to use the input u(t) to modify the eigenvalues of A to
changethesystemdynamics.
? Assumeafull-statefeedbackoftheform:
u = r?Kx
where r issome reference input andthe gain K isR
1×n
– If r =0,wecallthiscontrollera regulator
? Findtheclosed-loopdynamics:
˙x = Ax+B(r?Kx)
=(A?BK)x+Br
= A
cl
x+Br
y = Cx
Fall 2001 16.31 13–2
? Objective: Pick K sothat A
cl
hasthedesiredproperties, e.g.,
– A unstable,want A
cl
stable
– Put2polesat?2±2j
? Notethatthereare n parametersin K and n eigenvaluesin A,so
itlookspromising,butwhatcanweachieve?
? Example #1: Consider:
˙x =
bracketleftbigg
11
12
bracketrightbigg
x+
bracketleftbigg
1
0
bracketrightbigg
u
– Then
det(sI ?A)=(s?1)(s?2)?1=s
2
?3s+1=0
sothesystemisunstable.
– De?ne u =?
bracketleftbig
k
1
k
2
bracketrightbig
x =?Kx,then
A
cl
= A?BK =
bracketleftbigg
11
12
bracketrightbigg
?
bracketleftbigg
1
0
bracketrightbigg
bracketleftbig
k
1
k
2
bracketrightbig
=
bracketleftbigg
1?k
1
1?k
2
12
bracketrightbigg
– Sothenwehavethat
det(sI ?A
cl
)=s
2
+(k
1
?3)s+(1?2k
1
+k
2
)=0
– Thus, by choosing k
1
and k
2
, we can put λ
i
(A
cl
)anywherein
thecomplexplane(assumingcomplexconjugatepairsofpoles).
Fall 2001 16.31 13–3
? Toputthepolesats =?5, ?6,comparethedesired characteristic
equation
(s+5)(s+6)=s
2
+11s+30=0
withtheclosed-loopone
s
2
+(k
1
?3)x+(1?2k
1
+k
2
)=0
toconcludethat
k
1
?3=11
1?2k
1
+k
2
=30
bracerightbigg
k
1
=14
k
2
=57
sothat K =
bracketleftbig
14 57
bracketrightbig
,whichiscalled Pole Placement.
? Ofcourse,itisnotalwaysthiseasy,astheissueofcontrollability
mustbeaddressed.
? Example #2: Considerthissystem:
˙x =
bracketleftbigg
11
02
bracketrightbigg
x+
bracketleftbigg
1
0
bracketrightbigg
u
withthesamecontrolapproach
A
cl
= A?BK =
bracketleftbigg
11
02
bracketrightbigg
?
bracketleftbigg
1
0
bracketrightbigg
bracketleftbig
k
1
k
2
bracketrightbig
=
bracketleftbigg
1?k
1
1?k
2
02
bracketrightbigg
sothat
det(sI ?A
cl
)=(s?1+k
1
)(s?2)=0
Sothefeedbackcontrolcanmodifythepoleats =1,butitcannot
movethepoleat s =2.
? The system cannot be stabilized with full-state feed-
back control.
Fall 2001 16.31 13–4
? Whatisthereasonforthisproblem?
– Itisassociatedwithlossofcontrollabilityofthe e
2t
mode.
? Considerthebasiccontrollabilitytest:
M
c
=
bracketleftbig
B AB
bracketrightbig
=
bracketleftbiggbracketleftbigg
1
0
bracketrightbigg bracketleftbigg
11
02
bracketrightbiggbracketleftbigg
1
0
bracketrightbiggbracketrightbigg
Sothat rank M
c
=1< 2.
? Considerthe modal test todevelopalittlemoreinsight:
A =
bracketleftbigg
11
02
bracketrightbigg
,decomposeas AV = VΛ ?Λ=V
?1
AV
where
Λ=
bracketleftbigg
10
02
bracketrightbigg
V =
bracketleftbigg
11
01
bracketrightbigg
V
?1
=
bracketleftbigg
1 ?1
01
bracketrightbigg
Convert
˙x= Ax+Bu
z=V
?1
x
?→ ˙z =Λz+V
?1
Bu
where z =
bracketleftbig
z
1
z
2
bracketrightbig
T
.But:
V
?1
B =
bracketleftbigg
1 ?1
01
bracketrightbiggbracketleftbigg
1
0
bracketrightbigg
=
bracketleftbigg
1
0
bracketrightbigg
sothatthedynamicsinmodalformare:
˙z =
bracketleftbigg
10
02
bracketrightbigg
z+
bracketleftbigg
1
0
bracketrightbigg
u
? WiththiszerointhemodalB-matrix,caneasilyseethatthemode
associatedwiththe z
2
stateis uncontrollable.
– Must assume that the pair (A, B) are controllable.
Fall 2001 16.31 13–5
Ackermann’s Formula
? Thepreviousoutlinedadesignprocedureandshowedhowtodoit
byhandforsecond-ordersystems.
– Extendstohigherorder(controllable)systems,buttedious.
? Ackermann’s Formula gives us a method of doing this entire
designprocessisoneeasystep.
K =
bracketleftbig
0 ... 01
bracketrightbig
M
?1
c
Φ
d
(A)
where
–M
c
=
bracketleftbig
BAB...A
n?1
B
bracketrightbig
– Φ
d
(s) is the characteristic equation for the closed-loop poles,
whichwethenevaluatefors = A.
– Itisexplicitthatthesystem must be controllablebecause
weareinvertingthecontrollabilitymatrix.
? Revisitexample # 1: Φ
d
(s)=s
2
+11s+30
M
c
=
bracketleftbig
B AB
bracketrightbig
=
bracketleftbiggbracketleftbigg
1
0
bracketrightbigg bracketleftbigg
11
12
bracketrightbiggbracketleftbigg
1
0
bracketrightbiggbracketrightbigg
=
bracketleftbigg
11
01
bracketrightbigg
So
K =
bracketleftbig
01
bracketrightbig
bracketleftbigg
11
01
bracketrightbigg
?1
parenleftBigg
bracketleftbigg
11
12
bracketrightbigg
2
+11
bracketleftbigg
11
12
bracketrightbigg
+30I
parenrightBigg
=
bracketleftbig
01
bracketrightbig
parenleftbiggbracketleftbigg
43 14
14 57
bracketrightbiggparenrightbigg
=
bracketleftbig
14 57
bracketrightbig
?AutomatedinMatlab: place.m& acker.m(see polyvalm.mtoo)
Fall 2001 16.31 13–6
? Wheredidthisformulacomefrom?
? For simplicity, consider a third-order system (case #2), but this
extendstoanyorder.
A =
?
?
?a
1
?a
2
?a
3
100
010
?
?
B =
?
?
1
0
0
?
?
C =
bracketleftbig
b
1
b
2
b
3
bracketrightbig
– Seekeybene?tofusing control canonical state-spacemodel
– This form is useful because the characteristic equation for the
systemisobvious?det(sI ?A)=s
3
+a
1
s
2
+a
2
s+a
3
=0
? Canshowthat
A
cl
= A?BK =
?
?
?a
1
?a
2
?a
3
100
010
?
?
?
?
?
1
0
0
?
?
bracketleftbig
k
1
k
2
k
3
bracketrightbig
=
?
?
?a
1
?k
1
?a
2
?k
2
?a
3
?k
3
100
01
?
?
sothatthecharacteristicequationforthesystemisstillobvious:
Φ
cl
(s)=det(sI ?A
cl
)
= s
3
+(a
1
+k
1
)s
2
+(a
2
+k
2
)s+(a
3
+k
3
)=0
Fall 2001 16.31 13–7
? Wethencomparethis withthe desiredcharacteristicequationde-
velopedfromthedesiredclosed-looppolelocations:
Φ
d
(s)=s
3
+(α
1
)s
2
+(α
2
)s+(α
3
)=0
togetthat
a
1
+k
1
= α
1
.
.
.
a
n
+k
n
= α
n
?
?
?
k
1
= α
1
?a
1
.
.
.
k
n
= α
n
?a
n
? Togetthespeci?csoftheAckermannformula,wethen:
– TakeanarbitraryA,B andtransformittothecontrolcanonical
form(x?z = T
?1
x)
– Solve for the gains
?
K using the formulas above for the state z
(u =
?
Kz)
– Thenswitchbacktogainsneededforthestate x,sothat
K =
?
KT
?1
(u =
?
Kz= Kx)
? Pole placement is a very powerful tool and we will be using it for
mostofthiscourse.