You are on page 1of 5

ECCI-M6 – March 2012

Introduction to Nonlinear Control


Exercises

Problems 10 and 13 are 8 points each. Problems 18 and 19 are 12 points each. All other
problems are 4 points each. The total is 100 points. To validate the module, the student
needs to score 70/100. If the student does not achieve 70/100, he/she will receive comments
on the solution and will be allowed to resubmit the solution of some problems. The solution
has to be submitted as a pdf file and e-mailed to khalil@msu.edu by March 18, 2012. A
corrected solution has to be submitted within one week from receiving comments on the
earlier submission. The solution can be typed or handwritten.
1. Consider the system

ẋ1 = x2 , ẋ2 = x1 − 2 tan−1 (x1 + x2 )

(a) Find all equilibrium points and determine the type of each one.
(b) Construct the phase portrait using pplane and discuss the qualitative behavior of
the system.
2. Consider the negative resistance oscillator

ẋ1 = x2 , ẋ2 = −x1 − h0 (x1 )x2

with
h(v) = −v + v 3 − 15 v 5 + 1 7
105
v
Construct the phase portrait using pplane and discuss the qualitative behavior of
system.

3. Consider the system


ẋ1 = x2 , x2 = −g(x1 ) − h(x2 )
where g and h are continuously differentiable functions that satisfy g(0) = 0 and
h(0) = 0.

(a) Under what conditions on g and h will the origin be exponentially stable?
(b) Under what conditions on g and h will the origin be globally asymptotically
stable?
4. Euler equations for a rotating rigid spacecraft are given by

J1 ω̇1 = (J2 − J3 )ω2 ω3 + u1


J2 ω̇2 = (J3 − J1 )ω3 ω1 + u2
J3 ω̇3 = (J1 − J2 )ω1 ω2 + u3

where ω1 to ω3 are the components of the angular velocity vector ω along the principal
axes, u1 to u3 are the torque inputs applied about the principal axes, and J1 to J3 are
the principal moments of inertia.

1
(a) Show that with u1 = u2 = u3 = 0 the origin ω = 0 is stable. Is it asymptotically
stable?
(b) Suppose the torque inputs apply the feedback control ui = −ki ωi , where k1 to k3
are positive constants. Show that the origin of the closed-loop system is globally
asymptotically stable.

5. Consider the system

ẋ1 = −x2 x3 + 1, ẋ2 = x1 x3 − x2 , ẋ3 = x23 (1 − x3 )

(a) Show that the system has a unique equilibrium point.


(b) Show that the equilibrium point is asymptotically stable. Is it globally asymptot-
ically stable?

6. Consider a second-order system ẋ = f (x), together with a Lyapunov function V (x).


Suppose that V̇ (x) < 0 for all x21 + x22 ≥ a2 . The sketch below shows four different
directions of the vector field at a point on the circle x21 + x22 = a2 . Which of these
directions are possible and which are not? Justify your answer.

x2

2 1

3
4
x1

x2+x2=a2
1 2
V(x)=c

7. An RLC circuit with time-varying elements is represented by


1 1 R(t)
ẋ1 = x2 , ẋ2 = − x1 − x2
L(t) C(t) L(t)
where L(t), C(t), and R(t) are positive continuously differentiable functions. Consider
a Lyapunov function candidate
" #
2L(t) 2 2
V (t, x) = R(t) + x21 + 2x1 x2 + x
R(t)C(t) R(t) 2

Find conditions on L(t), C(t), R(t), L̇(t), Ċ(t), and Ṙ(t) that will ensure exponential
stability of the origin.

8. Investigate input-to-state stability of each of the following systems.


(a)

ẋ1 = x2 , ẋ2 = −x31 − x2 + u

2
(b)

ẋ1 = (x1 − x2 + u)(x21 − 1), ẋ2 = (x1 + x2 + u)(x21 − 1)

9. Consider the system

ẋ1 = x2 , ẋ2 = −h(x1 ) − ax2 + u, y = αx1 + x2

where 0 < α < a and h ∈ (0, ∞]. Show that the system is strictly passive.

10. Consider the feedback connection, shown below, with



(
 ẋ1
 = x2
ẋ3 = −x3 + e2
H1 : ẋ2 = −x1 − h1 (x2 ) + e1 and H2 :

 y2 = h2 (x3 )
y1 = x2

where h1 and h2 are locally Lipschitz functions, which satisfy h1 ∈ (0, ∞], h2 ∈ (0, ∞],
and |h2 (z)| ≥ |z|/(1 + z 2 ) for all z.

(a) Show that the feedback connection is passive.


(b) Show that the origin of the unforced system is globally asymptotically stable.

u1 - e
n 1-
y1 -
+ H1
−6

y2 e +
? u2
¾ ¾ 2 n¾+
H2

11. Study absolute stability for the transfer function


s
G(s) =
s2 −s+1
using the circle criterion. Find a sector [α, β] for which the system is absolutely stable.

12. Repeat the previous exercise using the Popov criterion.

13. Consider the system

ẋ1 = −x1 + x1 x2 , ẋ2 = x2 + x3 , ẋ3 = sin(x2 ) − x3 + u, y = x1 + x2

(a) Show that the system is input–output linearizable and transform it into the normal
form. Specify the region over which the transformation is valid.

3
(b) Is the system minimum phase?
(c) Design a state feedback control to stabilize the origin and estimate the region of
attraction.

14. Show that the system

ẋ1 = −x1 + x2 , ẋ2 = x1 − x2 − x1 x3 + u, ẋ3 = x1 + x1 x2 − 2x3 , y = x21 − 2x3

has relative degree 3 and design a state feedback control law to globally stabilize the
origin.

15. Using backstepping, design a state feedback controller to globally stabilize the origin
of the system
ẋ1 = x1 x2 , ẋ2 = x1 + u

16. Repeat Exercise 15 using passivity-based control.

17. Repeat Exercise 15 using continuously-implemented sliding mode control.

18. The equation of motion of a pendulum whose suspension point is subjected to a time-
varying, bounded, horizontal acceleration is given by

m`θ̈ + mg sin θ + k`θ̇ = T /` + mh(t) cos θ

where h is the horizontal acceleration and T is the torque input. Assume that

0.9 ≤ ` ≤ 1.1, 0.5 ≤ m ≤ 1.5, 0 ≤ k ≤ 0.2, |h(t)| ≤ 1

and g = 9.81. It is desired to stabilize the pendulum at θ = 0 for arbitrary initial


conditions θ(0) and θ̇(0).

(a) Design a continuously-implemented sliding mode state feedback controller to achieve


global ultimate boundedness with |θ| ≤ 0.01 and |θ̇| ≤ 0.01.
(b) Suppose you can only measure θ. Design a high-gain observer to implement the
controller of part (a) as output feedback controller. Can this controller achieve
global ultimate boundedness? If not, can it achieve semiglobal ultimate bound-
edness?.

19. The figure below shows a hydraulic system where liquid is stored in an open tank. The
cross-sectional area of the tank, A(h), is a function of h, the height Rof the liquid level
above the bottom of the tank. The liquid volume v is given by v = 0h A(λ) dλ. For a
liquid of density ρ, the absolute pressure p is given by p = ρgh + pa , where pa is the
atmospheric pressure (assumed constant) and g is the acceleration due to gravity. The
tank receives liquid at a flow rate√wi and loses liquid through a valve that obeys the
flow-pressure relationship wo = k ∆p. In the current case, ∆p = p − pa . Take u = wi
to be the control input and y = h to be the output.

4
(a) Using x = h as the state variable, show that the state model is given by
1 √
ẋ = (u − c x), y=x
A(x)

where c = k ρg.
(b) Find uss that is needed to maintain the output at a constant value r.
(c) Assuming that A(x) and c are precisely known, design a state feedback controller
to globally stabilize the output at r. 1
(d) Assuming that A(x) and c are uncertain but we know nominal models Â(x) and
ĉ and the following inequalities are satisfied for some known positive constants c1
to c4 . ¯ ¯
¯ A(x) − Â(x) ¯
¯ ¯
c1 ≤ A(x) ≤ c2 , |ĉ − c| ≤ c3 , ¯ ¯ ≤ c4 < 1
¯ A(x) ¯

Design a state feedback controller to ensure that the output y is globally regulated
to r.

wi $
C
AA
A
pa  
C AA 
AA 
A 
h AA 
A  k  wXo
AA p  
AA  @@
+p pa

You might also like