Scribd
Upload
116 views

Partial Feedback Linearization

1) The document discusses partial feedback linearization, which is a technique to transform a nonlinear system into a linear system through a change of variables and feedback. 2) One method presented is to transform the system such that one part is linearized while the other part depends only on the linearized part. 3) It is shown that if the origin of the independent part is asymptotically stable and the linearized part is controllable, then the origin of the overall system is asymptotically stable.

Uploaded by

zamzuri_rashid
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
116 views

Partial Feedback Linearization

1) The document discusses partial feedback linearization, which is a technique to transform a nonlinear system into a linear system through a change of variables and feedback. 2) One method presented is to transform the system such that one part is linearized while the other part depends only on the linearized part. 3) It is shown that if the origin of the independent part is asymptotically stable and the linearized part is controllable, then the origin of the overall system is asymptotically stable.

Uploaded by

zamzuri_rashid
Copyright
© Attribution Non-Commercial (BY-NC)
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 11

Nonlinear Systems and Control Lecture # 27 Stabilization Partial Feedback Linearization

p. 1/1

Consider the nonlinear system


x = f (x) + G(x)u [f (0) = 0]

Suppose there is a change of variables


z= = T (x) = T1 (x) T2 (x)

dened for all x D Rn , that transforms the system into


= f0 (, ) = A + B(x)[u (x)] (A, B) is controllable and (x) is nonsingular for all x D

p. 2/1

u = (x) + 1 (x)v = f0 (, ), = A + Bv

Suppose the origin of = f0 (, 0) is asymptotically stable


v = K,

where (A BK) is Hurwitz

Lemma 13.1: The origin of


= f0 (, ), = (A BK)

is asymptotically stable if the origin of = f0 (, 0) is asymptotically stable Proof:


V (, ) = V1 ()+k T P

p. 3/1

If the origin of = f0 (, 0) is globally asymptotically stable, will the origin of


= f0 (, ), = (A BK)

be globally asymptotically stable? In general No Example = + 2 , =v The origin of = is globally exponentially stable, but the origin of = + 2 , = k, k > 0 is not globally asymptotically stable. The region of attraction is { < 1 + k}

p. 4/1

Example
= 1 (1 + 2 ) 3 , 2 1 = 2 , 2 = v

The origin of = 1 3 is globally asymptotically stable 2


v = k 1 2k2 = K
2 def

ABK =

0 1 k2 2k

The eigenvalues of (A BK) are k and k kt kt (1 + kt)e te e(ABK)t = k2 tekt (1 kt)ekt

p. 5/1

Peaking Phenomenon:
max{k te
t 2 kt

}=

k e

as k

1 (0) = 1, 2 (0) = 0 2 (t) = k2 tekt =


1 2

1 k2 tekt 3 ,
2 0

(0) = 0

2 (t) =

2 1 + 0 [t + (1 + kt)ekt 1]

2 If 0 > 1, the system will have a nite escape time if k is chosen large enough

p. 6/1

Lemma 13.2: The origin of


= f0 (, ), = (A BK)

is globally asymptotically stable if the system = f0 (, ) is input-to-state stable Proof: Use Lemma 4.7: If x1 = f1 (x1 , x2 ) is ISS and the origin of x2 = f2 (x2 ) is globally asymptotically stable, then the origin of x1 = f1 (x1 , x2 ), x2 = f2 (x2 ) is globally asymptotically stable

p. 7/1

u = (x) 1 (x)KT2 (x)

What is the effect of uncertainty in , , and T2 ?


Let (x), (x), and T2 (x) be nominal models of (x), (x), and T2 (x) u = (x) 1 (x)K T2 (x) = f0 (, ), = (A BK) + B(z)

= [ + 1 KT2 1 K T2 ]

p. 8/1

Lemma 13.4 If (z) for all z and = f0 (, ) is input-to-state stable, then the state z is globally ultimately bounded by a class K function of If (z) k z in some neighborhood of z = 0, with sufciently small k, and the origin of = f0 (, 0) is exponentially stable, then z = 0 is an exponentially stable equilibrium point of the system
= f0 (, ), = (A BK) + B(z)

p. 9/1

ProofFirst Part: As in Lemma 13.3


(t)
c,

t t0
tt0

(t) 0 ( (t0 ) , t t0 ) + 0 (sup (t) ) (t) 0 ( (t0 ) , t t0 ) + 0 (c)

ProofSecond Part: c1
V1

V1 () c2
2

f0 (, 0) c3 V1 c4

p. 10/1

V (z) = bV1 () + T P V
T

Q=

bc3 (k P B + bc4 L/2) (k P B + bc4 L/2) 1 2k P B b=k Q is positive denite for sufciently small k

p. 11/1

You might also like